Move back to ARM format assembler files, but add a script to convert

from arm2gnu format. Update revised temporary makefile that now invokes
the conversion script as required.




git-svn-id: https://svn.xiph.org/branches/lowmem-branch/Tremolo@17254 0101bb08-14d6-0310-b084-bc0e0c8e3800

8 files changed, 2395 insertions, 2049 deletions

diff --git a/Makefile.rjw b/Makefile.rjw
new file mode 100644
index 0000000..c8acd47
--- /dev/null
+++ b/Makefile.rjw
@@ -0,0 +1,188 @@
+# Tremolo Makefile
+#
+# This is a temporary makefile used to test this branch until the merge
+# completes. It doesn't use the config system etc, so should be considered
+# a poor relation to doing it properly.
+#
+# Use:
+#
+#   make -f Makefile.rjw
+#
+# $URL$
+# $Id$
+
+srcdir    = .
+VPATH     = $(srcdir)
+
+CC     = arm-none-linux-gnueabi-gcc
+CXX    = arm-none-linux-gnueabi-g++
+LD     = arm-none-linux-gnueabi-g++
+AR     = arm-none-linux-gnueabi-ar cru
+RANLIB = arm-none-linux-gnueabi-ranlib
+STRIP  = arm-none-linux-gnueabi-strip
+WINDRES= arm-none-linux-gnueabi-windres
+MKDIR  = mkdir -p
+RM     = rm -f
+RM_REC = rm -rf
+ECHO   = echo -n
+CAT    = cat
+AS     = arm-none-linux-gnueabi-as
+
+DEFINES := 
+
+CFLAGS := -g -mcpu=cortex-a8 -mfpu=neon -I$(srcdir) -D__ARM__ -D_ARM_
+
+CXXFLAGS := $(CFLAGS)
+
+LDFLAGS :=
+LIBS :=
+
+OBJS := 
+MODULE_DIRS += .
+
+LIBOBJS := \
+	build/bitwise.o \
+	build/bitwiseARM.o \
+	build/codebook.o \
+	build/dsp.o \
+	build/floor0.o \
+	build/floor1.o \
+	build/floor1ARM.o \
+	build/floor_lookup.o \
+	build/framing.o \
+	build/info.o \
+	build/mapping0.o \
+	build/mdct.o \
+	build/mdctARM.o \
+	build/misc.o \
+	build/res012.o \
+	build/vorbisfile.o
+EXEOBJS := build/testtremor.o
+
+LIBOBJS_C := \
+	  build/bitwise.oc \
+	  build/codebook.oc \
+	  build/dsp.oc \
+	  build/floor0.oc \
+	  build/floor1.oc \
+	  build/floor_lookup.oc \
+	  build/framing.oc \
+	  build/info.oc \
+	  build/mapping0.oc \
+	  build/mdct.oc \
+	  build/misc.oc \
+	  build/res012.oc \
+	  build/vorbisfile.oc
+EXEOBJS_C := build/testtremor.oc
+
+LIBOBJS_L := \
+	  build/bitwise.ol \
+	  build/bitwiseARM.o \
+	  build/codebook.ol \
+	  build/dsp.ol \
+	  build/floor0.ol \
+	  build/floor1.ol \
+	  build/floor1LARM.o \
+	  build/floor_lookup.ol \
+	  build/framing.ol \
+	  build/info.ol \
+	  build/mapping0.ol \
+	  build/mdct.ol \
+	  build/mdctLARM.o \
+	  build/misc.ol \
+	  build/res012.ol \
+	  build/vorbisfile.ol
+EXEOBJS_L := build/testtremor.ol
+
+LIBOBJS_LC := \
+	   build/bitwise.olc \
+	   build/codebook.olc \
+	   build/dsp.olc \
+	   build/floor0.olc \
+	   build/floor1.olc \
+	   build/floor_lookup.olc \
+	   build/framing.olc \
+	   build/info.olc \
+	   build/mapping0.olc \
+	   build/mdct.olc \
+	   build/misc.olc \
+	   build/res012.olc \
+	   build/vorbisfile.olc
+EXEOBJS_LC := build/testtremor.olc
+
+# Rules
+.SUFFIXES: .o .oc .ol .olc
+
+# Compilation rule
+build/%.o: %.c
+	$(CC) $(CFLAGS) -c $< -o $@ -D_ARM_ASSEM_
+
+build/%.oc: %.c
+	$(CC) $(CFLAGS) -c $< -o $@
+
+build/%.ol: %.c
+	$(CC) $(CFLAGS) -c $< -o $@ -D_LOW_ACCURACY_ -D_ARM_ASSEM_
+
+build/%.olc: %.c
+	$(CC) $(CFLAGS) -c $< -o $@ -D_LOW_ACCURACY_
+
+# Assembly rule
+build/%.o: %.s
+	./arm2gnu.pl < $< > build/$*.S
+	$(CC) -Ibuild $(CFLAGS) -c build/$*.S -o $@
+	$(STRIP) -x $@
+
+all: libTremolo.lib libTremoloC.lib libTremoloL.lib libTremoloLC.lib \
+     testtremor.exe testtremorC.exe testtremorL.exe testtremorLC.exe
+
+libTremolo.lib: $(LIBOBJS)
+	$(AR) $@ $^
+	$(RANLIB) $@
+
+libTremoloC.lib: $(LIBOBJS_C)
+	$(AR) $@ $^
+	$(RANLIB) $@
+
+libTremoloL.lib: $(LIBOBJS_L)
+	$(AR) $@ $^
+	$(RANLIB) $@
+
+libTremoloLC.lib: $(LIBOBJS_LC)
+	$(AR) $@ $^
+	$(RANLIB) $@
+
+#bitwiseTEST.o: bitwise.c
+#	$(CC) $(CFLAGS) -c -o bitwiseTEST.o bitwise.c -D_V_BIT_TEST
+
+#bittest.exe: bitwiseTEST.o bitwiseARM.o
+#	$(LD) $^ $(LDFLAGS) $(LIBS) -o $@ -Wl,-Map,bittest.exe.map
+
+testtremor.exe: $(EXEOBJS) $(LIBOBJS)
+	$(LD) $^ $(LDFLAGS) $(LIBS) -o $@ -Wl,-Map,testtremor.exe.map
+
+testtremorC.exe: $(EXEOBJS_C) $(LIBOBJS_C)
+	$(LD) $^ $(LDFLAGS) $(LIBS) -o $@ -Wl,-Map,testtremorC.exe.map
+
+testtremorL.exe: $(EXEOBJS_L) $(LIBOBJS_L)
+	$(LD) $^ $(LDFLAGS) $(LIBS) -o $@ -Wl,-Map,testtremorL.exe.map
+
+testtremorLC.exe: $(EXEOBJS_LC) $(LIBOBJS_LC)
+	$(LD) $^ $(LDFLAGS) $(LIBS) -o $@ -Wl,-Map,testtremorLC.exe.map
+
+annotate.exe: annotate.c
+	gcc $^ -o $@
+
+clean:
+	@-rm build/* >& /dev/null
+	@-rm libTremolo.lib >& /dev/null
+	@-rm libTremoloC.lib >& /dev/null
+	@-rm libTremoloL.lib >& /dev/null
+	@-rm libTremoloLC.lib >& /dev/null
+	@-rm testtremor.exe >& /dev/null
+	@-rm testtremorC.exe >& /dev/null
+	@-rm testtremorL.exe >& /dev/null
+	@-rm testtremorLC.exe >& /dev/null
+	@-rm testtremor.exe.map >& /dev/null
+	@-rm testtremorC.exe.map >& /dev/null
+	@-rm testtremorL.exe.map >& /dev/null
+	@-rm testtremorLC.exe.map >& /dev/null
diff --git a/Makefile2 b/Makefile2
deleted file mode 100644
index 1b29142..0000000
--- a/Makefile2
+++ /dev/null
@@ -1,107 +0,0 @@
-# Tremolo Makefile for Windows CE port
-# Uses the VLC toolchain
-# $URL$
-# $Id$
-
-srcdir    = .
-VPATH     = $(srcdir)
-
-CC     = arm-none-linux-gnueabi-gcc
-CXX    = arm-none-linux-gnueabi-g++
-LD     = arm-none-linux-gnueabi-g++
-AR     = arm-none-linux-gnueabi-ar cru
-RANLIB = arm-none-linux-gnueabi-ranlib
-STRIP  = arm-none-linux-gnueabi-strip
-WINDRES= arm-none-linux-gnueabi-windres
-MKDIR  = mkdir -p
-RM     = rm -f
-RM_REC = rm -rf
-ECHO   = echo -n
-CAT    = cat
-AS     = arm-none-linux-gnueabi-as
-
-DEFINES := 
-
-CFLAGS := -g -mcpu=cortex-a8 -mfpu=neon -I$(srcdir) -D__ARM__ -D_ARM_
-
-CXXFLAGS := $(CFLAGS)
-
-#LDFLAGS := -L/opt/mingw32ce -Llibs/lib -lmingw32
-LDFLAGS :=
-LIBS :=
-
-OBJS := 
-MODULE_DIRS += .
-
-LIBOBJS := bitwise.o bitwiseARM.o codebook.o dsp.o floor0.o \
-           floor1.o floor1ARM.o floor_lookup.o framing.o info.o mapping0.o \
-           mdct.o mdctARM.o misc.o res012.o vorbisfile.o
-EXEOBJS := testtremor.o
-
-#LIBOBJS_C := bitwise.oc codebook.oc dsp.oc floor0.oc floor1.oc \
-#             floor_lookup.oc framing.oc info.oc mapping0.oc mdct.oc misc.oc \
-#             res012.oc vorbisfile.oc
-LIBOBJS_C := bitwise.oc codebook.oc dsp.oc floor0.oc \
-	     floor1.oc floor_lookup.oc framing.oc info.oc mapping0.oc \
-	     mdct.oc misc.oc res012.oc vorbisfile.oc
-EXEOBJS_C := testtremor.oc
-
-LIBOBJS_L := bitwise.ol bitwiseARM.o codebook.ol dsp.ol floor0.ol \
-             floor1.ol floor1LARM.o floor_lookup.ol framing.ol info.ol mapping0.ol \
-             mdct.ol mdctLARM.o misc.ol res012.ol vorbisfile.ol
-EXEOBJS_L := testtremor.ol
-
-LIBOBJS_LC := bitwise.olc codebook.olc dsp.olc floor0.olc floor1.olc \
-              floor_lookup.olc framing.olc info.olc mapping0.olc mdct.olc misc.olc \
-              res012.olc vorbisfile.olc
-EXEOBJS_LC := testtremor.olc
-
-# Rules
-.SUFFIXES: .o .oc .ol .olc
-
-.c.o:
-	$(CC) $(CFLAGS) -c $(<) -o $*.o -D_ARM_ASSEM_
-
-.c.oc:
-	$(CC) $(CFLAGS) -c $(<) -o $*.oc -DONLY_C
-
-.c.ol:
-	$(CC) $(CFLAGS) -c $(<) -o $*.ol -D_LOW_ACCURACY_ -D_ARM_ASSEM_
-
-.c.olc:
-	$(CC) $(CFLAGS) -c $(<) -o $*.olc -D_LOW_ACCURACY_ -DONLY_C
-.S.s:
-	./arm2gnu.pl < $(<) > $*.s
-
-all: libTremolo.lib testtremor.exe testtremorC.exe testtremorL.exe testtremorLC.exe
-
-libTremolo.lib: $(LIBOBJS)
-	$(AR) $@ $^
-	$(RANLIB) $@
-
-#bitwiseTEST.o: bitwise.c
-#	$(CC) $(CFLAGS) -c -o bitwiseTEST.o bitwise.c -D_V_BIT_TEST
-
-#bittest.exe: bitwiseTEST.o bitwiseARM.o
-#	$(LD) $^ $(LDFLAGS) $(LIBS) -o $@ -Wl,-Map,bittest.exe.map
-
-testtremor.exe: testtremor.o $(LIBOBJS)
-	$(LD) $^ $(LDFLAGS) $(LIBS) -o $@ -Wl,-Map,testtremor.exe.map
-
-testtremorC.exe: testtremor.oc $(LIBOBJS_C)
-	$(LD) $^ $(LDFLAGS) $(LIBS) -o $@ -Wl,-Map,testtremorC.exe.map
-
-testtremorL.exe: testtremor.ol $(LIBOBJS_L)
-	$(LD) $^ $(LDFLAGS) $(LIBS) -o $@ -Wl,-Map,testtremorL.exe.map
-
-testtremorLC.exe: testtremor.olc $(LIBOBJS_LC)
-	$(LD) $^ $(LDFLAGS) $(LIBS) -o $@ -Wl,-Map,testtremorLC.exe.map
-
-annotate.exe: annotate.c
-	gcc $^ -o $@
-
-clean:
-	rm `find . -name \*.o`
-	rm `find . -name \*.ol`
-	rm `find . -name \*.oc`
-	rm `find . -name \*.olc`
diff --git a/arm2gnu.pl b/arm2gnu.pl
new file mode 100755
index 0000000..ad12baa
--- /dev/null
+++ b/arm2gnu.pl
@@ -0,0 +1,265 @@
+#!/usr/bin/perl
+
+my $bigend;  # little/big endian
+
+eval 'exec /usr/local/bin/perl -S $0 ${1+"$@"}'
+    if $running_under_some_shell;
+
+while ($ARGV[0] =~ /^-/) {
+    $_ = shift;
+  last if /^--/;
+    if (/^-n/) {
+    $nflag++;
+    next;
+    }
+    die "I don't recognize this switch: $_\\n";
+}
+$printit++ unless $nflag;
+
+$\ = "\n";      # automatically add newline on print
+$n=0;
+
+$thumb = 0;     # ARM mode by default, not Thumb.
+
+LINE:
+while (<>) {
+
+    # For ADRLs we need to add a new line after the substituted one.
+    $addPadding = 0;
+
+    # First, we do not dare to touch *anything* inside double quotes, do we?
+    # Second, if you want a dollar character in the string,
+    # insert two of them -- that's how ARM C and assembler treat strings.
+    s/^([A-Za-z_]\w*)[ \t]+DCB[ \t]*\"/$1:   .ascii \"/   && do { s/\$\$/\$/g; next };
+    s/\bDCB\b[ \t]*\"/.ascii \"/                          && do { s/\$\$/\$/g; next };
+    s/^(\S+)\s+RN\s+(\S+)/$1 .req r$2/                    && do { s/\$\$/\$/g; next };
+    # If substituted -- leave immediately !
+
+    s/@/,:/;
+    s/;/@/;
+    while ( /@.*'/ ) {
+      s/(@.*)'/$1/g;
+    }
+    s/\{FALSE\}/0/g;
+    s/\{TRUE\}/1/g;
+    s/\{(\w\w\w\w+)\}/$1/g;
+    s/\bINCLUDE[ \t]*([^ \t\n]+)/.include \"$1\"/;
+    s/\bGET[ \t]*([^ \t\n]+)/.include \"$1\"/;
+    s/\bIMPORT\b/.extern/;
+    s/\bEXPORT\b/.global/;
+    s/^(\s+)\[/$1IF/;
+    s/^(\s+)\|/$1ELSE/;
+    s/^(\s+)\]/$1ENDIF/;
+    s/IF *:DEF:/ .ifdef/;
+    s/IF *:LNOT: *:DEF:/ .ifndef/;
+    s/ELSE/ .else/;
+    s/ENDIF/ .endif/;
+
+    if( /\bIF\b/ ) {
+      s/\bIF\b/ .if/;
+      s/=/==/;
+    }
+    if ( $n == 2) {
+        s/\$/\\/g;
+    }
+    if ($n == 1) {
+        s/\$//g;
+        s/label//g;
+    $n = 2;
+      }
+    if ( /MACRO/ ) {
+      s/MACRO *\n/.macro/;
+      $n=1;
+    }
+    if ( /\bMEND\b/ ) {
+      s/\bMEND\b/.endm/;
+      $n=0;
+    }
+
+    # ".rdata" doesn't work in 'as' version 2.13.2, as it is ".rodata" there.
+    #
+    if ( /\bAREA\b/ ) {
+        s/^(.+)CODE(.+)READONLY(.*)/    .text/;
+        s/^(.+)DATA(.+)READONLY(.*)/    .section .rdata\n    .align 2/;
+        s/^(.+)\|\|\.data\|\|(.+)/    .data\n    .align 2/;
+        s/^(.+)\|\|\.bss\|\|(.+)/    .bss/;
+    }
+
+    s/\|\|\.constdata\$(\d+)\|\|/.L_CONST$1/;       # ||.constdata$3||
+    s/\|\|\.bss\$(\d+)\|\|/.L_BSS$1/;               # ||.bss$2||
+    s/\|\|\.data\$(\d+)\|\|/.L_DATA$1/;             # ||.data$2||
+    s/\|\|([a-zA-Z0-9_]+)\@([a-zA-Z0-9_]+)\|\|/@ $&/;
+    s/^(\s+)\%(\s)/    .space $1/;
+
+    s/\|(.+)\.(\d+)\|/\.$1_$2/;                     # |L80.123| -> .L80_123
+    s/\bCODE32\b/.code 32/ && do {$thumb = 0};
+    s/\bCODE16\b/.code 16/ && do {$thumb = 1};
+    if (/\bPROC\b/)
+    {
+        print "    .thumb_func" if ($thumb);
+        s/\bPROC\b/@ $&/;
+    }
+    s/\bENDP\b/@ $&/;
+    s/\bSUBT\b/@ $&/;
+    s/\bDATA\b/@ $&/;   # DATA directive is deprecated -- Asm guide, p.7-25
+    s/\bKEEP\b/@ $&/;
+    s/\bEXPORTAS\b/@ $&/;
+    s/\|\|(.)+\bEQU\b/@ $&/;
+    s/\|\|([\w\$]+)\|\|/$1/;
+    s/\bENTRY\b/@ $&/;
+    s/\bASSERT\b/@ $&/;
+    s/\bGBLL\b/@ $&/;
+    s/\bGBLA\b/@ $&/;
+    s/^\W+OPT\b/@ $&/;
+    s/:OR:/|/g;
+    s/:SHL:/<</g;
+    s/:SHR:/>>/g;
+    s/:AND:/&/g;
+    s/:LAND:/&&/g;
+    s/CPSR/cpsr/;
+    s/SPSR/spsr/;
+    s/ALIGN$/.balign 4/;
+    s/psr_cxsf/psr_all/;
+    s/LTORG/.ltorg/;
+    s/^([A-Za-z_]\w*)[ \t]+EQU/ .set $1,/;
+    s/^([A-Za-z_]\w*)[ \t]+SETL/ .set $1,/;
+    s/^([A-Za-z_]\w*)[ \t]+SETA/ .set $1,/;
+    s/^([A-Za-z_]\w*)[ \t]+\*/ .set $1,/;
+
+    #  {PC} + 0xdeadfeed  -->  . + 0xdeadfeed
+    s/\{PC\} \+/ \. +/;
+
+    # Single hex constant on the line !
+    #
+    # >>> NOTE <<<
+    #   Double-precision floats in gcc are always mixed-endian, which means
+    #   bytes in two words are little-endian, but words are big-endian.
+    #   So, 0x0000deadfeed0000 would be stored as 0x0000dead at low address
+    #   and 0xfeed0000 at high address.
+    #
+    s/\bDCFD\b[ \t]+0x([a-fA-F0-9]{8})([a-fA-F0-9]{8})/.long 0x$1, 0x$2/;
+    # Only decimal constants on the line, no hex !
+    s/\bDCFD\b[ \t]+([0-9\.\-]+)/.double $1/;
+
+    # Single hex constant on the line !
+#    s/\bDCFS\b[ \t]+0x([a-f0-9]{8})([a-f0-9]{8})/.long 0x$1, 0x$2/;
+    # Only decimal constants on the line, no hex !
+#    s/\bDCFS\b[ \t]+([0-9\.\-]+)/.double $1/;
+    s/\bDCFS[ \t]+0x/.word 0x/;
+    s/\bDCFS\b/.float/;
+
+    s/^([A-Za-z_]\w*)[ \t]+DCD/$1 .word/;
+    s/\bDCD\b/.word/;
+    s/^([A-Za-z_]\w*)[ \t]+DCW/$1 .short/;
+    s/\bDCW\b/.short/;
+    s/^([A-Za-z_]\w*)[ \t]+DCB/$1 .byte/;
+    s/\bDCB\b/.byte/;
+    s/^([A-Za-z_]\w*)[ \t]+\%/.comm $1,/;
+    s/^[A-Za-z_\.]\w+/$&:/;
+    s/^(\d+)/$1:/;
+    s/\%(\d+)/$1b_or_f/;
+    s/\%[Bb](\d+)/$1b/;
+    s/\%[Ff](\d+)/$1f/;
+    s/\%[Ff][Tt](\d+)/$1f/;
+    s/&([\dA-Fa-f]+)/0x$1/;
+    if ( /\b2_[01]+\b/ ) {
+      s/\b2_([01]+)\b/conv$1&&&&/g;
+      while ( /[01][01][01][01]&&&&/ ) {
+        s/0000&&&&/&&&&0/g;
+        s/0001&&&&/&&&&1/g;
+        s/0010&&&&/&&&&2/g;
+        s/0011&&&&/&&&&3/g;
+        s/0100&&&&/&&&&4/g;
+        s/0101&&&&/&&&&5/g;
+        s/0110&&&&/&&&&6/g;
+        s/0111&&&&/&&&&7/g;
+        s/1000&&&&/&&&&8/g;
+        s/1001&&&&/&&&&9/g;
+        s/1010&&&&/&&&&A/g;
+        s/1011&&&&/&&&&B/g;
+        s/1100&&&&/&&&&C/g;
+        s/1101&&&&/&&&&D/g;
+        s/1110&&&&/&&&&E/g;
+        s/1111&&&&/&&&&F/g;
+      }
+      s/000&&&&/&&&&0/g;
+      s/001&&&&/&&&&1/g;
+      s/010&&&&/&&&&2/g;
+      s/011&&&&/&&&&3/g;
+      s/100&&&&/&&&&4/g;
+      s/101&&&&/&&&&5/g;
+      s/110&&&&/&&&&6/g;
+      s/111&&&&/&&&&7/g;
+      s/00&&&&/&&&&0/g;
+      s/01&&&&/&&&&1/g;
+      s/10&&&&/&&&&2/g;
+      s/11&&&&/&&&&3/g;
+      s/0&&&&/&&&&0/g;
+      s/1&&&&/&&&&1/g;
+      s/conv&&&&/0x/g;
+    }
+
+    if ( /commandline/)
+    {
+        if( /-bigend/)
+        {
+            $bigend=1;
+        }
+    }
+
+    if ( /\bDCDU\b/ )
+    {
+        my $cmd=$_;
+        my $value;
+        my $w1;
+        my $w2;
+        my $w3;
+        my $w4;
+
+        s/\s+DCDU\b/@ $&/;
+
+        $cmd =~ /\bDCDU\b\s+0x(\d+)/;
+        $value = $1;
+        $value =~ /(\w\w)(\w\w)(\w\w)(\w\w)/;
+        $w1 = $1;
+        $w2 = $2;
+        $w3 = $3;
+        $w4 = $4;
+
+        if( $bigend ne "")
+        {
+            # big endian
+
+            print "        .byte      0x".$w1;
+            print "        .byte      0x".$w2;
+            print "        .byte      0x".$w3;
+            print "        .byte      0x".$w4;
+        }
+        else
+        {
+            # little endian
+
+            print "        .byte      0x".$w4;
+            print "        .byte      0x".$w3;
+            print "        .byte      0x".$w2;
+            print "        .byte      0x".$w1;
+        }
+
+    }
+
+
+    if ( /\badrl\b/i )
+    {
+        s/\badrl\s+(\w+)\s*,\s*(\w+)/ldr $1,=$2/i;
+        $addPadding = 1;
+    }
+    s/\bEND\b/@ END/;
+} continue {
+    printf ("%s", $_) if $printit;
+    if ($addPadding != 0)
+    {
+        printf ("   mov r0,r0\n");
+        $addPadding = 0;
+    }
+}
+
diff --git a/bitwiseARM.s b/bitwiseARM.s
index 7a24aee..5043a02 100644
--- a/bitwiseARM.s
+++ b/bitwiseARM.s
@@ -1,80 +1,80 @@
-@ Tremolo library
-@ Copyright (C) 2009 Robin Watts for Pinknoise Productions Ltd
+; Tremolo library
+; Copyright (C) 2009 Robin Watts for Pinknoise Productions Ltd
 
-    .text
+	AREA	|.text|, CODE, READONLY
 
-	.global	oggpack_look
-	.global	oggpack_adv
-	.global	oggpack_readinit
-	.global	oggpack_read
+	EXPORT	oggpack_look
+	EXPORT	oggpack_adv
+	EXPORT	oggpack_readinit
+	EXPORT	oggpack_read
 
-oggpack_look:
-	@ r0 = oggpack_buffer *b
-	@ r1 = int             bits
+oggpack_look
+	; r0 = oggpack_buffer *b
+	; r1 = int             bits
 	STMFD	r13!,{r10,r11,r14}
 	LDMIA	r0,{r2,r3,r12}
-					@ r2 = bitsLeftInSegment
-					@ r3 = ptr
-					@ r12= bitsLeftInWord
-	SUBS	r2,r2,r1		@ bitsLeftinSegment -= bits
-	BLT	look_slow		@ Not enough bits in this segment for
-					@ this request. Do it slowly.
-	LDR	r10,[r3]		@ r10= ptr[0]
-	RSB	r14,r12,#32		@ r14= 32-bitsLeftInWord
-	SUBS	r12,r12,r1		@ r12= bitsLeftInWord -= bits
-	LDRLT	r11,[r3,#4]!		@ r11= ptr[1]
-	MOV	r10,r10,LSR r14		@ r10= ptr[0]>>(32-bitsLeftInWord)
-	ADDLE	r12,r12,#32		@ r12= bitsLeftInWord += 32
-	RSB	r14,r14,#32		@ r14= 32-bitsLeftInWord
-	ORRLT	r10,r10,r11,LSL r14	@ r10= Next 32 bits.
+					; r2 = bitsLeftInSegment
+					; r3 = ptr
+					; r12= bitsLeftInWord
+	SUBS	r2,r2,r1		; bitsLeftinSegment -= bits
+	BLT	look_slow		; Not enough bits in this segment for
+					; this request. Do it slowly.
+	LDR	r10,[r3]		; r10= ptr[0]
+	RSB	r14,r12,#32		; r14= 32-bitsLeftInWord
+	SUBS	r12,r12,r1		; r12= bitsLeftInWord -= bits
+	LDRLT	r11,[r3,#4]!		; r11= ptr[1]
+	MOV	r10,r10,LSR r14		; r10= ptr[0]>>(32-bitsLeftInWord)
+	ADDLE	r12,r12,#32		; r12= bitsLeftInWord += 32
+	RSB	r14,r14,#32		; r14= 32-bitsLeftInWord
+	ORRLT	r10,r10,r11,LSL r14	; r10= Next 32 bits.
 	MOV	r14,#1
 	RSB	r14,r14,r14,LSL r1
 	AND	r0,r10,r14
 	LDMFD	r13!,{r10,r11,PC}
 
-look_slow:
+look_slow
 	STMFD	r13!,{r5,r6}
-	ADDS	r10,r2,r1		@ r10= bitsLeftInSegment + bits (i.e.
-					@ the initial value of bitsLeftInSeg)
-	@ r10 = bitsLeftInSegment (initial)
-	@ r12 = bitsLeftInWord
-	RSB	r14,r12,#32		@ r14= 32-bitsLeftInWord
-	MOV	r5,r10			@ r5 = bitsLeftInSegment (initial)
+	ADDS	r10,r2,r1		; r10= bitsLeftInSegment + bits (i.e.
+					; the initial value of bitsLeftInSeg)
+	; r10 = bitsLeftInSegment (initial)
+	; r12 = bitsLeftInWord
+	RSB	r14,r12,#32		; r14= 32-bitsLeftInWord
+	MOV	r5,r10			; r5 = bitsLeftInSegment (initial)
 	BLT	look_overrun
-	BEQ	look_next_segment	@ r10= r12 = 0, if we branch
-	CMP	r12,r10			@ If bitsLeftInWord < bitsLeftInSeg
-					@ there must be more in the next word
-	LDR	r10,[r3],#4		@ r10= ptr[0]
-	LDRLT	r6,[r3]			@ r6 = ptr[1]
+	BEQ	look_next_segment	; r10= r12 = 0, if we branch
+	CMP	r12,r10			; If bitsLeftInWord < bitsLeftInSeg
+					; there must be more in the next word
+	LDR	r10,[r3],#4		; r10= ptr[0]
+	LDRLT	r6,[r3]			; r6 = ptr[1]
 	MOV	r11,#1
-	MOV	r10,r10,LSR r14		@ r10= first bitsLeftInWord bits
-	ORRLT	r10,r10,r6,LSL r12	@ r10= first bitsLeftInSeg bits+crap
-	RSB	r11,r11,r11,LSL r5	@ r11= mask
-	AND	r10,r10,r11		@ r10= first r5 bits
-	@ Load the next segments data
-look_next_segment:
-	@ At this point, r10 contains the first r5 bits of the result
-	LDR	r11,[r0,#12]		@ r11= head = b->head
-	@ Stall
-	@ Stall
-look_next_segment_2:
-	LDR	r11,[r11,#12]		@ r11= head = head->next
-	@ Stall
-	@ Stall
+	MOV	r10,r10,LSR r14		; r10= first bitsLeftInWord bits
+	ORRLT	r10,r10,r6,LSL r12	; r10= first bitsLeftInSeg bits+crap
+	RSB	r11,r11,r11,LSL r5	; r11= mask
+	AND	r10,r10,r11		; r10= first r5 bits
+	; Load the next segments data
+look_next_segment
+	; At this point, r10 contains the first r5 bits of the result
+	LDR	r11,[r0,#12]		; r11= head = b->head
+	; Stall
+	; Stall
+look_next_segment_2
+	LDR	r11,[r11,#12]		; r11= head = head->next
+	; Stall
+	; Stall
 	CMP	r11,#0
 	BEQ	look_out_of_data
-	LDMIA	r11,{r6,r12,r14}	@ r6 = buffer
-					@ r12= begin
-					@ r14= length
-	LDR	r6,[r6]			@ r6 = buffer->data
+	LDMIA	r11,{r6,r12,r14}	; r6 = buffer
+					; r12= begin
+					; r14= length
+	LDR	r6,[r6]			; r6 = buffer->data
 	CMP	r14,#0
 	BEQ	look_next_segment_2
-	ADD	r6,r6,r12		@ r6 = buffer->data+begin
-look_slow_loop:
-	LDRB	r12,[r6],#1		@ r12= *buffer
-	SUBS	r14,r14,#1		@ r14= length
-	@ Stall
-	ORR	r10,r10,r12,LSL r5	@ r10= first r5+8 bits
+	ADD	r6,r6,r12		; r6 = buffer->data+begin
+look_slow_loop
+	LDRB	r12,[r6],#1		; r12= *buffer
+	SUBS	r14,r14,#1		; r14= length
+	; Stall
+	ORR	r10,r10,r12,LSL r5	; r10= first r5+8 bits
 	ADD	r5,r5,#8
 	BLE	look_really_slow
 	CMP	r5,r1
@@ -85,7 +85,7 @@ look_slow_loop:
 	LDMFD	r13!,{r5,r6,r10,r11,PC}
 
 
-look_really_slow:
+look_really_slow
 	CMP	r5,r1
 	BLT	look_next_segment_2
 	MOV	r14,#1
@@ -93,208 +93,208 @@ look_really_slow:
 	AND	r0,r10,r14
 	LDMFD	r13!,{r5,r6,r10,r11,PC}
 
-look_out_of_data:
-	MVN	r0,#0			@ return -1
-	@MOV	r0,#0
+look_out_of_data
+	MVN	r0,#0			; return -1
+	;MOV	r0,#0
 	LDMFD	r13!,{r5,r6,r10,r11,PC}
 
-look_overrun:
-	@ We had overrun when we started, so we need to skip -r10 bits.
-	LDR	r11,[r0,#12]		@ r11 = head = b->head
-	@ stall
-	@ stall
-look_overrun_next_segment:
-	LDR	r11,[r11,#12]		@ r11 = head->next
-	@ stall
-	@ stall
+look_overrun
+	; We had overrun when we started, so we need to skip -r10 bits.
+	LDR	r11,[r0,#12]		; r11 = head = b->head
+	; stall
+	; stall
+look_overrun_next_segment
+	LDR	r11,[r11,#12]		; r11 = head->next
+	; stall
+	; stall
 	CMP	r11,#0
 	BEQ	look_out_of_data
-	LDMIA	r11,{r6,r7,r14}		@ r6 = buffer
-					@ r7 = begin
-					@ r14= length
-	LDR	r6,[r6]			@ r6 = buffer->data
-	@ stall
-	@ stall
-	ADD	r6,r6,r7		@ r6 = buffer->data+begin
-	MOV	r14,r14,LSL #3		@ r14= length in bits
-	ADDS	r14,r14,r10		@ r14= length in bits-bits to skip
+	LDMIA	r11,{r6,r7,r14}		; r6 = buffer
+					; r7 = begin
+					; r14= length
+	LDR	r6,[r6]			; r6 = buffer->data
+	; stall
+	; stall
+	ADD	r6,r6,r7		; r6 = buffer->data+begin
+	MOV	r14,r14,LSL #3		; r14= length in bits
+	ADDS	r14,r14,r10		; r14= length in bits-bits to skip
 	MOVLE	r10,r14
 	BLE	look_overrun_next_segment
-	RSB	r10,r10,#0		@ r10= bits to skip
-	ADD	r6,r10,r10,LSR #3	@ r6 = pointer to data
+	RSB	r10,r10,#0		; r10= bits to skip
+	ADD	r6,r10,r10,LSR #3	; r6 = pointer to data
 	MOV	r10,#0
 	B	look_slow_loop
 
-oggpack_adv:
-	@ r0 = oggpack_buffer *b
-	@ r1 = bits
+oggpack_adv
+	; r0 = oggpack_buffer *b
+	; r1 = bits
 	LDMIA	r0,{r2,r3,r12}
-					@ r2 = bitsLeftInSegment
-					@ r3 = ptr
-					@ r12= bitsLeftInWord
-	SUBS	r2,r2,r1		@ Does this run us out of bits in the
-	BLE	adv_slow		@ segment? If so, do it slowly
+					; r2 = bitsLeftInSegment
+					; r3 = ptr
+					; r12= bitsLeftInWord
+	SUBS	r2,r2,r1		; Does this run us out of bits in the
+	BLE	adv_slow		; segment? If so, do it slowly
 	SUBS	r12,r12,r1
 	ADDLE	r12,r12,#32
 	ADDLE	r3,r3,#4
 	STMIA	r0,{r2,r3,r12}
 	MOV	PC,R14
-adv_slow:
+adv_slow
 	STMFD	r13!,{r10,r14}
 
-	LDR	r14,[r0,#12]		@ r14= head
-	@ stall
-adv_slow_loop:
-	LDR	r1,[r0,#20]		@ r1 = count
-	LDR	r10,[r14,#8]		@ r10= head->length
-	LDR	r14,[r14,#12]		@ r14= head->next
-	@ stall
-	ADD	r1,r1,r10		@ r1 = count += head->length
+	LDR	r14,[r0,#12]		; r14= head
+	; stall
+adv_slow_loop
+	LDR	r1,[r0,#20]		; r1 = count
+	LDR	r10,[r14,#8]		; r10= head->length
+	LDR	r14,[r14,#12]		; r14= head->next
+	; stall
+	ADD	r1,r1,r10		; r1 = count += head->length
 	CMP	r14,#0
 	BEQ	adv_end
-	STR	r1,[r0,#20]		@ b->count = count
-	STR	r14,[r0,#12]		@ b->head = head
-	LDMIA	r14,{r3,r10,r12}	@ r3 = buffer
-					@ r10= begin
-					@ r12= length
-	LDR	r3,[r3]			@ r3 = buffer->data
-	ADD	r3,r3,r10		@ r3 = Pointer to start (byte)
-	AND	r10,r3,#3		@ r10= bytes to backtrk to word align
-	MOV	r10,r10,LSL #3		@ r10= bits to backtrk to word align
-	RSB	r10,r10,#32		@ r10= bits left in word
-	ADDS	r10,r10,r2		@ r10= bits left in word after skip
+	STR	r1,[r0,#20]		; b->count = count
+	STR	r14,[r0,#12]		; b->head = head
+	LDMIA	r14,{r3,r10,r12}	; r3 = buffer
+					; r10= begin
+					; r12= length
+	LDR	r3,[r3]			; r3 = buffer->data
+	ADD	r3,r3,r10		; r3 = Pointer to start (byte)
+	AND	r10,r3,#3		; r10= bytes to backtrk to word align
+	MOV	r10,r10,LSL #3		; r10= bits to backtrk to word align
+	RSB	r10,r10,#32		; r10= bits left in word
+	ADDS	r10,r10,r2		; r10= bits left in word after skip
 	ADDLE	r10,r10,#32
 	ADDLE	r3,r3,#4
-	BIC	r3,r3,#3		@ r3 = Pointer to start (word)
-	ADDS	r2,r2,r12,LSL #3	@ r2 = length in bits after advance
+	BIC	r3,r3,#3		; r3 = Pointer to start (word)
+	ADDS	r2,r2,r12,LSL #3	; r2 = length in bits after advance
 	BLE	adv_slow_loop
 	STMIA	r0,{r2,r3,r10}
 
 	LDMFD	r13!,{r10,PC}
-adv_end:
+adv_end
 	MOV	r2, #0
 	MOV	r12,#0
 	STMIA	r0,{r2,r3,r12}
 
 	LDMFD	r13!,{r10,PC}
 
-oggpack_readinit:
-	@ r0 = oggpack_buffer *b
-	@ r1 = oggreference   *r
-	STR	r1,[r0,#12]		@ b->head = r1
-	STR	r1,[r0,#16]		@ b->tail = r1
-	LDMIA	r1,{r2,r3,r12}		@ r2 = b->head->buffer
-					@ r3 = b->head->begin
-					@ r12= b->head->length
-	LDR	r2,[r2]			@ r2 = b->head->buffer->data
-	MOV	r1,r12,LSL #3		@ r1 = BitsInSegment
+oggpack_readinit
+	; r0 = oggpack_buffer *b
+	; r1 = oggreference   *r
+	STR	r1,[r0,#12]		; b->head = r1
+	STR	r1,[r0,#16]		; b->tail = r1
+	LDMIA	r1,{r2,r3,r12}		; r2 = b->head->buffer
+					; r3 = b->head->begin
+					; r12= b->head->length
+	LDR	r2,[r2]			; r2 = b->head->buffer->data
+	MOV	r1,r12,LSL #3		; r1 = BitsInSegment
 	MOV	r12,#0
-	ADD	r3,r2,r3		@ r3 = r2+b->head->begin
-	BIC	r2,r3,#3		@ r2 = b->headptr (word)
+	ADD	r3,r2,r3		; r3 = r2+b->head->begin
+	BIC	r2,r3,#3		; r2 = b->headptr (word)
 	AND	r3,r3,#3
 	MOV	r3,r3,LSL #3
-	RSB	r3,r3,#32		@ r3 = BitsInWord
+	RSB	r3,r3,#32		; r3 = BitsInWord
 	STMIA	r0,{r1,r2,r3}
 	STR	r12,[r0,#20]
 	MOV	PC,R14
 
-oggpack_read:
-	@ r0 = oggpack_buffer *b
-	@ r1 = int             bits
+oggpack_read
+	; r0 = oggpack_buffer *b
+	; r1 = int             bits
 	STMFD	r13!,{r10,r11,r14}
 	LDMIA	r0,{r2,r3,r12}
-					@ r2 = bitsLeftInSegment
-					@ r3 = ptr
-					@ r12= bitsLeftInWord
-	SUBS	r2,r2,r1		@ bitsLeftinSegment -= bits
-	BLT	read_slow		@ Not enough bits in this segment for
-					@ this request. Do it slowly.
-	LDR	r10,[r3]		@ r10= ptr[0]
-	RSB	r14,r12,#32		@ r14= 32-bitsLeftInWord
-	SUBS	r12,r12,r1		@ r12= bitsLeftInWord -= bits
+					; r2 = bitsLeftInSegment
+					; r3 = ptr
+					; r12= bitsLeftInWord
+	SUBS	r2,r2,r1		; bitsLeftinSegment -= bits
+	BLT	read_slow		; Not enough bits in this segment for
+					; this request. Do it slowly.
+	LDR	r10,[r3]		; r10= ptr[0]
+	RSB	r14,r12,#32		; r14= 32-bitsLeftInWord
+	SUBS	r12,r12,r1		; r12= bitsLeftInWord -= bits
 	ADDLE	r3,r3,#4
-	LDRLT	r11,[r3]		@ r11= ptr[1]
-	MOV	r10,r10,LSR r14		@ r10= ptr[0]>>(32-bitsLeftInWord)
-	ADDLE	r12,r12,#32		@ r12= bitsLeftInWord += 32
-	RSB	r14,r14,#32		@ r14= 32-bitsLeftInWord
-	ORRLT	r10,r10,r11,LSL r14	@ r10= Next 32 bits.
+	LDRLT	r11,[r3]		; r11= ptr[1]
+	MOV	r10,r10,LSR r14		; r10= ptr[0]>>(32-bitsLeftInWord)
+	ADDLE	r12,r12,#32		; r12= bitsLeftInWord += 32
+	RSB	r14,r14,#32		; r14= 32-bitsLeftInWord
+	ORRLT	r10,r10,r11,LSL r14	; r10= Next 32 bits.
 	STMIA	r0,{r2,r3,r12}
 	MOV	r14,#1
 	RSB	r14,r14,r14,LSL r1
 	AND	r0,r10,r14
 	LDMFD	r13!,{r10,r11,PC}
 
-read_slow:
+read_slow
 	STMFD	r13!,{r5,r6}
-	ADDS	r10,r2,r1		@ r10= bitsLeftInSegment + bits (i.e.
-					@ the initial value of bitsLeftInSeg)
-	@ r10 = bitsLeftInSegment (initial)
-	@ r12 = bitsLeftInWord
-	RSB	r14,r12,#32		@ r14= 32-bitsLeftInWord
-	MOV	r5,r10			@ r5 = bitsLeftInSegment (initial)
+	ADDS	r10,r2,r1		; r10= bitsLeftInSegment + bits (i.e.
+					; the initial value of bitsLeftInSeg)
+	; r10 = bitsLeftInSegment (initial)
+	; r12 = bitsLeftInWord
+	RSB	r14,r12,#32		; r14= 32-bitsLeftInWord
+	MOV	r5,r10			; r5 = bitsLeftInSegment (initial)
 	BLT	read_overrun
-	BEQ	read_next_segment	@ r10= r12 = 0, if we branch
-	CMP	r12,r10			@ If bitsLeftInWord < bitsLeftInSeg
-					@ there must be more in the next word
-	LDR	r10,[r3],#4		@ r10= ptr[0]
-	LDRLT	r6,[r3]			@ r6 = ptr[1]
+	BEQ	read_next_segment	; r10= r12 = 0, if we branch
+	CMP	r12,r10			; If bitsLeftInWord < bitsLeftInSeg
+					; there must be more in the next word
+	LDR	r10,[r3],#4		; r10= ptr[0]
+	LDRLT	r6,[r3]			; r6 = ptr[1]
 	MOV	r11,#1
-	MOV	r10,r10,LSR r14		@ r10= first bitsLeftInWord bits
-	ORRLT	r10,r10,r6,LSL r12	@ r10= first bitsLeftInSeg bits+crap
-	RSB	r11,r11,r11,LSL r5	@ r11= mask
-	AND	r10,r10,r11		@ r10= first r5 bits
-	@ Load the next segments data
-read_next_segment:
-	@ At this point, r10 contains the first r5 bits of the result
-	LDR	r11,[r0,#12]		@ r11= head = b->head
-	@ Stall
-read_next_segment_2:
-	@ r11 = head
-	LDR	r6,[r0,#20]		@ r6 = count
-	LDR	r12,[r11,#8]		@ r12= length
-	LDR	r11,[r11,#12]		@ r11= head = head->next
-	@ Stall
-	ADD	r6,r6,r12		@ count += length
+	MOV	r10,r10,LSR r14		; r10= first bitsLeftInWord bits
+	ORRLT	r10,r10,r6,LSL r12	; r10= first bitsLeftInSeg bits+crap
+	RSB	r11,r11,r11,LSL r5	; r11= mask
+	AND	r10,r10,r11		; r10= first r5 bits
+	; Load the next segments data
+read_next_segment
+	; At this point, r10 contains the first r5 bits of the result
+	LDR	r11,[r0,#12]		; r11= head = b->head
+	; Stall
+read_next_segment_2
+	; r11 = head
+	LDR	r6,[r0,#20]		; r6 = count
+	LDR	r12,[r11,#8]		; r12= length
+	LDR	r11,[r11,#12]		; r11= head = head->next
+	; Stall
+	ADD	r6,r6,r12		; count += length
 	CMP	r11,#0
 	BEQ	read_out_of_data
 	STR	r11,[r0,#12]
-	STR	r6,[r0,#20]		@ b->count = count
-	LDMIA	r11,{r6,r12,r14}	@ r6 = buffer
-					@ r12= begin
-					@ r14= length
-	LDR	r6,[r6]			@ r6 = buffer->data
+	STR	r6,[r0,#20]		; b->count = count
+	LDMIA	r11,{r6,r12,r14}	; r6 = buffer
+					; r12= begin
+					; r14= length
+	LDR	r6,[r6]			; r6 = buffer->data
 	CMP	r14,#0
 	BEQ	read_next_segment_2
-	ADD	r6,r6,r12		@ r6 = buffer->data+begin
-read_slow_loop:
-	LDRB	r12,[r6],#1		@ r12= *buffer
-	SUBS	r14,r14,#1		@ r14= length
-	@ Stall
-	ORR	r10,r10,r12,LSL r5	@ r10= first r5+8 bits
+	ADD	r6,r6,r12		; r6 = buffer->data+begin
+read_slow_loop
+	LDRB	r12,[r6],#1		; r12= *buffer
+	SUBS	r14,r14,#1		; r14= length
+	; Stall
+	ORR	r10,r10,r12,LSL r5	; r10= first r5+8 bits
 	ADD	r5,r5,#8
 	BLE	read_really_slow
 	CMP	r5,r1
 	BLT	read_slow_loop
-read_end:
+read_end
 	MOV	r12,#1
 	RSB	r12,r12,r12,LSL r1
 
-	@ Store back the new position
-	@ r2 = -number of bits to go from this segment
-	@ r6 = ptr
-	@ r14= bytesLeftInSegment
-	@ r11= New head value
-	LDMIA	r11,{r3,r6,r14}		@ r3 = buffer
-					@ r6 = begin
-					@ r14= length
-	LDR	r3,[r3]			@ r3 = buffer->data
-	ADD	r1,r2,r14,LSL #3	@ r1 = bitsLeftInSegment
-	@ stall
-	ADD	r6,r3,r6		@ r6 = pointer
-	AND	r3,r6,#3		@ r3 = bytes used in first word
-	RSB	r3,r2,r3,LSL #3		@ r3 = bits used in first word
-	BIC	r2,r6,#3		@ r2 = word ptr
-	RSBS	r3,r3,#32		@ r3 = bitsLeftInWord
+	; Store back the new position
+	; r2 = -number of bits to go from this segment
+	; r6 = ptr
+	; r14= bytesLeftInSegment
+	; r11= New head value
+	LDMIA	r11,{r3,r6,r14}		; r3 = buffer
+					; r6 = begin
+					; r14= length
+	LDR	r3,[r3]			; r3 = buffer->data
+	ADD	r1,r2,r14,LSL #3	; r1 = bitsLeftInSegment
+	; stall
+	ADD	r6,r3,r6		; r6 = pointer
+	AND	r3,r6,#3		; r3 = bytes used in first word
+	RSB	r3,r2,r3,LSL #3		; r3 = bits used in first word
+	BIC	r2,r6,#3		; r2 = word ptr
+	RSBS	r3,r3,#32		; r3 = bitsLeftInWord
 	ADDLE	r3,r3,#32
 	ADDLE	r2,r2,#4
 	STMIA	r0,{r1,r2,r3}
@@ -303,66 +303,66 @@ read_end:
 	LDMFD	r13!,{r5,r6,r10,r11,PC}
 
 
-read_really_slow:
+read_really_slow
 	CMP	r5,r1
 	BGE	read_end
-	LDR	r14,[r11,#8]		@ r14= length of segment just done
-	@ stall
-	@ stall
-	ADD	r2,r2,r14,LSL #3	@ r2 = -bits to use from next seg
+	LDR	r14,[r11,#8]		; r14= length of segment just done
+	; stall
+	; stall
+	ADD	r2,r2,r14,LSL #3	; r2 = -bits to use from next seg
 	B	read_next_segment_2
 
-read_out_of_data:
-	@ Store back the new position
-	@ r2 = -number of bits to go from this segment
-	@ r6 = ptr
-	@ r14= bytesLeftInSegment
-	@ RJW: This may be overkill - we leave the buffer empty, with -1
-	@ bits left in it. We might get away with just storing the
-	@ bitsLeftInSegment as -1.
-	LDR	r11,[r0,#12]		@ r11=head
+read_out_of_data
+	; Store back the new position
+	; r2 = -number of bits to go from this segment
+	; r6 = ptr
+	; r14= bytesLeftInSegment
+	; RJW: This may be overkill - we leave the buffer empty, with -1
+	; bits left in it. We might get away with just storing the
+	; bitsLeftInSegment as -1.
+	LDR	r11,[r0,#12]		; r11=head
 
-	LDMIA	r11,{r3,r6,r14}		@ r3 = buffer
-					@ r6 = begin
-					@ r14= length
-	LDR	r3,[r3]			@ r3 = buffer->data
-	ADD	r6,r3,r6		@ r6 = pointer
+	LDMIA	r11,{r3,r6,r14}		; r3 = buffer
+					; r6 = begin
+					; r14= length
+	LDR	r3,[r3]			; r3 = buffer->data
+	ADD	r6,r3,r6		; r6 = pointer
 	ADD	r6,r6,r14
-	AND	r3,r6,#3		@ r3 = bytes used in first word
-	MOV	r3,r3,LSL #3		@ r3 = bits used in first word
-	BIC	r2,r6,#3		@ r2 = word ptr
-	RSBS	r3,r3,#32		@ r3 = bitsLeftInWord
-	MVN	r1,#0			@ r1 = -1 = bitsLeftInSegment
+	AND	r3,r6,#3		; r3 = bytes used in first word
+	MOV	r3,r3,LSL #3		; r3 = bits used in first word
+	BIC	r2,r6,#3		; r2 = word ptr
+	RSBS	r3,r3,#32		; r3 = bitsLeftInWord
+	MVN	r1,#0			; r1 = -1 = bitsLeftInSegment
 	STMIA	r0,{r1,r2,r3}
-	MVN	r0,#0			@ return -1
-	@MOV	r0,#0
+	MVN	r0,#0			; return -1
+	;MOV	r0,#0
 	LDMFD	r13!,{r5,r6,r10,r11,PC}
 
-read_overrun:
-	@ We had overrun when we started, so we need to skip -r10 bits.
-	LDR	r11,[r0,#12]		@ r11 = head = b->head
-	@ stall
-	@ stall
-read_overrun_next_segment:
-	LDR	r11,[r11,#12]		@ r11 = head->next
-	@ stall
-	@ stall
+read_overrun
+	; We had overrun when we started, so we need to skip -r10 bits.
+	LDR	r11,[r0,#12]		; r11 = head = b->head
+	; stall
+	; stall
+read_overrun_next_segment
+	LDR	r11,[r11,#12]		; r11 = head->next
+	; stall
+	; stall
 	CMP	r11,#0
 	BEQ	read_out_of_data
-	LDMIA	r11,{r6,r7,r14}		@ r6 = buffer
-					@ r7 = begin
-					@ r14= length
-	LDR	r6,[r6]			@ r6 = buffer->data
-	@ stall
-	@ stall
-	ADD	r6,r6,r7		@ r6 = buffer->data+begin
-	MOV	r14,r14,LSL #3		@ r14= length in bits
-	ADDS	r14,r14,r10		@ r14= length in bits-bits to skip
+	LDMIA	r11,{r6,r7,r14}		; r6 = buffer
+					; r7 = begin
+					; r14= length
+	LDR	r6,[r6]			; r6 = buffer->data
+	; stall
+	; stall
+	ADD	r6,r6,r7		; r6 = buffer->data+begin
+	MOV	r14,r14,LSL #3		; r14= length in bits
+	ADDS	r14,r14,r10		; r14= length in bits-bits to skip
 	MOVLE	r10,r14
 	BLE	read_overrun_next_segment
-	RSB	r10,r10,#0		@ r10= bits to skip
-	ADD	r6,r10,r10,LSR #3	@ r6 = pointer to data
+	RSB	r10,r10,#0		; r10= bits to skip
+	ADD	r6,r10,r10,LSR #3	; r6 = pointer to data
 	MOV	r10,#0
 	B	read_slow_loop
 
-	@ END
+	END
diff --git a/floor1ARM.s b/floor1ARM.s
index f7f7ae1..689b5c5 100644
--- a/floor1ARM.s
+++ b/floor1ARM.s
@@ -1,36 +1,36 @@
-@ Tremolo library
-@ Copyright (C) 2009 Robin Watts for Pinknoise Productions Ltd
+; Tremolo library
+; Copyright (C) 2009 Robin Watts for Pinknoise Productions Ltd
 
-    .text
+	AREA	|.text|, CODE, READONLY
 
-	.global	render_line_arm
+	EXPORT	render_line_arm
 
-render_line_arm:
-	@ r0 = n
-	@ r1 = d
-	@ r2 = floor
-	@ r3 = base
-	@ <> = err
-	@ <> = adx
-	@ <> = ady
+render_line_arm
+	; r0 = n
+	; r1 = d
+	; r2 = floor
+	; r3 = base
+	; <> = err
+	; <> = adx
+	; <> = ady
 	MOV	r12,r13
 	STMFD	r13!,{r4-r6,r11,r14}
-	LDMFD	r12,{r11,r12,r14}	@ r11 = err
-					@ r12 = adx
-					@ r14 = ady
-rl_loop:
-	LDR	r4,[r1]			@ r4 = *d
-	LDR	r5,[r2],r3,LSL #2	@ r5 = *floor    r2 = floor+base
-	SUBS	r11,r11,r14		@ err -= ady
-	ADDLT	r11,r11,r12		@ if (err < 0) err+=adx
-	SMULL	r6, r5, r4, r5		@ (r6,r5) = *d * *floor
-	ADDLT	r2, r2, #4		@              floor+=1
+	LDMFD	r12,{r11,r12,r14}	; r11 = err
+					; r12 = adx
+					; r14 = ady
+rl_loop
+	LDR	r4,[r1]			; r4 = *d
+	LDR	r5,[r2],r3,LSL #2	; r5 = *floor    r2 = floor+base
+	SUBS	r11,r11,r14		; err -= ady
+	ADDLT	r11,r11,r12		; if (err < 0) err+=adx
+	SMULL	r6, r5, r4, r5		; (r6,r5) = *d * *floor
+	ADDLT	r2, r2, #4		;              floor+=1
 	MOVS	r6, r6, LSR #15
-	ADC	r5, r6, r5, LSL #17	@ r5 = MULT31_SHIFT15
+	ADC	r5, r6, r5, LSL #17	; r5 = MULT31_SHIFT15
 	STR	r5,[r1],#4
 	SUBS	r0, r0, #1
 	BGT	rl_loop
 
 	LDMFD	r13!,{r4-r6,r11,PC}
 
-	@ END
+	END
diff --git a/floor1LARM.s b/floor1LARM.s
index 21163ae..d7ead1d 100644
--- a/floor1LARM.s
+++ b/floor1LARM.s
@@ -1,35 +1,35 @@
-@ Tremolo library
-@ Copyright (C) 2009 Robin Watts for Pinknoise Productions Ltd
+; Tremolo library
+; Copyright (C) 2009 Robin Watts for Pinknoise Productions Ltd
 
-    .text
+	AREA	|.text|, CODE, READONLY
 
-	.global	render_line_arm_low
+	EXPORT	render_line_arm_low
 
-render_line_arm_low:
-	@ r0 = n
-	@ r1 = d
-	@ r2 = floor
-	@ r3 = base
-	@ <> = err
-	@ <> = adx
-	@ <> = ady
+render_line_arm_low
+	; r0 = n
+	; r1 = d
+	; r2 = floor
+	; r3 = base
+	; <> = err
+	; <> = adx
+	; <> = ady
 	MOV	r12,r13
 	STMFD	r13!,{r4-r6,r11,r14}
-	LDMFD	r12,{r11,r12,r14}	@ r11 = err
-					@ r12 = adx
-					@ r14 = ady
-rl_loop:
-	LDR	r4, [r1]		@ r4 = *d
-	LDR	r5, [r2], r3,LSL #2	@ r5 = *floor    r2 = floor+base
-	SUBS	r11,r11,r14		@ err -= ady
+	LDMFD	r12,{r11,r12,r14}	; r11 = err
+					; r12 = adx
+					; r14 = ady
+rl_loop
+	LDR	r4, [r1]		; r4 = *d
+	LDR	r5, [r2], r3,LSL #2	; r5 = *floor    r2 = floor+base
+	SUBS	r11,r11,r14		; err -= ady
 	MOV	r4, r4, ASR #6
-	MUL	r5, r4, r5		@ r5 = MULT31_SHIFT15
-	ADDLT	r11,r11,r12		@ if (err < 0) err+=adx
-	ADDLT	r2, r2, #4		@              floor+=1
+	MUL	r5, r4, r5		; r5 = MULT31_SHIFT15
+	ADDLT	r11,r11,r12		; if (err < 0) err+=adx
+	ADDLT	r2, r2, #4		;              floor+=1
 	SUBS	r0, r0, #1
 	STR	r5, [r1], #4
 	BGT	rl_loop
 
 	LDMFD	r13!,{r4-r6,r11,PC}
 
-	@ END
+	END
diff --git a/mdctARM.s b/mdctARM.s
index 436a527..a2dc5bb 100644
--- a/mdctARM.s
+++ b/mdctARM.s
@@ -1,92 +1,92 @@
-@ Tremolo library
-@ Copyright (C) 2009 Robin Watts for Pinknoise Productions Ltd
+; Tremolo library
+; Copyright (C) 2009 Robin Watts for Pinknoise Productions Ltd
 
-    .text
+	AREA	|.text|, CODE, READONLY
 
-	@ full accuracy version
+	; full accuracy version
 
-	.global mdct_backward_arm
-	.global mdct_shift_right_arm
-	.global mdct_unroll_prelap_arm
-	.global mdct_unroll_part2_arm
-	.global mdct_unroll_part3_arm
-	.global mdct_unroll_postlap_arm
+	EXPORT mdct_backward_arm
+	EXPORT mdct_shift_right_arm
+	EXPORT mdct_unroll_prelap_arm
+	EXPORT mdct_unroll_part2_arm
+	EXPORT mdct_unroll_part3_arm
+	EXPORT mdct_unroll_postlap_arm
 
-	.extern	sincos_lookup0
-	.extern	sincos_lookup1
+	IMPORT	sincos_lookup0
+	IMPORT	sincos_lookup1
 
-mdct_unroll_prelap_arm:
-	@ r0 = out
-	@ r1 = post
-	@ r2 = r
-	@ r3 = step
+mdct_unroll_prelap_arm
+	; r0 = out
+	; r1 = post
+	; r2 = r
+	; r3 = step
 	STMFD	r13!,{r4-r7,r14}
 	MVN	r4, #0x8000
 	MOV	r3, r3, LSL #1
-	SUB	r1, r2, r1		@ r1 = r - post
-	SUBS	r1, r1, #16		@ r1 = r - post - 16
+	SUB	r1, r2, r1		; r1 = r - post
+	SUBS	r1, r1, #16		; r1 = r - post - 16
 	BLT	unroll_over
-unroll_loop:
+unroll_loop
 	LDMDB	r2!,{r5,r6,r7,r12}
 
-	MOV	r5, r5, ASR #9		@ r5 = (*--r)>>9
-	MOV	r6, r6, ASR #9		@ r6 = (*--r)>>9
-	MOV	r7, r7, ASR #9		@ r7 = (*--r)>>9
-	MOV	r12,r12,ASR #9		@ r12= (*--r)>>9
+	MOV	r5, r5, ASR #9		; r5 = (*--r)>>9
+	MOV	r6, r6, ASR #9		; r6 = (*--r)>>9
+	MOV	r7, r7, ASR #9		; r7 = (*--r)>>9
+	MOV	r12,r12,ASR #9		; r12= (*--r)>>9
 
 	MOV	r14,r12,ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r12,r4, r14,ASR #31
 	STRH	r12,[r0], r3
 
 	MOV	r14,r7, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r7, r4, r14,ASR #31
 	STRH	r7, [r0], r3
 
 	MOV	r14,r6, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r6, r4, r14,ASR #31
 	STRH	r6, [r0], r3
 
 	MOV	r14,r5, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r5, r4, r14,ASR #31
 	STRH	r5, [r0], r3
 
 	SUBS	r1, r1, #16
 	BGE	unroll_loop
 
-unroll_over:
+unroll_over
 	ADDS	r1, r1, #16
 	BLE	unroll_end
-unroll_loop2:
+unroll_loop2
 	LDR	r5,[r2,#-4]!
-	@ stall
-	@ stall (Xscale)
-	MOV	r5, r5, ASR #9		@ r5 = (*--r)>>9
+	; stall
+	; stall (Xscale)
+	MOV	r5, r5, ASR #9		; r5 = (*--r)>>9
 	MOV	r14,r5, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r5, r4, r14,ASR #31
 	STRH	r5, [r0], r3
 	SUBS	r1, r1, #4
 	BGT	unroll_loop2
-unroll_end:
+unroll_end
 	LDMFD	r13!,{r4-r7,PC}
 
-mdct_unroll_postlap_arm:
-	@ r0 = out
-	@ r1 = post
-	@ r2 = l
-	@ r3 = step
+mdct_unroll_postlap_arm
+	; r0 = out
+	; r1 = post
+	; r2 = l
+	; r3 = step
 	STMFD	r13!,{r4-r7,r14}
 	MVN	r4, #0x8000
 	MOV	r3, r3, LSL #1
-	SUB	r1, r1, r2		@ r1 = post - l
-	MOV	r1, r1, ASR #1		@ r1 = (post - l)>>1
-	SUBS	r1, r1, #16		@ r1 = ((post - l)>>1) - 4
+	SUB	r1, r1, r2		; r1 = post - l
+	MOV	r1, r1, ASR #1		; r1 = (post - l)>>1
+	SUBS	r1, r1, #16		; r1 = ((post - l)>>1) - 4
 	BLT	unroll_over3
-unroll_loop3:
+unroll_loop3
 	LDR	r12,[r2],#8
 	LDR	r7, [r2],#8
 	LDR	r6, [r2],#8
@@ -97,142 +97,142 @@ unroll_loop3:
 	RSB	r6, r6, #0
 	RSB	r7, r7, #0
 
-	MOV	r12, r12,ASR #9		@ r12= (-*l)>>9
-	MOV	r5,  r5, ASR #9		@ r5 = (-*l)>>9
-	MOV	r6,  r6, ASR #9		@ r6 = (-*l)>>9
-	MOV	r7,  r7, ASR #9		@ r7 = (-*l)>>9
+	MOV	r12, r12,ASR #9		; r12= (-*l)>>9
+	MOV	r5,  r5, ASR #9		; r5 = (-*l)>>9
+	MOV	r6,  r6, ASR #9		; r6 = (-*l)>>9
+	MOV	r7,  r7, ASR #9		; r7 = (-*l)>>9
 
 	MOV	r14,r12,ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r12,r4, r14,ASR #31
 	STRH	r12,[r0], r3
 
 	MOV	r14,r7, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r7, r4, r14,ASR #31
 	STRH	r7, [r0], r3
 
 	MOV	r14,r6, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r6, r4, r14,ASR #31
 	STRH	r6, [r0], r3
 
 	MOV	r14,r5, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r5, r4, r14,ASR #31
 	STRH	r5, [r0], r3
 
 	SUBS	r1, r1, #16
 	BGE	unroll_loop3
 
-unroll_over3:
+unroll_over3
 	ADDS	r1, r1, #16
 	BLE	unroll_over4
-unroll_loop4:
+unroll_loop4
 	LDR	r5,[r2], #8
-	@ stall
-	@ stall (Xscale)
+	; stall
+	; stall (Xscale)
 	RSB	r5, r5, #0
-	MOV	r5, r5, ASR #9		@ r5 = (-*l)>>9
+	MOV	r5, r5, ASR #9		; r5 = (-*l)>>9
 	MOV	r14,r5, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r5, r4, r14,ASR #31
 	STRH	r5, [r0], r3
 	SUBS	r1, r1, #4
 	BGT	unroll_loop4
-unroll_over4:
+unroll_over4
 	LDMFD	r13!,{r4-r7,PC}
 
-mdct_unroll_part2_arm:
-	@ r0 = out
-	@ r1 = post
-	@ r2 = l
-	@ r3 = r
-	@ <> = step
-	@ <> = wL
-	@ <> = wR
+mdct_unroll_part2_arm
+	; r0 = out
+	; r1 = post
+	; r2 = l
+	; r3 = r
+	; <> = step
+	; <> = wL
+	; <> = wR
 	MOV	r12,r13
 	STMFD	r13!,{r4,r6-r11,r14}
-	LDMFD	r12,{r8,r9,r10}		@ r8 = step
-					@ r9 = wL
-					@ r10= wR
+	LDMFD	r12,{r8,r9,r10}		; r8 = step
+					; r9 = wL
+					; r10= wR
 	MVN	r4, #0x8000
 	MOV	r8, r8, LSL #1
-	SUBS	r1, r3, r1		@ r1 = (r - post)
+	SUBS	r1, r3, r1		; r1 = (r - post)
 	BLE	unroll_over5
-unroll_loop5:
-	LDR	r12,[r2, #-8]!		@ r12= *l       (but l -= 2 first)
-	LDR	r11,[r9],#4		@ r11= *wL++
-	LDR	r7, [r3, #-4]!		@ r7 = *--r
-	LDR	r6, [r10,#-4]!		@ r6 = *--wR
-
-	@ Can save a cycle here, at the cost of 1bit errors in rounding
-	SMULL	r14,r11,r12,r11		@ (r14,r11)  = *l   * *wL++
-	SMULL	r14,r6, r7, r6		@ (r14,r6)   = *--r * *--wR
+unroll_loop5
+	LDR	r12,[r2, #-8]!		; r12= *l       (but l -= 2 first)
+	LDR	r11,[r9],#4		; r11= *wL++
+	LDR	r7, [r3, #-4]!		; r7 = *--r
+	LDR	r6, [r10,#-4]!		; r6 = *--wR
+
+	; Can save a cycle here, at the cost of 1bit errors in rounding
+	SMULL	r14,r11,r12,r11		; (r14,r11)  = *l   * *wL++
+	SMULL	r14,r6, r7, r6		; (r14,r6)   = *--r * *--wR
 	ADD	r6, r6, r11
 	MOV	r6, r6, ASR #8
 	MOV	r14,r6, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r6, r4, r14,ASR #31
 	STRH	r6, [r0], r8
 
 	SUBS	r1, r1, #4
 	BGT	unroll_loop5
 
-unroll_over5:
+unroll_over5
 	LDMFD	r13!,{r4,r6-r11,PC}
 
-mdct_unroll_part3_arm:
-	@ r0 = out
-	@ r1 = post
-	@ r2 = l
-	@ r3 = r
-	@ <> = step
-	@ <> = wL
-	@ <> = wR
+mdct_unroll_part3_arm
+	; r0 = out
+	; r1 = post
+	; r2 = l
+	; r3 = r
+	; <> = step
+	; <> = wL
+	; <> = wR
 	MOV	r12,r13
 	STMFD	r13!,{r4,r6-r11,r14}
-	LDMFD	r12,{r8,r9,r10}		@ r8 = step
-					@ r9 = wL
-					@ r10= wR
+	LDMFD	r12,{r8,r9,r10}		; r8 = step
+					; r9 = wL
+					; r10= wR
 	MVN	r4, #0x8000
 	MOV	r8, r8, LSL #1
-	SUBS	r1, r1, r3		@ r1 = (post - r)
+	SUBS	r1, r1, r3		; r1 = (post - r)
 	BLE	unroll_over6
-unroll_loop6:
-	LDR	r12,[r2],#8		@ r12= *l       (but l += 2 first)
-	LDR	r11,[r9],#4		@ r11= *wL++
-	LDR	r7, [r3],#4		@ r7 = *r++
-	LDR	r6, [r10,#-4]!		@ r6 = *--wR
-
-	@ Can save a cycle here, at the cost of 1bit errors in rounding
-	SMULL	r14,r11,r12,r11		@ (r14,r11)  = *l   * *wL++
-	SMULL	r14,r6, r7, r6		@ (r14,r6)   = *--r * *--wR
+unroll_loop6
+	LDR	r12,[r2],#8		; r12= *l       (but l += 2 first)
+	LDR	r11,[r9],#4		; r11= *wL++
+	LDR	r7, [r3],#4		; r7 = *r++
+	LDR	r6, [r10,#-4]!		; r6 = *--wR
+
+	; Can save a cycle here, at the cost of 1bit errors in rounding
+	SMULL	r14,r11,r12,r11		; (r14,r11)  = *l   * *wL++
+	SMULL	r14,r6, r7, r6		; (r14,r6)   = *--r * *--wR
 	SUB	r6, r6, r11
 	MOV	r6, r6, ASR #8
 	MOV	r14,r6, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r6, r4, r14,ASR #31
 	STRH	r6, [r0], r8
 
 	SUBS	r1, r1, #4
 	BGT	unroll_loop6
 
-unroll_over6:
+unroll_over6
 	LDMFD	r13!,{r4,r6-r11,PC}
 
-mdct_shift_right_arm:
-	@ r0 = n
-	@ r1 = in
-	@ r2 = right
+mdct_shift_right_arm
+	; r0 = n
+	; r1 = in
+	; r2 = right
 	STMFD	r13!,{r4-r11,r14}
 
-	MOV	r0, r0, LSR #2		@ n >>= 2
+	MOV	r0, r0, LSR #2		; n >>= 2
 	ADD	r1, r1, #4
 
 	SUBS	r0, r0,	#8
 	BLT	sr_less_than_8
-sr_loop:
+sr_loop
 	LDR	r3, [r1], #8
 	LDR	r4, [r1], #8
 	LDR	r5, [r1], #8
@@ -244,135 +244,135 @@ sr_loop:
 	SUBS	r0, r0, #8
 	STMIA	r2!,{r3,r4,r5,r6,r7,r8,r12,r14}
 	BGE	sr_loop
-sr_less_than_8:
+sr_less_than_8
 	ADDS	r0, r0, #8
 	BEQ	sr_end
-sr_loop2:
+sr_loop2
 	LDR	r3, [r1], #8
 	SUBS	r0, r0, #1
 	STR	r3, [r2], #4
 	BGT	sr_loop2
-sr_end:
+sr_end
 	LDMFD	r13!,{r4-r11,PC}
 
-mdct_backward_arm:
-	@ r0 = n
-	@ r1 = in
+mdct_backward_arm
+	; r0 = n
+	; r1 = in
 	STMFD	r13!,{r4-r11,r14}
 
-	MOV	r2,#1<<4	@ r2 = 1<<shift
-	MOV	r3,#13-4	@ r3 = 13-shift
-find_shift_loop:
-	TST	r0,r2		@ if (n & (1<<shift)) == 0
+	MOV	r2,#1<<4	; r2 = 1<<shift
+	MOV	r3,#13-4	; r3 = 13-shift
+find_shift_loop
+	TST	r0,r2		; if (n & (1<<shift)) == 0
 	MOV	r2,r2,LSL #1
-	SUBEQ	r3,r3,#1	@ shift--
+	SUBEQ	r3,r3,#1	; shift--
 	BEQ	find_shift_loop
 	MOV	r2,#2
-	MOV	r2,r2,LSL r3	@ r2 = step = 2<<shift
-
-	@ presymmetry
-	@ r0 = n (a multiple of 4)
-	@ r1 = in
-	@ r2 = step
-	@ r3 = shift
-
-	ADD	r4, r1, r0, LSL #1	@ r4 = aX = in+(n>>1)
-	ADD	r14,r1, r0		@ r14= in+(n>>2)
-	SUB	r4, r4, #3*4		@ r4 = aX = in+n2-3
-	LDR	r5, =sincos_lookup0	@ r5 = T=sincos_lookup0
-
-presymmetry_loop1:
-	LDR	r7, [r4,#8]		@ r6 = s2 = aX[2]
-	LDR	r11,[r5,#4]		@ r11= T[1]
-	LDR	r6, [r4]		@ r6 = s0 = aX[0]
-	LDR	r10,[r5],r2,LSL #2	@ r10= T[0]   T += step
-
-	@ XPROD31(s0, s2, T[0], T[1], 0xaX[0], &ax[2])
-	SMULL	r8, r9, r7, r11		@ (r8, r9)   = s2*T[1]
-	@ stall
-	@ stall ?
-	SMLAL	r8, r9, r6, r10		@ (r8, r9)  += s0*T[0]
+	MOV	r2,r2,LSL r3	; r2 = step = 2<<shift
+
+	; presymmetry
+	; r0 = n (a multiple of 4)
+	; r1 = in
+	; r2 = step
+	; r3 = shift
+
+	ADD	r4, r1, r0, LSL #1	; r4 = aX = in+(n>>1)
+	ADD	r14,r1, r0		; r14= in+(n>>2)
+	SUB	r4, r4, #3*4		; r4 = aX = in+n2-3
+	LDR	r5, =sincos_lookup0	; r5 = T=sincos_lookup0
+
+presymmetry_loop1
+	LDR	r7, [r4,#8]		; r6 = s2 = aX[2]
+	LDR	r11,[r5,#4]		; r11= T[1]
+	LDR	r6, [r4]		; r6 = s0 = aX[0]
+	LDR	r10,[r5],r2,LSL #2	; r10= T[0]   T += step
+
+	; XPROD31(s0, s2, T[0], T[1], &aX[0], &ax[2])
+	SMULL	r8, r9, r7, r11		; (r8, r9)   = s2*T[1]
+	; stall
+	; stall ?
+	SMLAL	r8, r9, r6, r10		; (r8, r9)  += s0*T[0]
 	RSB	r6, r6, #0
-	@ stall ?
-	SMULL	r8, r12,r7, r10		@ (r8, r12)  = s2*T[0]
+	; stall ?
+	SMULL	r8, r12,r7, r10		; (r8, r12)  = s2*T[0]
 	MOV	r9, r9, LSL #1
-	@ stall ?
-	SMLAL	r8, r12,r6, r11		@ (r8, r12) -= s0*T[1]
-	STR	r9, [r4],#-16		@ aX[0] = r9
+	; stall ?
+	SMLAL	r8, r12,r6, r11		; (r8, r12) -= s0*T[1]
+	STR	r9, [r4],#-16		; aX[0] = r9
 	CMP	r4,r14
 	MOV	r12,r12,LSL #1
-	STR	r12,[r4,#8+16]		@ aX[2] = r12
+	STR	r12,[r4,#8+16]		; aX[2] = r12
 
-	BGE	presymmetry_loop1	@ while (aX >= in+n4)
+	BGE	presymmetry_loop1	; while (aX >= in+n4)
 
-presymmetry_loop2:
-	LDR	r6,[r4]			@ r6 = s0 = aX[0]
-	LDR	r10,[r5,#4]		@ r10= T[1]
-	LDR	r7,[r4,#8]		@ r6 = s2 = aX[2]
-	LDR	r11,[r5],-r2,LSL #2	@ r11= T[0]   T -= step
+presymmetry_loop2
+	LDR	r6,[r4]			; r6 = s0 = aX[0]
+	LDR	r10,[r5,#4]		; r10= T[1]
+	LDR	r7,[r4,#8]		; r6 = s2 = aX[2]
+	LDR	r11,[r5],-r2,LSL #2	; r11= T[0]   T -= step
 
-	@ XPROD31(s0, s2, T[1], T[0], 0xaX[0], &ax[2])
-	SMULL	r8, r9, r6, r10		@ (r8, r9)   = s0*T[1]
-	@ stall
-	@ stall ?
-	SMLAL	r8, r9, r7, r11		@ (r8, r9)  += s2*T[0]
+	; XPROD31(s0, s2, T[1], T[0], &aX[0], &ax[2])
+	SMULL	r8, r9, r6, r10		; (r8, r9)   = s0*T[1]
+	; stall
+	; stall ?
+	SMLAL	r8, r9, r7, r11		; (r8, r9)  += s2*T[0]
 	RSB	r6, r6, #0
-	@ stall ?
-	SMULL	r8, r12,r7, r10		@ (r8, r12)  = s2*T[1]
+	; stall ?
+	SMULL	r8, r12,r7, r10		; (r8, r12)  = s2*T[1]
 	MOV	r9, r9, LSL #1
-	@ stall ?
-	SMLAL	r8, r12,r6, r11		@ (r8, r12) -= s0*T[0]
-	STR	r9, [r4],#-16		@ aX[0] = r9
+	; stall ?
+	SMLAL	r8, r12,r6, r11		; (r8, r12) -= s0*T[0]
+	STR	r9, [r4],#-16		; aX[0] = r9
 	CMP	r4,r1
 	MOV	r12,r12,LSL #1
-	STR	r12,[r4,#8+16]		@ aX[2] = r12
+	STR	r12,[r4,#8+16]		; aX[2] = r12
 
-	BGE	presymmetry_loop2	@ while (aX >= in)
+	BGE	presymmetry_loop2	; while (aX >= in)
 
-	@ r0 = n
-	@ r1 = in
-	@ r2 = step
-	@ r3 = shift
+	; r0 = n
+	; r1 = in
+	; r2 = step
+	; r3 = shift
 	STMFD	r13!,{r3}
-	LDR	r5, =sincos_lookup0	@ r5 = T=sincos_lookup0
-	ADD	r4, r1, r0, LSL #1	@ r4 = aX = in+(n>>1)
-	SUB	r4, r4, #4*4		@ r4 = aX = in+(n>>1)-4
-	LDR	r11,[r5,#4]		@ r11= T[1]
-	LDR	r10,[r5],r2, LSL #2	@ r10= T[0]    T += step
-presymmetry_loop3:
-	LDR	r8,[r1],#16 		@ r8 = ro0 = bX[0]
-	LDR	r9,[r1,#8-16]		@ r9 = ro2 = bX[2]
-	LDR	r6,[r4]			@ r6 = ri0 = aX[0]
-
-	@ XNPROD31( ro2, ro0, T[1], T[0], 0xaX[0], &aX[2] )
-	@ aX[0] = (ro2*T[1] - ro0*T[0])>>31 aX[2] = (ro0*T[1] + ro2*T[0])>>31
-	SMULL	r14,r12,r8, r11		@ (r14,r12)  = ro0*T[1]
-	RSB	r8,r8,#0		@ r8 = -ro0
-	@ Stall ?
-	SMLAL	r14,r12,r9, r10		@ (r14,r12) += ro2*T[0]
-	LDR	r7,[r4,#8]		@ r7 = ri2 = aX[2]
-	@ Stall ?
-	SMULL	r14,r3, r9, r11		@ (r14,r3)   = ro2*T[1]
+	LDR	r5, =sincos_lookup0	; r5 = T=sincos_lookup0
+	ADD	r4, r1, r0, LSL #1	; r4 = aX = in+(n>>1)
+	SUB	r4, r4, #4*4		; r4 = aX = in+(n>>1)-4
+	LDR	r11,[r5,#4]		; r11= T[1]
+	LDR	r10,[r5],r2, LSL #2	; r10= T[0]    T += step
+presymmetry_loop3
+	LDR	r8,[r1],#16 		; r8 = ro0 = bX[0]
+	LDR	r9,[r1,#8-16]		; r9 = ro2 = bX[2]
+	LDR	r6,[r4]			; r6 = ri0 = aX[0]
+
+	; XNPROD31( ro2, ro0, T[1], T[0], &aX[0], &aX[2] )
+	; aX[0] = (ro2*T[1] - ro0*T[0])>>31 aX[2] = (ro0*T[1] + ro2*T[0])>>31
+	SMULL	r14,r12,r8, r11		; (r14,r12)  = ro0*T[1]
+	RSB	r8,r8,#0		; r8 = -ro0
+	; Stall ?
+	SMLAL	r14,r12,r9, r10		; (r14,r12) += ro2*T[0]
+	LDR	r7,[r4,#8]		; r7 = ri2 = aX[2]
+	; Stall ?
+	SMULL	r14,r3, r9, r11		; (r14,r3)   = ro2*T[1]
 	MOV	r12,r12,LSL #1
-	LDR	r11,[r5,#4]		@ r11= T[1]
-	SMLAL	r14,r3, r8, r10		@ (r14,r3)  -= ro0*T[0]
-	LDR	r10,[r5],r2, LSL #2	@ r10= T[0]    T += step
+	LDR	r11,[r5,#4]		; r11= T[1]
+	SMLAL	r14,r3, r8, r10		; (r14,r3)  -= ro0*T[0]
+	LDR	r10,[r5],r2, LSL #2	; r10= T[0]    T += step
 	STR	r12,[r4,#8]
 	MOV	r3, r3, LSL #1
 	STR	r3, [r4],#-16
 
-	@ XNPROD31( ri2, ri0, T[0], T[1], 0xbX[0], &bX[2] )
-	@ bX[0] = (ri2*T[0] - ri0*T[1])>>31 bX[2] = (ri0*T[0] + ri2*T[1])>>31
-	SMULL	r14,r12,r6, r10		@ (r14,r12)  = ri0*T[0]
-	RSB	r6,r6,#0		@ r6 = -ri0
-	@ stall ?
-	SMLAL	r14,r12,r7, r11		@ (r14,r12) += ri2*T[1]
-	@ stall ?
-	@ stall ?
-	SMULL	r14,r3, r7, r10		@ (r14,r3)   = ri2*T[0]
+	; XNPROD31( ri2, ri0, T[0], T[1], &bX[0], &bX[2] )
+	; bX[0] = (ri2*T[0] - ri0*T[1])>>31 bX[2] = (ri0*T[0] + ri2*T[1])>>31
+	SMULL	r14,r12,r6, r10		; (r14,r12)  = ri0*T[0]
+	RSB	r6,r6,#0		; r6 = -ri0
+	; stall ?
+	SMLAL	r14,r12,r7, r11		; (r14,r12) += ri2*T[1]
+	; stall ?
+	; stall ?
+	SMULL	r14,r3, r7, r10		; (r14,r3)   = ri2*T[0]
 	MOV	r12,r12,LSL #1
-	@ stall ?
-	SMLAL	r14,r3, r6, r11		@ (r14,r3)  -= ri0*T[1]
+	; stall ?
+	SMLAL	r14,r3, r6, r11		; (r14,r3)  -= ri0*T[1]
 	CMP	r4,r1
 	STR	r12,[r1,#8-16]
 	MOV	r3, r3, LSL #1
@@ -380,571 +380,571 @@ presymmetry_loop3:
 
 	BGE	presymmetry_loop3
 
-	SUB	r1,r1,r0		@ r1 = in -= n>>2 (i.e. restore in)
+	SUB	r1,r1,r0		; r1 = in -= n>>2 (i.e. restore in)
 
 	LDR	r3,[r13]
 	STR	r2,[r13,#-4]!
 
-	@ mdct_butterflies
-	@ r0 = n  = (points * 2)
-	@ r1 = in = x
-	@ r2 = i
-	@ r3 = shift
+	; mdct_butterflies
+	; r0 = n  = (points * 2)
+	; r1 = in = x
+	; r2 = i
+	; r3 = shift
 	STMFD	r13!,{r0-r1}
-	RSBS	r4,r3,#6		@ r4 = stages = 7-shift then --stages
+	RSBS	r4,r3,#6		; r4 = stages = 7-shift then --stages
 	LDR	r5,=sincos_lookup0
 	BLE	no_generics
-	MOV	r14,#4			@ r14= 4               (i=0)
-	MOV	r6, r14,LSL r3		@ r6 = (4<<i)<<shift
-mdct_butterflies_loop1:
-	MOV	r0, r0, LSR #1		@ r0 = points>>i = POINTS
-	MOV	r2, r14,LSR #2		@ r2 = (1<<i)-j        (j=0)
+	MOV	r14,#4			; r14= 4               (i=0)
+	MOV	r6, r14,LSL r3		; r6 = (4<<i)<<shift
+mdct_butterflies_loop1
+	MOV	r0, r0, LSR #1		; r0 = points>>i = POINTS
+	MOV	r2, r14,LSR #2		; r2 = (1<<i)-j        (j=0)
 	STMFD	r13!,{r4,r14}
-mdct_butterflies_loop2:
-
-	@ mdct_butterfly_generic(x+POINTS*j, POINTS, 4<<(i+shift))
-	@ mdct_butterfly_generic(r1, r0, r6)
-	@ r0 = points
-	@ r1 = x
-	@ preserve r2 (external loop counter)
-	@ preserve r3
-	@ preserve r4 (external loop counter)
-	@ r5 = T = sincos_lookup0
-	@ r6 = step
-	@ preserve r14
-
-	STR	r2,[r13,#-4]!		@ stack r2
-	ADD	r1,r1,r0,LSL #1		@ r1 = x2+4 = x + (POINTS>>1)
-	ADD	r7,r1,r0,LSL #1		@ r7 = x1+4 = x + POINTS
-	ADD	r12,r5,#1024*4		@ r12= sincos_lookup0+1024
-
-mdct_bufferfly_generic_loop1:
-	LDMDB	r7!,{r2,r3,r8,r11}	@ r2 = x1[0]
-					@ r3 = x1[1]
-					@ r8 = x1[2]
-					@ r11= x1[3]    x1 -= 4
-	LDMDB	r1!,{r4,r9,r10,r14}	@ r4 = x2[0]
-					@ r9 = x2[1]
-					@ r10= x2[2]
-					@ r14= x2[3]    x2 -= 4
-
-	SUB	r2, r2, r3		@ r2 = s0 = x1[0] - x1[1]
-	ADD	r3, r2, r3, LSL #1	@ r3 =      x1[0] + x1[1] (-> x1[0])
-	SUB	r11,r11,r8		@ r11= s1 = x1[3] - x1[2]
-	ADD	r8, r11,r8, LSL #1	@ r8 =      x1[3] + x1[2] (-> x1[2])
-	SUB	r9, r9, r4		@ r9 = s2 = x2[1] - x2[0]
-	ADD	r4, r9, r4, LSL #1	@ r4 =      x2[1] + x2[0] (-> x1[1])
-	SUB	r14,r14,r10		@ r14= s3 = x2[3] - x2[2]
-	ADD	r10,r14,r10,LSL #1	@ r10=      x2[3] + x2[2] (-> x1[3])
+mdct_butterflies_loop2
+
+	; mdct_butterfly_generic(x+POINTS*j, POINTS, 4<<(i+shift))
+	; mdct_butterfly_generic(r1, r0, r6)
+	; r0 = points
+	; r1 = x
+	; preserve r2 (external loop counter)
+	; preserve r3
+	; preserve r4 (external loop counter)
+	; r5 = T = sincos_lookup0
+	; r6 = step
+	; preserve r14
+
+	STR	r2,[r13,#-4]!		; stack r2
+	ADD	r1,r1,r0,LSL #1		; r1 = x2+4 = x + (POINTS>>1)
+	ADD	r7,r1,r0,LSL #1		; r7 = x1+4 = x + POINTS
+	ADD	r12,r5,#1024*4		; r12= sincos_lookup0+1024
+
+mdct_bufferfly_generic_loop1
+	LDMDB	r7!,{r2,r3,r8,r11}	; r2 = x1[0]
+					; r3 = x1[1]
+					; r8 = x1[2]
+					; r11= x1[3]    x1 -= 4
+	LDMDB	r1!,{r4,r9,r10,r14}	; r4 = x2[0]
+					; r9 = x2[1]
+					; r10= x2[2]
+					; r14= x2[3]    x2 -= 4
+
+	SUB	r2, r2, r3		; r2 = s0 = x1[0] - x1[1]
+	ADD	r3, r2, r3, LSL #1	; r3 =      x1[0] + x1[1] (-> x1[0])
+	SUB	r11,r11,r8		; r11= s1 = x1[3] - x1[2]
+	ADD	r8, r11,r8, LSL #1	; r8 =      x1[3] + x1[2] (-> x1[2])
+	SUB	r9, r9, r4		; r9 = s2 = x2[1] - x2[0]
+	ADD	r4, r9, r4, LSL #1	; r4 =      x2[1] + x2[0] (-> x1[1])
+	SUB	r14,r14,r10		; r14= s3 = x2[3] - x2[2]
+	ADD	r10,r14,r10,LSL #1	; r10=      x2[3] + x2[2] (-> x1[3])
 	STMIA	r7,{r3,r4,r8,r10}
 
-	@ r0 = points
-	@ r1 = x2
-	@ r2 = s0
-	@ r3 free
-	@ r4 free
-	@ r5 = T
-	@ r6 = step
-	@ r7 = x1
-	@ r8 free
-	@ r9 = s2
-	@ r10 free
-	@ r11= s1
-	@ r12= limit
-	@ r14= s3
-
-	LDR	r8, [r5,#4]		@ r8 = T[1]
-	LDR	r10,[r5],r6,LSL #2	@ r10= T[0]		T += step
-
-	@ XPROD31(s1, s0, T[0], T[1], &x2[0], &x2[2])
-	@ x2[0] = (s1*T[0] + s0*T[1])>>31     x2[2] = (s0*T[0] - s1*T[1])>>31
-	@ stall Xscale
-	SMULL	r4, r3, r2, r8		@ (r4, r3)   = s0*T[1]
-	SMLAL	r4, r3, r11,r10		@ (r4, r3)  += s1*T[0]
+	; r0 = points
+	; r1 = x2
+	; r2 = s0
+	; r3 free
+	; r4 free
+	; r5 = T
+	; r6 = step
+	; r7 = x1
+	; r8 free
+	; r9 = s2
+	; r10 free
+	; r11= s1
+	; r12= limit
+	; r14= s3
+
+	LDR	r8, [r5,#4]		; r8 = T[1]
+	LDR	r10,[r5],r6,LSL #2	; r10= T[0]		T += step
+
+	; XPROD31(s1, s0, T[0], T[1], &x2[0], &x2[2])
+	; x2[0] = (s1*T[0] + s0*T[1])>>31     x2[2] = (s0*T[0] - s1*T[1])>>31
+	; stall Xscale
+	SMULL	r4, r3, r2, r8		; (r4, r3)   = s0*T[1]
+	SMLAL	r4, r3, r11,r10		; (r4, r3)  += s1*T[0]
 	RSB	r11,r11,#0
-	SMULL	r11,r4, r8, r11		@ (r11,r4)   = -s1*T[1]
-	SMLAL	r11,r4, r2, r10		@ (r11,r4)  += s0*T[0]
-	MOV	r2, r3, LSL #1		@ r2 = r3<<1 = Value for x2[0]
-
-	@ XPROD31(s2, s3, T[0], T[1], &x2[1], &x2[3])
-	@ x2[1] = (s2*T[0] + s3*T[1])>>31     x2[3] = (s3*T[0] - s2*T[1])>>31
-	SMULL	r11,r3, r9, r10		@ (r11,r3)   = s2*T[0]
-	MOV	r4, r4, LSL #1		@ r4 = r4<<1 = Value for x2[2]
-	SMLAL	r11,r3, r14,r8		@ (r11,r3)  += s3*T[1]
+	SMULL	r11,r4, r8, r11		; (r11,r4)   = -s1*T[1]
+	SMLAL	r11,r4, r2, r10		; (r11,r4)  += s0*T[0]
+	MOV	r2, r3, LSL #1		; r2 = r3<<1 = Value for x2[0]
+
+	; XPROD31(s2, s3, T[0], T[1], &x2[1], &x2[3])
+	; x2[1] = (s2*T[0] + s3*T[1])>>31     x2[3] = (s3*T[0] - s2*T[1])>>31
+	SMULL	r11,r3, r9, r10		; (r11,r3)   = s2*T[0]
+	MOV	r4, r4, LSL #1		; r4 = r4<<1 = Value for x2[2]
+	SMLAL	r11,r3, r14,r8		; (r11,r3)  += s3*T[1]
 	RSB	r9, r9, #0
-	SMULL	r10,r11,r14,r10		@ (r10,r11)  = s3*T[0]
-	MOV	r3, r3, LSL #1		@ r3 = r3<<1 = Value for x2[1]
-	SMLAL	r10,r11,r9,r8		@ (r10,r11) -= s2*T[1]
+	SMULL	r10,r11,r14,r10		; (r10,r11)  = s3*T[0]
+	MOV	r3, r3, LSL #1		; r3 = r3<<1 = Value for x2[1]
+	SMLAL	r10,r11,r9,r8		; (r10,r11) -= s2*T[1]
 	CMP	r5, r12
-	MOV	r11,r11,LSL #1		@ r11= r11<<1 = Value for x2[3]
+	MOV	r11,r11,LSL #1		; r11= r11<<1 = Value for x2[3]
 
 	STMIA	r1,{r2,r3,r4,r11}
 
 	BLT	mdct_bufferfly_generic_loop1
 
 	SUB	r12,r12,#1024*4
-mdct_bufferfly_generic_loop2:
-	LDMDB	r7!,{r2,r3,r9,r10}	@ r2 = x1[0]
-					@ r3 = x1[1]
-					@ r9 = x1[2]
-					@ r10= x1[3]    x1 -= 4
-	LDMDB	r1!,{r4,r8,r11,r14}	@ r4 = x2[0]
-					@ r8 = x2[1]
-					@ r11= x2[2]
-					@ r14= x2[3]    x2 -= 4
-
-	SUB	r2, r2, r3		@ r2 = s0 = x1[0] - x1[1]
-	ADD	r3, r2, r3, LSL #1	@ r3 =      x1[0] + x1[1] (-> x1[0])
-	SUB	r9, r9,r10		@ r9 = s1 = x1[2] - x1[3]
-	ADD	r10,r9,r10, LSL #1	@ r10=      x1[2] + x1[3] (-> x1[2])
-	SUB	r4, r4, r8		@ r4 = s2 = x2[0] - x2[1]
-	ADD	r8, r4, r8, LSL #1	@ r8 =      x2[0] + x2[1] (-> x1[1])
-	SUB	r14,r14,r11		@ r14= s3 = x2[3] - x2[2]
-	ADD	r11,r14,r11,LSL #1	@ r11=      x2[3] + x2[2] (-> x1[3])
+mdct_bufferfly_generic_loop2
+	LDMDB	r7!,{r2,r3,r9,r10}	; r2 = x1[0]
+					; r3 = x1[1]
+					; r9 = x1[2]
+					; r10= x1[3]    x1 -= 4
+	LDMDB	r1!,{r4,r8,r11,r14}	; r4 = x2[0]
+					; r8 = x2[1]
+					; r11= x2[2]
+					; r14= x2[3]    x2 -= 4
+
+	SUB	r2, r2, r3		; r2 = s0 = x1[0] - x1[1]
+	ADD	r3, r2, r3, LSL #1	; r3 =      x1[0] + x1[1] (-> x1[0])
+	SUB	r9, r9,r10		; r9 = s1 = x1[2] - x1[3]
+	ADD	r10,r9,r10, LSL #1	; r10=      x1[2] + x1[3] (-> x1[2])
+	SUB	r4, r4, r8		; r4 = s2 = x2[0] - x2[1]
+	ADD	r8, r4, r8, LSL #1	; r8 =      x2[0] + x2[1] (-> x1[1])
+	SUB	r14,r14,r11		; r14= s3 = x2[3] - x2[2]
+	ADD	r11,r14,r11,LSL #1	; r11=      x2[3] + x2[2] (-> x1[3])
 	STMIA	r7,{r3,r8,r10,r11}
 
-	@ r0 = points
-	@ r1 = x2
-	@ r2 = s0
-	@ r3 free
-	@ r4 = s2
-	@ r5 = T
-	@ r6 = step
-	@ r7 = x1
-	@ r8 free
-	@ r9 = s1
-	@ r10 free
-	@ r11 free
-	@ r12= limit
-	@ r14= s3
-
-	LDR	r8, [r5,#4]		@ r8 = T[1]
-	LDR	r10,[r5],-r6,LSL #2	@ r10= T[0]		T -= step
-
-	@ XNPROD31(s0, s1, T[0], T[1], &x2[0], &x2[2])
-	@ x2[0] = (s0*T[0] - s1*T[1])>>31     x2[2] = (s1*T[0] + s0*T[1])>>31
-	@ stall Xscale
-	SMULL	r3, r11,r2, r8		@ (r3, r11)  = s0*T[1]
-	SMLAL	r3, r11,r9, r10		@ (r3, r11) += s1*T[0]
+	; r0 = points
+	; r1 = x2
+	; r2 = s0
+	; r3 free
+	; r4 = s2
+	; r5 = T
+	; r6 = step
+	; r7 = x1
+	; r8 free
+	; r9 = s1
+	; r10 free
+	; r11 free
+	; r12= limit
+	; r14= s3
+
+	LDR	r8, [r5,#4]		; r8 = T[1]
+	LDR	r10,[r5],-r6,LSL #2	; r10= T[0]		T -= step
+
+	; XNPROD31(s0, s1, T[0], T[1], &x2[0], &x2[2])
+	; x2[0] = (s0*T[0] - s1*T[1])>>31     x2[2] = (s1*T[0] + s0*T[1])>>31
+	; stall Xscale
+	SMULL	r3, r11,r2, r8		; (r3, r11)  = s0*T[1]
+	SMLAL	r3, r11,r9, r10		; (r3, r11) += s1*T[0]
 	RSB	r9, r9, #0
-	SMULL	r3, r2, r10,r2		@ (r3, r2)   = s0*T[0]
-	SMLAL	r3, r2, r9, r8		@ (r3, r2)  += -s1*T[1]
-	MOV	r9, r11,LSL #1		@ r9 = r11<<1 = Value for x2[2]
-
-	@ XNPROD31(s3, s2, T[0], T[1], &x2[1], &x2[3])
-	@ x2[1] = (s3*T[0] - s2*T[1])>>31     x2[3] = (s2*T[0] + s3*T[1])>>31
-	SMULL	r3, r11,r4, r10		@ (r3,r11)   = s2*T[0]
-	MOV	r2, r2, LSL #1		@ r2 = r2<<1  = Value for x2[0]
-	SMLAL	r3, r11,r14,r8		@ (r3,r11)  += s3*T[1]
+	SMULL	r3, r2, r10,r2		; (r3, r2)   = s0*T[0]
+	SMLAL	r3, r2, r9, r8		; (r3, r2)  += -s1*T[1]
+	MOV	r9, r11,LSL #1		; r9 = r11<<1 = Value for x2[2]
+
+	; XNPROD31(s3, s2, T[0], T[1], &x2[1], &x2[3])
+	; x2[1] = (s3*T[0] - s2*T[1])>>31     x2[3] = (s2*T[0] + s3*T[1])>>31
+	SMULL	r3, r11,r4, r10		; (r3,r11)   = s2*T[0]
+	MOV	r2, r2, LSL #1		; r2 = r2<<1  = Value for x2[0]
+	SMLAL	r3, r11,r14,r8		; (r3,r11)  += s3*T[1]
 	RSB	r4, r4, #0
-	SMULL	r10,r3,r14,r10		@ (r10,r3)   = s3*T[0]
-	MOV	r11,r11,LSL #1		@ r11= r11<<1 = Value for x2[3]
-	SMLAL	r10,r3, r4, r8		@ (r10,r3)  -= s2*T[1]
+	SMULL	r10,r3,r14,r10		; (r10,r3)   = s3*T[0]
+	MOV	r11,r11,LSL #1		; r11= r11<<1 = Value for x2[3]
+	SMLAL	r10,r3, r4, r8		; (r10,r3)  -= s2*T[1]
 	CMP	r5, r12
-	MOV	r3, r3, LSL #1		@ r3 = r3<<1  = Value for x2[1]
+	MOV	r3, r3, LSL #1		; r3 = r3<<1  = Value for x2[1]
 
 	STMIA	r1,{r2,r3,r9,r11}
 
 	BGT	mdct_bufferfly_generic_loop2
 
-	LDR	r2,[r13],#4		@ unstack r2
-	ADD	r1, r1, r0, LSL #2	@ r1 = x+POINTS*j
-	@ stall Xscale
-	SUBS	r2, r2, #1		@ r2--                 (j++)
+	LDR	r2,[r13],#4		; unstack r2
+	ADD	r1, r1, r0, LSL #2	; r1 = x+POINTS*j
+	; stall Xscale
+	SUBS	r2, r2, #1		; r2--                 (j++)
 	BGT	mdct_butterflies_loop2
 
 	LDMFD	r13!,{r4,r14}
 
 	LDR	r1,[r13,#4]
 
-	SUBS	r4, r4, #1		@ stages--
-	MOV	r14,r14,LSL #1		@ r14= 4<<i            (i++)
-	MOV	r6, r6, LSL #1		@ r6 = step <<= 1      (i++)
+	SUBS	r4, r4, #1		; stages--
+	MOV	r14,r14,LSL #1		; r14= 4<<i            (i++)
+	MOV	r6, r6, LSL #1		; r6 = step <<= 1      (i++)
 	BGE	mdct_butterflies_loop1
 	LDMFD	r13,{r0-r1}
-no_generics:
-	@ mdct_butterflies part2 (loop around mdct_bufferfly_32)
-	@ r0 = points
-	@ r1 = in
-	@ r2 = step
-	@ r3 = shift
-
-mdct_bufferflies_loop3:
-	@ mdct_bufferfly_32
-
-	@ block1
-	ADD	r4, r1, #16*4		@ r4 = &in[16]
-	LDMIA	r4,{r5,r6,r9,r10}	@ r5 = x[16]
-					@ r6 = x[17]
-					@ r9 = x[18]
-					@ r10= x[19]
-	LDMIA	r1,{r7,r8,r11,r12}	@ r7 = x[0]
-					@ r8 = x[1]
-					@ r11= x[2]
-					@ r12= x[3]
-	SUB	r5, r5, r6		@ r5 = s0 = x[16] - x[17]
-	ADD	r6, r5, r6, LSL #1	@ r6 =      x[16] + x[17]  -> x[16]
-	SUB	r9, r9, r10		@ r9 = s1 = x[18] - x[19]
-	ADD	r10,r9, r10,LSL #1	@ r10=      x[18] + x[19]  -> x[18]
-	SUB	r8, r8, r7		@ r8 = s2 = x[ 1] - x[ 0]
-	ADD	r7, r8, r7, LSL #1	@ r7 =      x[ 1] + x[ 0]  -> x[17]
-	SUB	r12,r12,r11		@ r12= s3 = x[ 3] - x[ 2]
-	ADD	r11,r12,r11, LSL #1	@ r11=      x[ 3] + x[ 2]  -> x[19]
+no_generics
+	; mdct_butterflies part2 (loop around mdct_bufferfly_32)
+	; r0 = points
+	; r1 = in
+	; r2 = step
+	; r3 = shift
+
+mdct_bufferflies_loop3
+	; mdct_bufferfly_32
+
+	; block1
+	ADD	r4, r1, #16*4		; r4 = &in[16]
+	LDMIA	r4,{r5,r6,r9,r10}	; r5 = x[16]
+					; r6 = x[17]
+					; r9 = x[18]
+					; r10= x[19]
+	LDMIA	r1,{r7,r8,r11,r12}	; r7 = x[0]
+					; r8 = x[1]
+					; r11= x[2]
+					; r12= x[3]
+	SUB	r5, r5, r6		; r5 = s0 = x[16] - x[17]
+	ADD	r6, r5, r6, LSL #1	; r6 =      x[16] + x[17]  -> x[16]
+	SUB	r9, r9, r10		; r9 = s1 = x[18] - x[19]
+	ADD	r10,r9, r10,LSL #1	; r10=      x[18] + x[19]  -> x[18]
+	SUB	r8, r8, r7		; r8 = s2 = x[ 1] - x[ 0]
+	ADD	r7, r8, r7, LSL #1	; r7 =      x[ 1] + x[ 0]  -> x[17]
+	SUB	r12,r12,r11		; r12= s3 = x[ 3] - x[ 2]
+	ADD	r11,r12,r11, LSL #1	; r11=      x[ 3] + x[ 2]  -> x[19]
 	STMIA	r4!,{r6,r7,r10,r11}
 
 	LDR	r6,cPI1_8
 	LDR	r7,cPI3_8
 
-	@ XNPROD31( s0, s1, cPI3_8, cPI1_8, &x[ 0], &x[ 2] )
-	@ x[0] = s0*cPI3_8 - s1*cPI1_8     x[2] = s1*cPI3_8 + s0*cPI1_8
-	@ stall Xscale
-	SMULL	r14,r11,r5, r6		@ (r14,r11)  = s0*cPI1_8
-	SMLAL	r14,r11,r9, r7		@ (r14,r11) += s1*cPI3_8
+	; XNPROD31( s0, s1, cPI3_8, cPI1_8, &x[ 0], &x[ 2] )
+	; x[0] = s0*cPI3_8 - s1*cPI1_8     x[2] = s1*cPI3_8 + s0*cPI1_8
+	; stall Xscale
+	SMULL	r14,r11,r5, r6		; (r14,r11)  = s0*cPI1_8
+	SMLAL	r14,r11,r9, r7		; (r14,r11) += s1*cPI3_8
 	RSB	r9, r9, #0
-	SMULL	r14,r5, r7, r5		@ (r14,r5)   = s0*cPI3_8
-	SMLAL	r14,r5, r9, r6		@ (r14,r5)  -= s1*cPI1_8
+	SMULL	r14,r5, r7, r5		; (r14,r5)   = s0*cPI3_8
+	SMLAL	r14,r5, r9, r6		; (r14,r5)  -= s1*cPI1_8
 	MOV	r11,r11,LSL #1
 	MOV	r5, r5, LSL #1
 
-	@ XPROD31 ( s2, s3, cPI1_8, cPI3_8, &x[ 1], &x[ 3] )
-	@ x[1] = s2*cPI1_8 + s3*cPI3_8     x[3] = s3*cPI1_8 - s2*cPI3_8
-	SMULL	r14,r9, r8, r6		@ (r14,r9)   = s2*cPI1_8
-	SMLAL	r14,r9, r12,r7		@ (r14,r9)  += s3*cPI3_8
+	; XPROD31 ( s2, s3, cPI1_8, cPI3_8, &x[ 1], &x[ 3] )
+	; x[1] = s2*cPI1_8 + s3*cPI3_8     x[3] = s3*cPI1_8 - s2*cPI3_8
+	SMULL	r14,r9, r8, r6		; (r14,r9)   = s2*cPI1_8
+	SMLAL	r14,r9, r12,r7		; (r14,r9)  += s3*cPI3_8
 	RSB	r8,r8,#0
-	SMULL	r14,r12,r6, r12		@ (r14,r12)  = s3*cPI1_8
-	SMLAL	r14,r12,r8, r7		@ (r14,r12) -= s2*cPI3_8
+	SMULL	r14,r12,r6, r12		; (r14,r12)  = s3*cPI1_8
+	SMLAL	r14,r12,r8, r7		; (r14,r12) -= s2*cPI3_8
 	MOV	r9, r9, LSL #1
 	MOV	r12,r12,LSL #1
 	STMIA	r1!,{r5,r9,r11,r12}
 
-	@ block2
-	LDMIA	r4,{r5,r6,r9,r10}	@ r5 = x[20]
-					@ r6 = x[21]
-					@ r9 = x[22]
-					@ r10= x[23]
-	LDMIA	r1,{r7,r8,r11,r12}	@ r7 = x[4]
-					@ r8 = x[5]
-					@ r11= x[6]
-					@ r12= x[7]
-	SUB	r5, r5, r6		@ r5 = s0 = x[20] - x[21]
-	ADD	r6, r5, r6, LSL #1	@ r6 =      x[20] + x[21]  -> x[20]
-	SUB	r9, r9, r10		@ r9 = s1 = x[22] - x[23]
-	ADD	r10,r9, r10,LSL #1	@ r10=      x[22] + x[23]  -> x[22]
-	SUB	r8, r8, r7		@ r8 = s2 = x[ 5] - x[ 4]
-	ADD	r7, r8, r7, LSL #1	@ r7 =      x[ 5] + x[ 4]  -> x[21]
-	SUB	r12,r12,r11		@ r12= s3 = x[ 7] - x[ 6]
-	ADD	r11,r12,r11, LSL #1	@ r11=      x[ 7] + x[ 6]  -> x[23]
+	; block2
+	LDMIA	r4,{r5,r6,r9,r10}	; r5 = x[20]
+					; r6 = x[21]
+					; r9 = x[22]
+					; r10= x[23]
+	LDMIA	r1,{r7,r8,r11,r12}	; r7 = x[4]
+					; r8 = x[5]
+					; r11= x[6]
+					; r12= x[7]
+	SUB	r5, r5, r6		; r5 = s0 = x[20] - x[21]
+	ADD	r6, r5, r6, LSL #1	; r6 =      x[20] + x[21]  -> x[20]
+	SUB	r9, r9, r10		; r9 = s1 = x[22] - x[23]
+	ADD	r10,r9, r10,LSL #1	; r10=      x[22] + x[23]  -> x[22]
+	SUB	r8, r8, r7		; r8 = s2 = x[ 5] - x[ 4]
+	ADD	r7, r8, r7, LSL #1	; r7 =      x[ 5] + x[ 4]  -> x[21]
+	SUB	r12,r12,r11		; r12= s3 = x[ 7] - x[ 6]
+	ADD	r11,r12,r11, LSL #1	; r11=      x[ 7] + x[ 6]  -> x[23]
 	LDR	r14,cPI2_8
 	STMIA	r4!,{r6,r7,r10,r11}
 
-	SUB	r5, r5, r9		@ r5 = s0 - s1
-	ADD	r9, r5, r9, LSL #1	@ r9 = s0 + s1
-	SMULL	r6, r5, r14,r5		@ (r6,r5)  = (s0-s1)*cPI2_8
-	SUB	r12,r12,r8		@ r12= s3 - s2
-	ADD	r8, r12,r8, LSL #1	@ r8 = s3 + s2
+	SUB	r5, r5, r9		; r5 = s0 - s1
+	ADD	r9, r5, r9, LSL #1	; r9 = s0 + s1
+	SMULL	r6, r5, r14,r5		; (r6,r5)  = (s0-s1)*cPI2_8
+	SUB	r12,r12,r8		; r12= s3 - s2
+	ADD	r8, r12,r8, LSL #1	; r8 = s3 + s2
 
-	SMULL	r6, r8, r14,r8		@ (r6,r8)  = (s3+s2)*cPI2_8
+	SMULL	r6, r8, r14,r8		; (r6,r8)  = (s3+s2)*cPI2_8
 	MOV	r5, r5, LSL #1
-	SMULL	r6, r9, r14,r9		@ (r6,r9)  = (s0+s1)*cPI2_8
+	SMULL	r6, r9, r14,r9		; (r6,r9)  = (s0+s1)*cPI2_8
 	MOV	r8, r8, LSL #1
-	SMULL	r6, r12,r14,r12		@ (r6,r12) = (s3-s2)*cPI2_8
+	SMULL	r6, r12,r14,r12		; (r6,r12) = (s3-s2)*cPI2_8
 	MOV	r9, r9, LSL #1
 	MOV	r12,r12,LSL #1
 	STMIA	r1!,{r5,r8,r9,r12}
 
-	@ block3
-	LDMIA	r4,{r5,r6,r9,r10}	@ r5 = x[24]
-					@ r6 = x[25]
-					@ r9 = x[25]
-					@ r10= x[26]
-	LDMIA	r1,{r7,r8,r11,r12}	@ r7 = x[8]
-					@ r8 = x[9]
-					@ r11= x[10]
-					@ r12= x[11]
-	SUB	r5, r5, r6		@ r5 = s0 = x[24] - x[25]
-	ADD	r6, r5, r6, LSL #1	@ r6 =      x[24] + x[25]  -> x[25]
-	SUB	r9, r9, r10		@ r9 = s1 = x[26] - x[27]
-	ADD	r10,r9, r10,LSL #1	@ r10=      x[26] + x[27]  -> x[26]
-	SUB	r8, r8, r7		@ r8 = s2 = x[ 9] - x[ 8]
-	ADD	r7, r8, r7, LSL #1	@ r7 =      x[ 9] + x[ 8]  -> x[25]
-	SUB	r12,r12,r11		@ r12= s3 = x[11] - x[10]
-	ADD	r11,r12,r11, LSL #1	@ r11=      x[11] + x[10]  -> x[27]
+	; block3
+	LDMIA	r4,{r5,r6,r9,r10}	; r5 = x[24]
+					; r6 = x[25]
+					; r9 = x[25]
+					; r10= x[26]
+	LDMIA	r1,{r7,r8,r11,r12}	; r7 = x[8]
+					; r8 = x[9]
+					; r11= x[10]
+					; r12= x[11]
+	SUB	r5, r5, r6		; r5 = s0 = x[24] - x[25]
+	ADD	r6, r5, r6, LSL #1	; r6 =      x[24] + x[25]  -> x[25]
+	SUB	r9, r9, r10		; r9 = s1 = x[26] - x[27]
+	ADD	r10,r9, r10,LSL #1	; r10=      x[26] + x[27]  -> x[26]
+	SUB	r8, r8, r7		; r8 = s2 = x[ 9] - x[ 8]
+	ADD	r7, r8, r7, LSL #1	; r7 =      x[ 9] + x[ 8]  -> x[25]
+	SUB	r12,r12,r11		; r12= s3 = x[11] - x[10]
+	ADD	r11,r12,r11, LSL #1	; r11=      x[11] + x[10]  -> x[27]
 	STMIA	r4!,{r6,r7,r10,r11}
 
 	LDR	r6,cPI3_8
 	LDR	r7,cPI1_8
 
-	@ XNPROD31( s0, s1, cPI1_8, cPI3_8, &x[ 8], &x[10] )
-	@ x[8] = s0*cPI1_8 - s1*cPI3_8     x[10] = s1*cPI1_8 + s0*cPI3_8
-	@ stall Xscale
-	SMULL	r14,r11,r5, r6		@ (r14,r11)  = s0*cPI3_8
-	SMLAL	r14,r11,r9, r7		@ (r14,r11) += s1*cPI1_8
+	; XNPROD31( s0, s1, cPI1_8, cPI3_8, &x[ 8], &x[10] )
+	; x[8] = s0*cPI1_8 - s1*cPI3_8     x[10] = s1*cPI1_8 + s0*cPI3_8
+	; stall Xscale
+	SMULL	r14,r11,r5, r6		; (r14,r11)  = s0*cPI3_8
+	SMLAL	r14,r11,r9, r7		; (r14,r11) += s1*cPI1_8
 	RSB	r9, r9, #0
-	SMULL	r14,r5, r7, r5		@ (r14,r5)   = s0*cPI1_8
-	SMLAL	r14,r5, r9, r6		@ (r14,r5)  -= s1*cPI3_8
+	SMULL	r14,r5, r7, r5		; (r14,r5)   = s0*cPI1_8
+	SMLAL	r14,r5, r9, r6		; (r14,r5)  -= s1*cPI3_8
 	MOV	r11,r11,LSL #1
 	MOV	r5, r5, LSL #1
 
-	@ XPROD31 ( s2, s3, cPI3_8, cPI1_8, &x[ 9], &x[11] )
-	@ x[9] = s2*cPI3_8 + s3*cPI1_8     x[11] = s3*cPI3_8 - s2*cPI1_8
-	SMULL	r14,r9, r8, r6		@ (r14,r9)   = s2*cPI3_8
-	SMLAL	r14,r9, r12,r7		@ (r14,r9)  += s3*cPI1_8
+	; XPROD31 ( s2, s3, cPI3_8, cPI1_8, &x[ 9], &x[11] )
+	; x[9] = s2*cPI3_8 + s3*cPI1_8     x[11] = s3*cPI3_8 - s2*cPI1_8
+	SMULL	r14,r9, r8, r6		; (r14,r9)   = s2*cPI3_8
+	SMLAL	r14,r9, r12,r7		; (r14,r9)  += s3*cPI1_8
 	RSB	r8,r8,#0
-	SMULL	r14,r12,r6, r12		@ (r14,r12)  = s3*cPI3_8
-	SMLAL	r14,r12,r8, r7		@ (r14,r12) -= s2*cPI1_8
+	SMULL	r14,r12,r6, r12		; (r14,r12)  = s3*cPI3_8
+	SMLAL	r14,r12,r8, r7		; (r14,r12) -= s2*cPI1_8
 	MOV	r9, r9, LSL #1
 	MOV	r12,r12,LSL #1
 	STMIA	r1!,{r5,r9,r11,r12}
 
-	@ block4
-	LDMIA	r4,{r5,r6,r10,r11}	@ r5 = x[28]
-					@ r6 = x[29]
-					@ r10= x[30]
-					@ r11= x[31]
-	LDMIA	r1,{r8,r9,r12,r14}	@ r8 = x[12]
-					@ r9 = x[13]
-					@ r12= x[14]
-					@ r14= x[15]
-	SUB	r5, r5, r6		@ r5 = s0 = x[28] - x[29]
-	ADD	r6, r5, r6, LSL #1	@ r6 =      x[28] + x[29]  -> x[28]
-	SUB	r7, r14,r12		@ r7 = s3 = x[15] - x[14]
-	ADD	r12,r7, r12, LSL #1	@ r12=      x[15] + x[14]  -> x[31]
-	SUB	r10,r10,r11		@ r10= s1 = x[30] - x[31]
-	ADD	r11,r10,r11,LSL #1	@ r11=      x[30] + x[31]  -> x[30]
-	SUB	r14, r8, r9		@ r14= s2 = x[12] - x[13]
-	ADD	r9, r14, r9, LSL #1	@ r9 =      x[12] + x[13]  -> x[29]
+	; block4
+	LDMIA	r4,{r5,r6,r10,r11}	; r5 = x[28]
+					; r6 = x[29]
+					; r10= x[30]
+					; r11= x[31]
+	LDMIA	r1,{r8,r9,r12,r14}	; r8 = x[12]
+					; r9 = x[13]
+					; r12= x[14]
+					; r14= x[15]
+	SUB	r5, r5, r6		; r5 = s0 = x[28] - x[29]
+	ADD	r6, r5, r6, LSL #1	; r6 =      x[28] + x[29]  -> x[28]
+	SUB	r7, r14,r12		; r7 = s3 = x[15] - x[14]
+	ADD	r12,r7, r12, LSL #1	; r12=      x[15] + x[14]  -> x[31]
+	SUB	r10,r10,r11		; r10= s1 = x[30] - x[31]
+	ADD	r11,r10,r11,LSL #1	; r11=      x[30] + x[31]  -> x[30]
+	SUB	r14, r8, r9		; r14= s2 = x[12] - x[13]
+	ADD	r9, r14, r9, LSL #1	; r9 =      x[12] + x[13]  -> x[29]
 	STMIA	r4!,{r6,r9,r11,r12}
 	STMIA	r1!,{r5,r7,r10,r14}
 
-	@ mdct_butterfly16 (1st version)
-	@ block 1
+	; mdct_butterfly16 (1st version)
+	; block 1
 	SUB	r1,r1,#16*4
 	ADD	r4,r1,#8*4
-	LDMIA	r4,{r5,r6,r9,r10}	@ r5 = x[ 8]
-					@ r6 = x[ 9]
-					@ r9 = x[10]
-					@ r10= x[11]
-	LDMIA	r1,{r7,r8,r11,r12}	@ r7 = x[0]
-					@ r8 = x[1]
-					@ r11= x[2]
-					@ r12= x[3]
-	SUB	r5, r5, r6		@ r5 = s0 = x[ 8] - x[ 9]
-	ADD	r6, r5, r6, LSL #1	@ r6 =      x[ 8] + x[ 9]  -> x[ 8]
-	SUB	r9, r9, r10		@ r9 = s1 = x[10] - x[11]
-	ADD	r10,r9, r10,LSL #1	@ r10=      x[10] + x[11]  -> x[10]
-	SUB	r8, r8, r7		@ r8 = s2 = x[ 1] - x[ 0]
-	ADD	r7, r8, r7, LSL #1	@ r7 =      x[ 1] + x[ 0]  -> x[ 9]
-	SUB	r12,r12,r11		@ r12= s3 = x[ 3] - x[ 2]
-	ADD	r11,r12,r11, LSL #1	@ r11=      x[ 3] + x[ 2]  -> x[11]
+	LDMIA	r4,{r5,r6,r9,r10}	; r5 = x[ 8]
+					; r6 = x[ 9]
+					; r9 = x[10]
+					; r10= x[11]
+	LDMIA	r1,{r7,r8,r11,r12}	; r7 = x[0]
+					; r8 = x[1]
+					; r11= x[2]
+					; r12= x[3]
+	SUB	r5, r5, r6		; r5 = s0 = x[ 8] - x[ 9]
+	ADD	r6, r5, r6, LSL #1	; r6 =      x[ 8] + x[ 9]  -> x[ 8]
+	SUB	r9, r9, r10		; r9 = s1 = x[10] - x[11]
+	ADD	r10,r9, r10,LSL #1	; r10=      x[10] + x[11]  -> x[10]
+	SUB	r8, r8, r7		; r8 = s2 = x[ 1] - x[ 0]
+	ADD	r7, r8, r7, LSL #1	; r7 =      x[ 1] + x[ 0]  -> x[ 9]
+	SUB	r12,r12,r11		; r12= s3 = x[ 3] - x[ 2]
+	ADD	r11,r12,r11, LSL #1	; r11=      x[ 3] + x[ 2]  -> x[11]
 	LDR	r14,cPI2_8
 	STMIA	r4!,{r6,r7,r10,r11}
 
-	SUB	r5, r5, r9		@ r5 = s0 - s1
-	ADD	r9, r5, r9, LSL #1	@ r9 = s0 + s1
-	SMULL	r6, r5, r14,r5		@ (r6,r5)  = (s0-s1)*cPI2_8
-	SUB	r12,r12,r8		@ r12= s3 - s2
-	ADD	r8, r12,r8, LSL #1	@ r8 = s3 + s2
+	SUB	r5, r5, r9		; r5 = s0 - s1
+	ADD	r9, r5, r9, LSL #1	; r9 = s0 + s1
+	SMULL	r6, r5, r14,r5		; (r6,r5)  = (s0-s1)*cPI2_8
+	SUB	r12,r12,r8		; r12= s3 - s2
+	ADD	r8, r12,r8, LSL #1	; r8 = s3 + s2
 
-	SMULL	r6, r8, r14,r8		@ (r6,r8)  = (s3+s2)*cPI2_8
+	SMULL	r6, r8, r14,r8		; (r6,r8)  = (s3+s2)*cPI2_8
 	MOV	r5, r5, LSL #1
-	SMULL	r6, r9, r14,r9		@ (r6,r9)  = (s0+s1)*cPI2_8
+	SMULL	r6, r9, r14,r9		; (r6,r9)  = (s0+s1)*cPI2_8
 	MOV	r8, r8, LSL #1
-	SMULL	r6, r12,r14,r12		@ (r6,r12) = (s3-s2)*cPI2_8
+	SMULL	r6, r12,r14,r12		; (r6,r12) = (s3-s2)*cPI2_8
 	MOV	r9, r9, LSL #1
 	MOV	r12,r12,LSL #1
 	STMIA	r1!,{r5,r8,r9,r12}
 
-	@ block4
-	LDMIA	r4,{r5,r6,r9,r10}	@ r5 = x[12]
-					@ r6 = x[13]
-					@ r9 = x[14]
-					@ r10= x[15]
-	LDMIA	r1,{r7,r8,r11,r12}	@ r7 = x[ 4]
-					@ r8 = x[ 5]
-					@ r11= x[ 6]
-					@ r12= x[ 7]
-	SUB	r14,r7, r8		@ r14= s0 = x[ 4] - x[ 5]
-	ADD	r8, r14,r8, LSL #1	@ r8 =      x[ 4] + x[ 5]  -> x[13]
-	SUB	r7, r12,r11		@ r7 = s1 = x[ 7] - x[ 6]
-	ADD	r11,r7, r11, LSL #1	@ r11=      x[ 7] + x[ 6]  -> x[15]
-	SUB	r5, r5, r6		@ r5 = s2 = x[12] - x[13]
-	ADD	r6, r5, r6, LSL #1	@ r6 =      x[12] + x[13]  -> x[12]
-	SUB	r12,r9, r10		@ r12= s3 = x[14] - x[15]
-	ADD	r10,r12,r10,LSL #1	@ r10=      x[14] + x[15]  -> x[14]
+	; block4
+	LDMIA	r4,{r5,r6,r9,r10}	; r5 = x[12]
+					; r6 = x[13]
+					; r9 = x[14]
+					; r10= x[15]
+	LDMIA	r1,{r7,r8,r11,r12}	; r7 = x[ 4]
+					; r8 = x[ 5]
+					; r11= x[ 6]
+					; r12= x[ 7]
+	SUB	r14,r7, r8		; r14= s0 = x[ 4] - x[ 5]
+	ADD	r8, r14,r8, LSL #1	; r8 =      x[ 4] + x[ 5]  -> x[13]
+	SUB	r7, r12,r11		; r7 = s1 = x[ 7] - x[ 6]
+	ADD	r11,r7, r11, LSL #1	; r11=      x[ 7] + x[ 6]  -> x[15]
+	SUB	r5, r5, r6		; r5 = s2 = x[12] - x[13]
+	ADD	r6, r5, r6, LSL #1	; r6 =      x[12] + x[13]  -> x[12]
+	SUB	r12,r9, r10		; r12= s3 = x[14] - x[15]
+	ADD	r10,r12,r10,LSL #1	; r10=      x[14] + x[15]  -> x[14]
 	STMIA	r4!,{r6,r8,r10,r11}
 	STMIA	r1!,{r5,r7,r12,r14}
 
-	@ mdct_butterfly_8
+	; mdct_butterfly_8
 	LDMDB	r1,{r6,r7,r8,r9,r10,r11,r12,r14}
-					@ r6 = x[0]
-					@ r7 = x[1]
-					@ r8 = x[2]
-					@ r9 = x[3]
-					@ r10= x[4]
-					@ r11= x[5]
-					@ r12= x[6]
-					@ r14= x[7]
-	ADD	r6, r6, r7		@ r6 = s0 = x[0] + x[1]
-	SUB	r7, r6, r7, LSL #1	@ r7 = s1 = x[0] - x[1]
-	ADD	r8, r8, r9		@ r8 = s2 = x[2] + x[3]
-	SUB	r9, r8, r9, LSL #1	@ r9 = s3 = x[2] - x[3]
-	ADD	r10,r10,r11		@ r10= s4 = x[4] + x[5]
-	SUB	r11,r10,r11,LSL #1	@ r11= s5 = x[4] - x[5]
-	ADD	r12,r12,r14		@ r12= s6 = x[6] + x[7]
-	SUB	r14,r12,r14,LSL #1	@ r14= s7 = x[6] - x[7]
-
-	ADD	r2, r11,r9		@ r2 = x[0] = s5 + s3
-	SUB	r4, r2, r9, LSL #1	@ r4 = x[2] = s5 - s3
-	SUB	r3, r14,r7		@ r3 = x[1] = s7 - s1
-	ADD	r5, r3, r7, LSL #1	@ r5 = x[3] = s7 + s1
-	SUB	r10,r10,r6		@ r10= x[4] = s4 - s0
-	SUB	r11,r12,r8		@ r11= x[5] = s6 - s2
-	ADD	r12,r10,r6, LSL #1	@ r12= x[6] = s4 + s0
-	ADD	r14,r11,r8, LSL #1	@ r14= x[7] = s6 + s2
+					; r6 = x[0]
+					; r7 = x[1]
+					; r8 = x[2]
+					; r9 = x[3]
+					; r10= x[4]
+					; r11= x[5]
+					; r12= x[6]
+					; r14= x[7]
+	ADD	r6, r6, r7		; r6 = s0 = x[0] + x[1]
+	SUB	r7, r6, r7, LSL #1	; r7 = s1 = x[0] - x[1]
+	ADD	r8, r8, r9		; r8 = s2 = x[2] + x[3]
+	SUB	r9, r8, r9, LSL #1	; r9 = s3 = x[2] - x[3]
+	ADD	r10,r10,r11		; r10= s4 = x[4] + x[5]
+	SUB	r11,r10,r11,LSL #1	; r11= s5 = x[4] - x[5]
+	ADD	r12,r12,r14		; r12= s6 = x[6] + x[7]
+	SUB	r14,r12,r14,LSL #1	; r14= s7 = x[6] - x[7]
+
+	ADD	r2, r11,r9		; r2 = x[0] = s5 + s3
+	SUB	r4, r2, r9, LSL #1	; r4 = x[2] = s5 - s3
+	SUB	r3, r14,r7		; r3 = x[1] = s7 - s1
+	ADD	r5, r3, r7, LSL #1	; r5 = x[3] = s7 + s1
+	SUB	r10,r10,r6		; r10= x[4] = s4 - s0
+	SUB	r11,r12,r8		; r11= x[5] = s6 - s2
+	ADD	r12,r10,r6, LSL #1	; r12= x[6] = s4 + s0
+	ADD	r14,r11,r8, LSL #1	; r14= x[7] = s6 + s2
 	STMDB	r1,{r2,r3,r4,r5,r10,r11,r12,r14}
 
-	@ mdct_butterfly_8
+	; mdct_butterfly_8
 	LDMIA	r1,{r6,r7,r8,r9,r10,r11,r12,r14}
-					@ r6 = x[0]
-					@ r7 = x[1]
-					@ r8 = x[2]
-					@ r9 = x[3]
-					@ r10= x[4]
-					@ r11= x[5]
-					@ r12= x[6]
-					@ r14= x[7]
-	ADD	r6, r6, r7		@ r6 = s0 = x[0] + x[1]
-	SUB	r7, r6, r7, LSL #1	@ r7 = s1 = x[0] - x[1]
-	ADD	r8, r8, r9		@ r8 = s2 = x[2] + x[3]
-	SUB	r9, r8, r9, LSL #1	@ r9 = s3 = x[2] - x[3]
-	ADD	r10,r10,r11		@ r10= s4 = x[4] + x[5]
-	SUB	r11,r10,r11,LSL #1	@ r11= s5 = x[4] - x[5]
-	ADD	r12,r12,r14		@ r12= s6 = x[6] + x[7]
-	SUB	r14,r12,r14,LSL #1	@ r14= s7 = x[6] - x[7]
-
-	ADD	r2, r11,r9		@ r2 = x[0] = s5 + s3
-	SUB	r4, r2, r9, LSL #1	@ r4 = x[2] = s5 - s3
-	SUB	r3, r14,r7		@ r3 = x[1] = s7 - s1
-	ADD	r5, r3, r7, LSL #1	@ r5 = x[3] = s7 + s1
-	SUB	r10,r10,r6		@ r10= x[4] = s4 - s0
-	SUB	r11,r12,r8		@ r11= x[5] = s6 - s2
-	ADD	r12,r10,r6, LSL #1	@ r12= x[6] = s4 + s0
-	ADD	r14,r11,r8, LSL #1	@ r14= x[7] = s6 + s2
+					; r6 = x[0]
+					; r7 = x[1]
+					; r8 = x[2]
+					; r9 = x[3]
+					; r10= x[4]
+					; r11= x[5]
+					; r12= x[6]
+					; r14= x[7]
+	ADD	r6, r6, r7		; r6 = s0 = x[0] + x[1]
+	SUB	r7, r6, r7, LSL #1	; r7 = s1 = x[0] - x[1]
+	ADD	r8, r8, r9		; r8 = s2 = x[2] + x[3]
+	SUB	r9, r8, r9, LSL #1	; r9 = s3 = x[2] - x[3]
+	ADD	r10,r10,r11		; r10= s4 = x[4] + x[5]
+	SUB	r11,r10,r11,LSL #1	; r11= s5 = x[4] - x[5]
+	ADD	r12,r12,r14		; r12= s6 = x[6] + x[7]
+	SUB	r14,r12,r14,LSL #1	; r14= s7 = x[6] - x[7]
+
+	ADD	r2, r11,r9		; r2 = x[0] = s5 + s3
+	SUB	r4, r2, r9, LSL #1	; r4 = x[2] = s5 - s3
+	SUB	r3, r14,r7		; r3 = x[1] = s7 - s1
+	ADD	r5, r3, r7, LSL #1	; r5 = x[3] = s7 + s1
+	SUB	r10,r10,r6		; r10= x[4] = s4 - s0
+	SUB	r11,r12,r8		; r11= x[5] = s6 - s2
+	ADD	r12,r10,r6, LSL #1	; r12= x[6] = s4 + s0
+	ADD	r14,r11,r8, LSL #1	; r14= x[7] = s6 + s2
 	STMIA	r1,{r2,r3,r4,r5,r10,r11,r12,r14}
 
-	@ block 2
+	; block 2
 	ADD	r1,r1,#16*4-8*4
 	ADD	r4,r1,#8*4
-	LDMIA	r4,{r5,r6,r9,r10}	@ r5 = x[ 8]
-					@ r6 = x[ 9]
-					@ r9 = x[10]
-					@ r10= x[11]
-	LDMIA	r1,{r7,r8,r11,r12}	@ r7 = x[0]
-					@ r8 = x[1]
-					@ r11= x[2]
-					@ r12= x[3]
-	SUB	r5, r5, r6		@ r5 = s0 = x[ 8] - x[ 9]
-	ADD	r6, r5, r6, LSL #1	@ r6 =      x[ 8] + x[ 9]  -> x[ 8]
-	SUB	r9, r9, r10		@ r9 = s1 = x[10] - x[11]
-	ADD	r10,r9, r10,LSL #1	@ r10=      x[10] + x[11]  -> x[10]
-	SUB	r8, r8, r7		@ r8 = s2 = x[ 1] - x[ 0]
-	ADD	r7, r8, r7, LSL #1	@ r7 =      x[ 1] + x[ 0]  -> x[ 9]
-	SUB	r12,r12,r11		@ r12= s3 = x[ 3] - x[ 2]
-	ADD	r11,r12,r11, LSL #1	@ r11=      x[ 3] + x[ 2]  -> x[11]
+	LDMIA	r4,{r5,r6,r9,r10}	; r5 = x[ 8]
+					; r6 = x[ 9]
+					; r9 = x[10]
+					; r10= x[11]
+	LDMIA	r1,{r7,r8,r11,r12}	; r7 = x[0]
+					; r8 = x[1]
+					; r11= x[2]
+					; r12= x[3]
+	SUB	r5, r5, r6		; r5 = s0 = x[ 8] - x[ 9]
+	ADD	r6, r5, r6, LSL #1	; r6 =      x[ 8] + x[ 9]  -> x[ 8]
+	SUB	r9, r9, r10		; r9 = s1 = x[10] - x[11]
+	ADD	r10,r9, r10,LSL #1	; r10=      x[10] + x[11]  -> x[10]
+	SUB	r8, r8, r7		; r8 = s2 = x[ 1] - x[ 0]
+	ADD	r7, r8, r7, LSL #1	; r7 =      x[ 1] + x[ 0]  -> x[ 9]
+	SUB	r12,r12,r11		; r12= s3 = x[ 3] - x[ 2]
+	ADD	r11,r12,r11, LSL #1	; r11=      x[ 3] + x[ 2]  -> x[11]
 	LDR	r14,cPI2_8
 	STMIA	r4!,{r6,r7,r10,r11}
 
-	SUB	r5, r5, r9		@ r5 = s0 - s1
-	ADD	r9, r5, r9, LSL #1	@ r9 = s0 + s1
-	SMULL	r6, r5, r14,r5		@ (r6,r5)  = (s0-s1)*cPI2_8
-	SUB	r12,r12,r8		@ r12= s3 - s2
-	ADD	r8, r12,r8, LSL #1	@ r8 = s3 + s2
+	SUB	r5, r5, r9		; r5 = s0 - s1
+	ADD	r9, r5, r9, LSL #1	; r9 = s0 + s1
+	SMULL	r6, r5, r14,r5		; (r6,r5)  = (s0-s1)*cPI2_8
+	SUB	r12,r12,r8		; r12= s3 - s2
+	ADD	r8, r12,r8, LSL #1	; r8 = s3 + s2
 
-	SMULL	r6, r8, r14,r8		@ (r6,r8)  = (s3+s2)*cPI2_8
+	SMULL	r6, r8, r14,r8		; (r6,r8)  = (s3+s2)*cPI2_8
 	MOV	r5, r5, LSL #1
-	SMULL	r6, r9, r14,r9		@ (r6,r9)  = (s0+s1)*cPI2_8
+	SMULL	r6, r9, r14,r9		; (r6,r9)  = (s0+s1)*cPI2_8
 	MOV	r8, r8, LSL #1
-	SMULL	r6, r12,r14,r12		@ (r6,r12) = (s3-s2)*cPI2_8
+	SMULL	r6, r12,r14,r12		; (r6,r12) = (s3-s2)*cPI2_8
 	MOV	r9, r9, LSL #1
 	MOV	r12,r12,LSL #1
 	STMIA	r1!,{r5,r8,r9,r12}
 
-	@ block4
-	LDMIA	r4,{r5,r6,r9,r10}	@ r5 = x[12]
-					@ r6 = x[13]
-					@ r9 = x[14]
-					@ r10= x[15]
-	LDMIA	r1,{r7,r8,r11,r12}	@ r7 = x[ 4]
-					@ r8 = x[ 5]
-					@ r11= x[ 6]
-					@ r12= x[ 7]
-	SUB	r5, r5, r6		@ r5 = s2 = x[12] - x[13]
-	ADD	r6, r5, r6, LSL #1	@ r6 =      x[12] + x[13]  -> x[12]
-	SUB	r9, r9, r10		@ r9 = s3 = x[14] - x[15]
-	ADD	r10,r9, r10,LSL #1	@ r10=      x[14] + x[15]  -> x[14]
-	SUB	r14,r7, r8		@ r14= s0 = x[ 4] - x[ 5]
-	ADD	r8, r14,r8, LSL #1	@ r8 =      x[ 4] + x[ 5]  -> x[13]
-	SUB	r7, r12,r11		@ r7 = s1 = x[ 7] - x[ 6]
-	ADD	r11,r7, r11, LSL #1	@ r11=      x[ 7] + x[ 6]  -> x[15]
+	; block4
+	LDMIA	r4,{r5,r6,r9,r10}	; r5 = x[12]
+					; r6 = x[13]
+					; r9 = x[14]
+					; r10= x[15]
+	LDMIA	r1,{r7,r8,r11,r12}	; r7 = x[ 4]
+					; r8 = x[ 5]
+					; r11= x[ 6]
+					; r12= x[ 7]
+	SUB	r5, r5, r6		; r5 = s2 = x[12] - x[13]
+	ADD	r6, r5, r6, LSL #1	; r6 =      x[12] + x[13]  -> x[12]
+	SUB	r9, r9, r10		; r9 = s3 = x[14] - x[15]
+	ADD	r10,r9, r10,LSL #1	; r10=      x[14] + x[15]  -> x[14]
+	SUB	r14,r7, r8		; r14= s0 = x[ 4] - x[ 5]
+	ADD	r8, r14,r8, LSL #1	; r8 =      x[ 4] + x[ 5]  -> x[13]
+	SUB	r7, r12,r11		; r7 = s1 = x[ 7] - x[ 6]
+	ADD	r11,r7, r11, LSL #1	; r11=      x[ 7] + x[ 6]  -> x[15]
 	STMIA	r4!,{r6,r8,r10,r11}
 	STMIA	r1!,{r5,r7,r9,r14}
 
-	@ mdct_butterfly_8
+	; mdct_butterfly_8
 	LDMDB	r1,{r6,r7,r8,r9,r10,r11,r12,r14}
-					@ r6 = x[0]
-					@ r7 = x[1]
-					@ r8 = x[2]
-					@ r9 = x[3]
-					@ r10= x[4]
-					@ r11= x[5]
-					@ r12= x[6]
-					@ r14= x[7]
-	ADD	r6, r6, r7		@ r6 = s0 = x[0] + x[1]
-	SUB	r7, r6, r7, LSL #1	@ r7 = s1 = x[0] - x[1]
-	ADD	r8, r8, r9		@ r8 = s2 = x[2] + x[3]
-	SUB	r9, r8, r9, LSL #1	@ r9 = s3 = x[2] - x[3]
-	ADD	r10,r10,r11		@ r10= s4 = x[4] + x[5]
-	SUB	r11,r10,r11,LSL #1	@ r11= s5 = x[4] - x[5]
-	ADD	r12,r12,r14		@ r12= s6 = x[6] + x[7]
-	SUB	r14,r12,r14,LSL #1	@ r14= s7 = x[6] - x[7]
-
-	ADD	r2, r11,r9		@ r2 = x[0] = s5 + s3
-	SUB	r4, r2, r9, LSL #1	@ r4 = x[2] = s5 - s3
-	SUB	r3, r14,r7		@ r3 = x[1] = s7 - s1
-	ADD	r5, r3, r7, LSL #1	@ r5 = x[3] = s7 + s1
-	SUB	r10,r10,r6		@ r10= x[4] = s4 - s0
-	SUB	r11,r12,r8		@ r11= x[5] = s6 - s2
-	ADD	r12,r10,r6, LSL #1	@ r12= x[6] = s4 + s0
-	ADD	r14,r11,r8, LSL #1	@ r14= x[7] = s6 + s2
+					; r6 = x[0]
+					; r7 = x[1]
+					; r8 = x[2]
+					; r9 = x[3]
+					; r10= x[4]
+					; r11= x[5]
+					; r12= x[6]
+					; r14= x[7]
+	ADD	r6, r6, r7		; r6 = s0 = x[0] + x[1]
+	SUB	r7, r6, r7, LSL #1	; r7 = s1 = x[0] - x[1]
+	ADD	r8, r8, r9		; r8 = s2 = x[2] + x[3]
+	SUB	r9, r8, r9, LSL #1	; r9 = s3 = x[2] - x[3]
+	ADD	r10,r10,r11		; r10= s4 = x[4] + x[5]
+	SUB	r11,r10,r11,LSL #1	; r11= s5 = x[4] - x[5]
+	ADD	r12,r12,r14		; r12= s6 = x[6] + x[7]
+	SUB	r14,r12,r14,LSL #1	; r14= s7 = x[6] - x[7]
+
+	ADD	r2, r11,r9		; r2 = x[0] = s5 + s3
+	SUB	r4, r2, r9, LSL #1	; r4 = x[2] = s5 - s3
+	SUB	r3, r14,r7		; r3 = x[1] = s7 - s1
+	ADD	r5, r3, r7, LSL #1	; r5 = x[3] = s7 + s1
+	SUB	r10,r10,r6		; r10= x[4] = s4 - s0
+	SUB	r11,r12,r8		; r11= x[5] = s6 - s2
+	ADD	r12,r10,r6, LSL #1	; r12= x[6] = s4 + s0
+	ADD	r14,r11,r8, LSL #1	; r14= x[7] = s6 + s2
 	STMDB	r1,{r2,r3,r4,r5,r10,r11,r12,r14}
 
-	@ mdct_butterfly_8
+	; mdct_butterfly_8
 	LDMIA	r1,{r6,r7,r8,r9,r10,r11,r12,r14}
-					@ r6 = x[0]
-					@ r7 = x[1]
-					@ r8 = x[2]
-					@ r9 = x[3]
-					@ r10= x[4]
-					@ r11= x[5]
-					@ r12= x[6]
-					@ r14= x[7]
-	ADD	r6, r6, r7		@ r6 = s0 = x[0] + x[1]
-	SUB	r7, r6, r7, LSL #1	@ r7 = s1 = x[0] - x[1]
-	ADD	r8, r8, r9		@ r8 = s2 = x[2] + x[3]
-	SUB	r9, r8, r9, LSL #1	@ r9 = s3 = x[2] - x[3]
-	ADD	r10,r10,r11		@ r10= s4 = x[4] + x[5]
-	SUB	r11,r10,r11,LSL #1	@ r11= s5 = x[4] - x[5]
-	ADD	r12,r12,r14		@ r12= s6 = x[6] + x[7]
-	SUB	r14,r12,r14,LSL #1	@ r14= s7 = x[6] - x[7]
-
-	ADD	r2, r11,r9		@ r2 = x[0] = s5 + s3
-	SUB	r4, r2, r9, LSL #1	@ r4 = x[2] = s5 - s3
-	SUB	r3, r14,r7		@ r3 = x[1] = s7 - s1
-	ADD	r5, r3, r7, LSL #1	@ r5 = x[3] = s7 + s1
-	SUB	r10,r10,r6		@ r10= x[4] = s4 - s0
-	SUB	r11,r12,r8		@ r11= x[5] = s6 - s2
-	ADD	r12,r10,r6, LSL #1	@ r12= x[6] = s4 + s0
-	ADD	r14,r11,r8, LSL #1	@ r14= x[7] = s6 + s2
+					; r6 = x[0]
+					; r7 = x[1]
+					; r8 = x[2]
+					; r9 = x[3]
+					; r10= x[4]
+					; r11= x[5]
+					; r12= x[6]
+					; r14= x[7]
+	ADD	r6, r6, r7		; r6 = s0 = x[0] + x[1]
+	SUB	r7, r6, r7, LSL #1	; r7 = s1 = x[0] - x[1]
+	ADD	r8, r8, r9		; r8 = s2 = x[2] + x[3]
+	SUB	r9, r8, r9, LSL #1	; r9 = s3 = x[2] - x[3]
+	ADD	r10,r10,r11		; r10= s4 = x[4] + x[5]
+	SUB	r11,r10,r11,LSL #1	; r11= s5 = x[4] - x[5]
+	ADD	r12,r12,r14		; r12= s6 = x[6] + x[7]
+	SUB	r14,r12,r14,LSL #1	; r14= s7 = x[6] - x[7]
+
+	ADD	r2, r11,r9		; r2 = x[0] = s5 + s3
+	SUB	r4, r2, r9, LSL #1	; r4 = x[2] = s5 - s3
+	SUB	r3, r14,r7		; r3 = x[1] = s7 - s1
+	ADD	r5, r3, r7, LSL #1	; r5 = x[3] = s7 + s1
+	SUB	r10,r10,r6		; r10= x[4] = s4 - s0
+	SUB	r11,r12,r8		; r11= x[5] = s6 - s2
+	ADD	r12,r10,r6, LSL #1	; r12= x[6] = s4 + s0
+	ADD	r14,r11,r8, LSL #1	; r14= x[7] = s6 + s2
 	STMIA	r1,{r2,r3,r4,r5,r10,r11,r12,r14}
 
 	ADD	r1,r1,#8*4
@@ -953,117 +953,117 @@ mdct_bufferflies_loop3:
 
 	LDMFD	r13,{r0-r3}
 
-mdct_bitreverse_arm:
-	@ r0 = points = n
-	@ r1 = in
-	@ r2 = step
-	@ r3 = shift
+mdct_bitreverse_arm
+	; r0 = points = n
+	; r1 = in
+	; r2 = step
+	; r3 = shift
 
-	MOV	r4, #0			@ r4 = bit = 0
-	ADD	r5, r1, r0, LSL #1	@ r5 = w = x + (n>>1)
+	MOV	r4, #0			; r4 = bit = 0
+	ADD	r5, r1, r0, LSL #1	; r5 = w = x + (n>>1)
 	ADR	r6, bitrev
 	SUB	r5, r5, #8
-brev_lp:
+brev_lp
 	LDRB	r7, [r6, r4, LSR #6]
 	AND	r8, r4, #0x3f
 	LDRB	r8, [r6, r8]
-	ADD	r4, r4, #1		@ bit++
-	@ stall XScale
-	ORR	r7, r7, r8, LSL #6	@ r7 = bitrev[bit]
+	ADD	r4, r4, #1		; bit++
+	; stall XScale
+	ORR	r7, r7, r8, LSL #6	; r7 = bitrev[bit]
 	MOV	r7, r7, LSR r3
-	ADD	r9, r1, r7, LSL #2	@ r9 = xx = x + (b>>shift)
-	CMP	r5, r9			@ if (w > xx)
-	LDR	r10,[r5],#-8		@   r10 = w[0]		w -= 2
-	LDRGT	r11,[r5,#12]		@   r11 = w[1]
-	LDRGT	r12,[r9]		@   r12 = xx[0]
-	LDRGT	r14,[r9,#4]		@   r14 = xx[1]
-	STRGT	r10,[r9]		@   xx[0]= w[0]
-	STRGT	r11,[r9,#4]		@   xx[1]= w[1]
-	STRGT	r12,[r5,#8]		@   w[0] = xx[0]
-	STRGT	r14,[r5,#12]		@   w[1] = xx[1]
+	ADD	r9, r1, r7, LSL #2	; r9 = xx = x + (b>>shift)
+	CMP	r5, r9			; if (w > xx)
+	LDR	r10,[r5],#-8		;   r10 = w[0]		w -= 2
+	LDRGT	r11,[r5,#12]		;   r11 = w[1]
+	LDRGT	r12,[r9]		;   r12 = xx[0]
+	LDRGT	r14,[r9,#4]		;   r14 = xx[1]
+	STRGT	r10,[r9]		;   xx[0]= w[0]
+	STRGT	r11,[r9,#4]		;   xx[1]= w[1]
+	STRGT	r12,[r5,#8]		;   w[0] = xx[0]
+	STRGT	r14,[r5,#12]		;   w[1] = xx[1]
 	CMP	r5,r1
 	BGT	brev_lp
 
-	@ mdct_step7
-	@ r0 = points
-	@ r1 = in
-	@ r2 = step
-	@ r3 = shift
-
-	CMP	r2, #4			@ r5 = T = (step>=4) ?
-	LDRGE	r5, =sincos_lookup0	@          sincos_lookup0 +
-	LDRLT	r5, =sincos_lookup1	@          sincos_lookup0 +
-	ADD	r7, r1, r0, LSL #1	@ r7 = w1 = x + (n>>1)
-	ADDGE	r5, r5, r2, LSL #1	@		            (step>>1)
-	ADD	r8, r5, #1024*4		@ r8 = Ttop
-step7_loop1:
-	LDR	r6, [r1]		@ r6 = w0[0]
-	LDR	r9, [r1,#4]		@ r9 = w0[1]
-	LDR	r10,[r7,#-8]!		@ r10= w1[0]	w1 -= 2
-	LDR	r11,[r7,#4]		@ r11= w1[1]
-	LDR	r14,[r5,#4]		@ r14= T[1]
-	LDR	r12,[r5],r2,LSL #2	@ r12= T[0]	T += step
-
-	ADD	r6, r6, r10		@ r6 = s0 = w0[0] + w1[0]
-	SUB	r10,r6, r10,LSL #1	@ r10= s1b= w0[0] - w1[0]
-	SUB	r11,r11,r9		@ r11= s1 = w1[1] - w0[1]
-	ADD	r9, r11,r9, LSL #1	@ r9 = s0b= w1[1] + w0[1]
-
-	@ Can save 1 cycle by using SMULL SMLAL - at the cost of being
-	@ 1 off.
-	SMULL	r0, r3, r6, r14		@ (r0,r3)   = s0*T[1]
-	SMULL	r0, r4, r11,r12		@ (r0,r4)  += s1*T[0] = s2
+	; mdct_step7
+	; r0 = points
+	; r1 = in
+	; r2 = step
+	; r3 = shift
+
+	CMP	r2, #4			; r5 = T = (step>=4) ?
+	LDRGE	r5, =sincos_lookup0	;          sincos_lookup0 +
+	LDRLT	r5, =sincos_lookup1	;          sincos_lookup0 +
+	ADD	r7, r1, r0, LSL #1	; r7 = w1 = x + (n>>1)
+	ADDGE	r5, r5, r2, LSL #1	;		            (step>>1)
+	ADD	r8, r5, #1024*4		; r8 = Ttop
+step7_loop1
+	LDR	r6, [r1]		; r6 = w0[0]
+	LDR	r9, [r1,#4]		; r9 = w0[1]
+	LDR	r10,[r7,#-8]!		; r10= w1[0]	w1 -= 2
+	LDR	r11,[r7,#4]		; r11= w1[1]
+	LDR	r14,[r5,#4]		; r14= T[1]
+	LDR	r12,[r5],r2,LSL #2	; r12= T[0]	T += step
+
+	ADD	r6, r6, r10		; r6 = s0 = w0[0] + w1[0]
+	SUB	r10,r6, r10,LSL #1	; r10= s1b= w0[0] - w1[0]
+	SUB	r11,r11,r9		; r11= s1 = w1[1] - w0[1]
+	ADD	r9, r11,r9, LSL #1	; r9 = s0b= w1[1] + w0[1]
+
+	; Can save 1 cycle by using SMULL SMLAL - at the cost of being
+	; 1 off.
+	SMULL	r0, r3, r6, r14		; (r0,r3)   = s0*T[1]
+	SMULL	r0, r4, r11,r12		; (r0,r4)  += s1*T[0] = s2
 	ADD	r3, r3, r4
-	SMULL	r0, r14,r11,r14		@ (r0,r14)  = s1*T[1]
-	SMULL	r0, r12,r6, r12		@ (r0,r12) += s0*T[0] = s3
+	SMULL	r0, r14,r11,r14		; (r0,r14)  = s1*T[1]
+	SMULL	r0, r12,r6, r12		; (r0,r12) += s0*T[0] = s3
 	SUB	r14,r14,r12
 
-	@ r9 = s0b<<1
-	@ r10= s1b<<1
-	ADD	r9, r3, r9, ASR #1	@ r9 = s0b + s2
-	SUB	r3, r9, r3, LSL #1	@ r3 = s0b - s2
+	; r9 = s0b<<1
+	; r10= s1b<<1
+	ADD	r9, r3, r9, ASR #1	; r9 = s0b + s2
+	SUB	r3, r9, r3, LSL #1	; r3 = s0b - s2
 
-	SUB	r12,r14,r10,ASR #1	@ r12= s3  - s1b
-	ADD	r10,r14,r10,ASR #1	@ r10= s3  + s1b
+	SUB	r12,r14,r10,ASR #1	; r12= s3  - s1b
+	ADD	r10,r14,r10,ASR #1	; r10= s3  + s1b
 	STR	r9, [r1],#4
-	STR	r10,[r1],#4		@ w0 += 2
+	STR	r10,[r1],#4		; w0 += 2
 	STR	r3, [r7]
 	STR	r12,[r7,#4]
 
 	CMP	r5,r8
 	BLT	step7_loop1
 
-step7_loop2:
-	LDR	r6, [r1]		@ r6 = w0[0]
-	LDR	r9, [r1,#4]		@ r9 = w0[1]
-	LDR	r10,[r7,#-8]!		@ r10= w1[0]	w1 -= 2
-	LDR	r11,[r7,#4]		@ r11= w1[1]
-	LDR	r14,[r5,-r2,LSL #2]!	@ r12= T[1]	T -= step
-	LDR	r12,[r5,#4]		@ r14= T[0]
-
-	ADD	r6, r6, r10		@ r6 = s0 = w0[0] + w1[0]
-	SUB	r10,r6, r10,LSL #1	@ r10= s1b= w0[0] - w1[0]
-	SUB	r11,r11,r9		@ r11= s1 = w1[1] - w0[1]
-	ADD	r9, r11,r9, LSL #1	@ r9 = s0b= w1[1] + w0[1]
-
-	@ Can save 1 cycle by using SMULL SMLAL - at the cost of being
-	@ 1 off.
-	SMULL	r0, r3, r6, r14		@ (r0,r3)   = s0*T[0]
-	SMULL	r0, r4, r11,r12		@ (r0,r4)  += s1*T[1] = s2
+step7_loop2
+	LDR	r6, [r1]		; r6 = w0[0]
+	LDR	r9, [r1,#4]		; r9 = w0[1]
+	LDR	r10,[r7,#-8]!		; r10= w1[0]	w1 -= 2
+	LDR	r11,[r7,#4]		; r11= w1[1]
+	LDR	r14,[r5,-r2,LSL #2]!	; r12= T[1]	T -= step
+	LDR	r12,[r5,#4]		; r14= T[0]
+
+	ADD	r6, r6, r10		; r6 = s0 = w0[0] + w1[0]
+	SUB	r10,r6, r10,LSL #1	; r10= s1b= w0[0] - w1[0]
+	SUB	r11,r11,r9		; r11= s1 = w1[1] - w0[1]
+	ADD	r9, r11,r9, LSL #1	; r9 = s0b= w1[1] + w0[1]
+
+	; Can save 1 cycle by using SMULL SMLAL - at the cost of being
+	; 1 off.
+	SMULL	r0, r3, r6, r14		; (r0,r3)   = s0*T[0]
+	SMULL	r0, r4, r11,r12		; (r0,r4)  += s1*T[1] = s2
 	ADD	r3, r3, r4
-	SMULL	r0, r14,r11,r14		@ (r0,r14)  = s1*T[0]
-	SMULL	r0, r12,r6, r12		@ (r0,r12) += s0*T[1] = s3
+	SMULL	r0, r14,r11,r14		; (r0,r14)  = s1*T[0]
+	SMULL	r0, r12,r6, r12		; (r0,r12) += s0*T[1] = s3
 	SUB	r14,r14,r12
 
-	@ r9 = s0b<<1
-	@ r10= s1b<<1
-	ADD	r9, r3, r9, ASR #1	@ r9 = s0b + s2
-	SUB	r3, r9, r3, LSL #1	@ r3 = s0b - s2
+	; r9 = s0b<<1
+	; r10= s1b<<1
+	ADD	r9, r3, r9, ASR #1	; r9 = s0b + s2
+	SUB	r3, r9, r3, LSL #1	; r3 = s0b - s2
 
-	SUB	r12,r14,r10,ASR #1	@ r12= s3  - s1b
-	ADD	r10,r14,r10,ASR #1	@ r10= s3  + s1b
+	SUB	r12,r14,r10,ASR #1	; r12= s3  - s1b
+	ADD	r10,r14,r10,ASR #1	; r10= s3  + s1b
 	STR	r9, [r1],#4
-	STR	r10,[r1],#4		@ w0 += 2
+	STR	r10,[r1],#4		; w0 += 2
 	STR	r3, [r7]
 	STR	r12,[r7,#4]
 
@@ -1072,116 +1072,116 @@ step7_loop2:
 
 	LDMFD	r13!,{r0-r3}
 
-	@ r0 = points
-	@ r1 = in
-	@ r2 = step
-	@ r3 = shift
-	MOV	r2, r2, ASR #2		@ r2 = step >>= 2
+	; r0 = points
+	; r1 = in
+	; r2 = step
+	; r3 = shift
+	MOV	r2, r2, ASR #2		; r2 = step >>= 2
 	CMP	r2, #0
 	CMPNE	r2, #1
 	BEQ	mdct_end
 
-	@ step > 1 (default case)
-	CMP	r2, #4			@ r5 = T = (step>=4) ?
-	LDRGE	r5, =sincos_lookup0	@          sincos_lookup0 +
-	LDRLT	r5, =sincos_lookup1	@          sincos_lookup1
-	ADD	r7, r1, r0, LSL #1	@ r7 = iX = x + (n>>1)
-	ADDGE	r5, r5, r2, LSL #1	@		            (step>>1)
-mdct_step8_default:
-	LDR	r6, [r1],#4		@ r6 =  s0 = x[0]
-	LDR	r8, [r1],#4		@ r8 = -s1 = x[1]
-	LDR	r12,[r5,#4]       	@ r12= T[1]
-	LDR	r14,[r5],r2,LSL #2	@ r14= T[0]	T += step
-	RSB	r8, r8, #0		@ r8 = s1
-
-	@ XPROD31(s0, s1, T[0], T[1], x, x+1)
-	@ x[0] = s0 * T[0] + s1 * T[1]      x[1] = s1 * T[0] - s0 * T[1]
-	SMULL	r9, r10, r8, r12	@ (r9,r10)  = s1 * T[1]
+	; step > 1 (default case)
+	CMP	r2, #4			; r5 = T = (step>=4) ?
+	LDRGE	r5, =sincos_lookup0	;          sincos_lookup0 +
+	LDRLT	r5, =sincos_lookup1	;          sincos_lookup1
+	ADD	r7, r1, r0, LSL #1	; r7 = iX = x + (n>>1)
+	ADDGE	r5, r5, r2, LSL #1	;		            (step>>1)
+mdct_step8_default
+	LDR	r6, [r1],#4		; r6 =  s0 = x[0]
+	LDR	r8, [r1],#4		; r8 = -s1 = x[1]
+	LDR	r12,[r5,#4]       	; r12= T[1]
+	LDR	r14,[r5],r2,LSL #2	; r14= T[0]	T += step
+	RSB	r8, r8, #0		; r8 = s1
+
+	; XPROD31(s0, s1, T[0], T[1], x, x+1)
+	; x[0] = s0 * T[0] + s1 * T[1]      x[1] = s1 * T[0] - s0 * T[1]
+	SMULL	r9, r10, r8, r12	; (r9,r10)  = s1 * T[1]
 	CMP	r1, r7
-	SMLAL	r9, r10, r6, r14	@ (r9,r10) += s0 * T[0]
-	RSB	r6, r6, #0		@ r6 = -s0
-	SMULL	r9, r11, r8, r14	@ (r9,r11)  = s1 * T[0]
+	SMLAL	r9, r10, r6, r14	; (r9,r10) += s0 * T[0]
+	RSB	r6, r6, #0		; r6 = -s0
+	SMULL	r9, r11, r8, r14	; (r9,r11)  = s1 * T[0]
 	MOV	r10,r10,LSL #1
-	SMLAL	r9, r11, r6, r12	@ (r9,r11) -= s0 * T[1]
+	SMLAL	r9, r11, r6, r12	; (r9,r11) -= s0 * T[1]
 	STR	r10,[r1,#-8]
 	MOV	r11,r11,LSL #1
 	STR	r11,[r1,#-4]
 	BLT	mdct_step8_default
 
-mdct_end:
+mdct_end
 	MOV	r0, r2
 	LDMFD	r13!,{r4-r11,PC}
 
-cPI1_8:
-	.word	0x7641af3d
-cPI2_8:
-	.word	0x5a82799a
-cPI3_8:
-	.word	0x30fbc54d
-bitrev:
-	.byte	0
-	.byte	32
-	.byte	16
-	.byte	48
-	.byte	8
-	.byte	40
-	.byte	24
-	.byte	56
-	.byte	4
-	.byte	36
-	.byte	20
-	.byte	52
-	.byte	12
-	.byte	44
-	.byte	28
-	.byte	60
-	.byte	2
-	.byte	34
-	.byte	18
-	.byte	50
-	.byte	10
-	.byte	42
-	.byte	26
-	.byte	58
-	.byte	6
-	.byte	38
-	.byte	22
-	.byte	54
-	.byte	14
-	.byte	46
-	.byte	30
-	.byte	62
-	.byte	1
-	.byte	33
-	.byte	17
-	.byte	49
-	.byte	9
-	.byte	41
-	.byte	25
-	.byte	57
-	.byte	5
-	.byte	37
-	.byte	21
-	.byte	53
-	.byte	13
-	.byte	45
-	.byte	29
-	.byte	61
-	.byte	3
-	.byte	35
-	.byte	19
-	.byte	51
-	.byte	11
-	.byte	43
-	.byte	27
-	.byte	59
-	.byte	7
-	.byte	39
-	.byte	23
-	.byte	55
-	.byte	15
-	.byte	47
-	.byte	31
-	.byte	63
-
-	@ END
+cPI1_8
+	DCD	0x7641af3d
+cPI2_8
+	DCD	0x5a82799a
+cPI3_8
+	DCD	0x30fbc54d
+bitrev
+	DCB	0
+	DCB	32
+	DCB	16
+	DCB	48
+	DCB	8
+	DCB	40
+	DCB	24
+	DCB	56
+	DCB	4
+	DCB	36
+	DCB	20
+	DCB	52
+	DCB	12
+	DCB	44
+	DCB	28
+	DCB	60
+	DCB	2
+	DCB	34
+	DCB	18
+	DCB	50
+	DCB	10
+	DCB	42
+	DCB	26
+	DCB	58
+	DCB	6
+	DCB	38
+	DCB	22
+	DCB	54
+	DCB	14
+	DCB	46
+	DCB	30
+	DCB	62
+	DCB	1
+	DCB	33
+	DCB	17
+	DCB	49
+	DCB	9
+	DCB	41
+	DCB	25
+	DCB	57
+	DCB	5
+	DCB	37
+	DCB	21
+	DCB	53
+	DCB	13
+	DCB	45
+	DCB	29
+	DCB	61
+	DCB	3
+	DCB	35
+	DCB	19
+	DCB	51
+	DCB	11
+	DCB	43
+	DCB	27
+	DCB	59
+	DCB	7
+	DCB	39
+	DCB	23
+	DCB	55
+	DCB	15
+	DCB	47
+	DCB	31
+	DCB	63
+
+	END
diff --git a/mdctLARM.s b/mdctLARM.s
index 9444b73..9fb397e 100644
--- a/mdctLARM.s
+++ b/mdctLARM.s
@@ -1,92 +1,92 @@
-@ Tremolo library
-@ Copyright (C) 2009 Robin Watts for Pinknoise Productions Ltd
+; Tremolo library
+; Copyright (C) 2009 Robin Watts for Pinknoise Productions Ltd
 
-    .text
+	AREA	|.text|, CODE, READONLY
 
-	@ low accuracy version
+	; low accuracy version
 
-	.global mdct_backward_arm_low
-	.global mdct_shift_right_arm_low
-	.global mdct_unroll_prelap_arm_low
-	.global mdct_unroll_part2_arm_low
-	.global mdct_unroll_part3_arm_low
-	.global mdct_unroll_postlap_arm_low
+	EXPORT mdct_backward_arm_low
+	EXPORT mdct_shift_right_arm_low
+	EXPORT mdct_unroll_prelap_arm_low
+	EXPORT mdct_unroll_part2_arm_low
+	EXPORT mdct_unroll_part3_arm_low
+	EXPORT mdct_unroll_postlap_arm_low
 
-	.extern	sincos_lookup0
-	.extern	sincos_lookup1
+	IMPORT	sincos_lookup0
+	IMPORT	sincos_lookup1
 
-mdct_unroll_prelap_arm_low:
-	@ r0 = out
-	@ r1 = post
-	@ r2 = r
-	@ r3 = step
+mdct_unroll_prelap_arm_low
+	; r0 = out
+	; r1 = post
+	; r2 = r
+	; r3 = step
 	STMFD	r13!,{r4-r7,r14}
 	MVN	r4, #0x8000
 	MOV	r3, r3, LSL #1
-	SUB	r1, r2, r1		@ r1 = r - post
-	SUBS	r1, r1, #16		@ r1 = r - post - 16
+	SUB	r1, r2, r1		; r1 = r - post
+	SUBS	r1, r1, #16		; r1 = r - post - 16
 	BLT	unroll_over
-unroll_loop:
+unroll_loop
 	LDMDB	r2!,{r5,r6,r7,r12}
 
-	MOV	r5, r5, ASR #9		@ r5 = (*--r)>>9
-	MOV	r6, r6, ASR #9		@ r6 = (*--r)>>9
-	MOV	r7, r7, ASR #9		@ r7 = (*--r)>>9
-	MOV	r12,r12,ASR #9		@ r12= (*--r)>>9
+	MOV	r5, r5, ASR #9		; r5 = (*--r)>>9
+	MOV	r6, r6, ASR #9		; r6 = (*--r)>>9
+	MOV	r7, r7, ASR #9		; r7 = (*--r)>>9
+	MOV	r12,r12,ASR #9		; r12= (*--r)>>9
 
 	MOV	r14,r12,ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r12,r4, r14,ASR #31
 	STRH	r12,[r0], r3
 
 	MOV	r14,r7, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r7, r4, r14,ASR #31
 	STRH	r7, [r0], r3
 
 	MOV	r14,r6, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r6, r4, r14,ASR #31
 	STRH	r6, [r0], r3
 
 	MOV	r14,r5, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r5, r4, r14,ASR #31
 	STRH	r5, [r0], r3
 
 	SUBS	r1, r1, #16
 	BGE	unroll_loop
 
-unroll_over:
+unroll_over
 	ADDS	r1, r1, #16
 	BLE	unroll_end
-unroll_loop2:
+unroll_loop2
 	LDR	r5,[r2,#-4]!
-	@ stall
-	@ stall (Xscale)
-	MOV	r5, r5, ASR #9		@ r5 = (*--r)>>9
+	; stall
+	; stall (Xscale)
+	MOV	r5, r5, ASR #9		; r5 = (*--r)>>9
 	MOV	r14,r5, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r5, r4, r14,ASR #31
 	STRH	r5, [r0], r3
 	SUBS	r1, r1, #4
 	BGT	unroll_loop2
-unroll_end:
+unroll_end
 	LDMFD	r13!,{r4-r7,PC}
 
-mdct_unroll_postlap_arm_low:
-	@ r0 = out
-	@ r1 = post
-	@ r2 = l
-	@ r3 = step
+mdct_unroll_postlap_arm_low
+	; r0 = out
+	; r1 = post
+	; r2 = l
+	; r3 = step
 	STMFD	r13!,{r4-r7,r14}
 	MVN	r4, #0x8000
 	MOV	r3, r3, LSL #1
-	SUB	r1, r1, r2		@ r1 = post - l
-	MOV	r1, r1, ASR #1		@ r1 = (post - l)>>1
-	SUBS	r1, r1, #16		@ r1 = ((post - l)>>1) - 4
+	SUB	r1, r1, r2		; r1 = post - l
+	MOV	r1, r1, ASR #1		; r1 = (post - l)>>1
+	SUBS	r1, r1, #16		; r1 = ((post - l)>>1) - 4
 	BLT	unroll_over3
-unroll_loop3:
+unroll_loop3
 	LDR	r12,[r2],#8
 	LDR	r7, [r2],#8
 	LDR	r6, [r2],#8
@@ -97,145 +97,145 @@ unroll_loop3:
 	RSB	r6, r6, #0
 	RSB	r7, r7, #0
 
-	MOV	r12, r12,ASR #9		@ r12= (-*l)>>9
-	MOV	r5,  r5, ASR #9		@ r5 = (-*l)>>9
-	MOV	r6,  r6, ASR #9		@ r6 = (-*l)>>9
-	MOV	r7,  r7, ASR #9		@ r7 = (-*l)>>9
+	MOV	r12, r12,ASR #9		; r12= (-*l)>>9
+	MOV	r5,  r5, ASR #9		; r5 = (-*l)>>9
+	MOV	r6,  r6, ASR #9		; r6 = (-*l)>>9
+	MOV	r7,  r7, ASR #9		; r7 = (-*l)>>9
 
 	MOV	r14,r12,ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r12,r4, r14,ASR #31
 	STRH	r12,[r0], r3
 
 	MOV	r14,r7, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r7, r4, r14,ASR #31
 	STRH	r7, [r0], r3
 
 	MOV	r14,r6, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r6, r4, r14,ASR #31
 	STRH	r6, [r0], r3
 
 	MOV	r14,r5, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r5, r4, r14,ASR #31
 	STRH	r5, [r0], r3
 
 	SUBS	r1, r1, #16
 	BGE	unroll_loop3
 
-unroll_over3:
+unroll_over3
 	ADDS	r1, r1, #16
 	BLE	unroll_over4
-unroll_loop4:
+unroll_loop4
 	LDR	r5,[r2], #8
-	@ stall
-	@ stall (Xscale)
+	; stall
+	; stall (Xscale)
 	RSB	r5, r5, #0
-	MOV	r5, r5, ASR #9		@ r5 = (-*l)>>9
+	MOV	r5, r5, ASR #9		; r5 = (-*l)>>9
 	MOV	r14,r5, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r5, r4, r14,ASR #31
 	STRH	r5, [r0], r3
 	SUBS	r1, r1, #4
 	BGT	unroll_loop4
-unroll_over4:
+unroll_over4
 	LDMFD	r13!,{r4-r7,PC}
 
-mdct_unroll_part2_arm_low:
-	@ r0 = out
-	@ r1 = post
-	@ r2 = l
-	@ r3 = r
-	@ <> = step
-	@ <> = wL
-	@ <> = wR
+mdct_unroll_part2_arm_low
+	; r0 = out
+	; r1 = post
+	; r2 = l
+	; r3 = r
+	; <> = step
+	; <> = wL
+	; <> = wR
 	MOV	r12,r13
 	STMFD	r13!,{r4,r6-r11,r14}
-	LDMFD	r12,{r8,r9,r10}		@ r8 = step
-					@ r9 = wL
-					@ r10= wR
+	LDMFD	r12,{r8,r9,r10}		; r8 = step
+					; r9 = wL
+					; r10= wR
 	MVN	r4, #0x8000
 	MOV	r8, r8, LSL #1
-	SUBS	r1, r3, r1		@ r1 = (r - post)
+	SUBS	r1, r3, r1		; r1 = (r - post)
 	BLE	unroll_over5
-unroll_loop5:
-	LDR	r12,[r2, #-8]!		@ r12= *l       (but l -= 2 first)
-	LDR	r7, [r3, #-4]!		@ r7 = *--r
-	LDRB	r6, [r10,#-1]!		@ r6 = *--wR
-	LDRB	r11,[r9],#1		@ r11= *wL++
+unroll_loop5
+	LDR	r12,[r2, #-8]!		; r12= *l       (but l -= 2 first)
+	LDR	r7, [r3, #-4]!		; r7 = *--r
+	LDRB	r6, [r10,#-1]!		; r6 = *--wR
+	LDRB	r11,[r9],#1		; r11= *wL++
 
 	MOV	r12, r12, ASR #8
-	@ Can save a cycle here, at the cost of 1bit errors in rounding
-	MUL	r11,r12,r11		@ r11  = *l   * *wL++
+	; Can save a cycle here, at the cost of 1bit errors in rounding
+	MUL	r11,r12,r11		; r11  = *l   * *wL++
 	MOV	r7, r7, ASR #8
-	MLA	r6, r7, r6, r11		@ r6   = *--r * *--wR
+	MLA	r6, r7, r6, r11		; r6   = *--r * *--wR
 	MOV	r6, r6, ASR #9
 	MOV	r14,r6, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r6, r4, r14,ASR #31
 	STRH	r6, [r0], r8
 
 	SUBS	r1, r1, #4
 	BGT	unroll_loop5
 
-unroll_over5:
+unroll_over5
 	LDMFD	r13!,{r4,r6-r11,PC}
 
-mdct_unroll_part3_arm_low:
-	@ r0 = out
-	@ r1 = post
-	@ r2 = l
-	@ r3 = r
-	@ <> = step
-	@ <> = wL
-	@ <> = wR
+mdct_unroll_part3_arm_low
+	; r0 = out
+	; r1 = post
+	; r2 = l
+	; r3 = r
+	; <> = step
+	; <> = wL
+	; <> = wR
 	MOV	r12,r13
 	STMFD	r13!,{r4,r6-r11,r14}
-	LDMFD	r12,{r8,r9,r10}		@ r8 = step
-					@ r9 = wL
-					@ r10= wR
+	LDMFD	r12,{r8,r9,r10}		; r8 = step
+					; r9 = wL
+					; r10= wR
 	MVN	r4, #0x8000
 	MOV	r8, r8, LSL #1
-	SUBS	r1, r1, r3		@ r1 = (post - r)
+	SUBS	r1, r1, r3		; r1 = (post - r)
 	BLE	unroll_over6
-unroll_loop6:
-	LDR	r12,[r2],#8		@ r12= *l       (but l += 2 first)
-	LDR	r7, [r3],#4		@ r7 = *r++
-	LDRB	r11,[r9],#1		@ r11= *wL++
-	LDRB	r6, [r10,#-1]!		@ r6 = *--wR
+unroll_loop6
+	LDR	r12,[r2],#8		; r12= *l       (but l += 2 first)
+	LDR	r7, [r3],#4		; r7 = *r++
+	LDRB	r11,[r9],#1		; r11= *wL++
+	LDRB	r6, [r10,#-1]!		; r6 = *--wR
 
-	@ Can save a cycle here, at the cost of 1bit errors in rounding
+	; Can save a cycle here, at the cost of 1bit errors in rounding
 	MOV	r12,r12,ASR #8
-	MUL	r11,r12,r11		@ (r14,r11)  = *l   * *wL++
+	MUL	r11,r12,r11		; (r14,r11)  = *l   * *wL++
 	MOV	r7, r7, ASR #8
-	MUL	r6, r7, r6		@ (r14,r6)   = *--r * *--wR
+	MUL	r6, r7, r6		; (r14,r6)   = *--r * *--wR
 	SUB	r6, r6, r11
 	MOV	r6, r6, ASR #9
 	MOV	r14,r6, ASR #15
-	TEQ	r14,r14,ASR #31		@ if r14==0 || r14==-1 then in range
+	TEQ	r14,r14,ASR #31		; if r14==0 || r14==-1 then in range
 	EORNE	r6, r4, r14,ASR #31
 	STRH	r6, [r0], r8
 
 	SUBS	r1, r1, #4
 	BGT	unroll_loop6
 
-unroll_over6:
+unroll_over6
 	LDMFD	r13!,{r4,r6-r11,PC}
 
-mdct_shift_right_arm_low:
-	@ r0 = n
-	@ r1 = in
-	@ r2 = right
+mdct_shift_right_arm_low
+	; r0 = n
+	; r1 = in
+	; r2 = right
 	STMFD	r13!,{r4-r11,r14}
 
-	MOV	r0, r0, LSR #2		@ n >>= 2
+	MOV	r0, r0, LSR #2		; n >>= 2
 	ADD	r1, r1, #4
 
 	SUBS	r0, r0,	#8
 	BLT	sr_less_than_8
-sr_loop:
+sr_loop
 	LDR	r3, [r1], #8
 	LDR	r4, [r1], #8
 	LDR	r5, [r1], #8
@@ -247,225 +247,225 @@ sr_loop:
 	SUBS	r0, r0, #8
 	STMIA	r2!,{r3,r4,r5,r6,r7,r8,r12,r14}
 	BGE	sr_loop
-sr_less_than_8:
+sr_less_than_8
 	ADDS	r0, r0, #8
 	BEQ	sr_end
-sr_loop2:
+sr_loop2
 	LDR	r3, [r1], #8
 	SUBS	r0, r0, #1
 	STR	r3, [r2], #4
 	BGT	sr_loop2
-sr_end:
+sr_end
 	LDMFD	r13!,{r4-r11,PC}
 
-mdct_backward_arm_low:
-	@ r0 = n
-	@ r1 = in
+mdct_backward_arm_low
+	; r0 = n
+	; r1 = in
 	STMFD	r13!,{r4-r11,r14}
 
-	MOV	r2, #1<<4	@ r2 = 1<<shift
-	MOV	r3, #13-4	@ r3 = 13-shift
-find_shift_loop:
-	TST	r0, r2		@ if (n & (1<<shift)) == 0
+	MOV	r2, #1<<4	; r2 = 1<<shift
+	MOV	r3, #13-4	; r3 = 13-shift
+find_shift_loop
+	TST	r0, r2		; if (n & (1<<shift)) == 0
 	MOV	r2, r2, LSL #1
-	SUBEQ	r3, r3, #1	@ shift--
+	SUBEQ	r3, r3, #1	; shift--
 	BEQ	find_shift_loop
 	MOV	r2, #2
-	MOV	r2, r2, LSL r3	@ r2 = step = 2<<shift
-
-	@ presymmetry
-	@ r0 = n (a multiple of 4)
-	@ r1 = in
-	@ r2 = step
-	@ r3 = shift
-
-	ADD	r4, r1, r0, LSL #1	@ r4 = aX = in+(n>>1)
-	ADD	r14,r1, r0		@ r14= in+(n>>2)
-	SUB	r4, r4, #3*4		@ r4 = aX = in+n2-3
-	LDR	r5, =sincos_lookup0	@ r5 = T=sincos_lookup0
-
-presymmetry_loop1:
-	LDR	r7, [r4,#8]		@ r6 = s2 = aX[2]
-	LDRB	r11,[r5,#1]		@ r11= T[1]
-	LDR	r6, [r4],#-16		@ r6 = s0 = aX[0]
-	LDRB	r10,[r5],r2		@ r10= T[0]   T += step
+	MOV	r2, r2, LSL r3	; r2 = step = 2<<shift
+
+	; presymmetry
+	; r0 = n (a multiple of 4)
+	; r1 = in
+	; r2 = step
+	; r3 = shift
+
+	ADD	r4, r1, r0, LSL #1	; r4 = aX = in+(n>>1)
+	ADD	r14,r1, r0		; r14= in+(n>>2)
+	SUB	r4, r4, #3*4		; r4 = aX = in+n2-3
+	LDR	r5, =sincos_lookup0	; r5 = T=sincos_lookup0
+
+presymmetry_loop1
+	LDR	r7, [r4,#8]		; r6 = s2 = aX[2]
+	LDRB	r11,[r5,#1]		; r11= T[1]
+	LDR	r6, [r4],#-16		; r6 = s0 = aX[0]
+	LDRB	r10,[r5],r2		; r10= T[0]   T += step
 	MOV	r6, r6, ASR #8
 	MOV	r7, r7, ASR #8
 
-	@ XPROD31(s0, s2, T[0], T[1], 0xaX[0], &ax[2])
-	MUL	r9, r6, r10		@ r9   = s0*T[0]
+	; XPROD31(s0, s2, T[0], T[1], &aX[0], &ax[2])
+	MUL	r9, r6, r10		; r9   = s0*T[0]
 	RSB	r6, r6, #0
-	MLA	r9, r7, r11,r9		@ r9  += s2*T[1]
+	MLA	r9, r7, r11,r9		; r9  += s2*T[1]
 	CMP	r4, r14
-	MUL	r12,r7, r10		@ r12  = s2*T[0]
-	STR	r9, [r4,#16]		@ aX[0] = r9
-	MLA	r12,r6, r11,r12		@ r12 -= s0*T[1]
-	STR	r12,[r4,#8+16]		@ aX[2] = r12
-
-	BGE	presymmetry_loop1	@ while (aX >= in+n4)
-
-presymmetry_loop2:
-	LDR	r6, [r4],#-16		@ r6 = s0 = aX[0]
-	LDRB	r10,[r5,#1]		@ r10= T[1]
-	LDR	r7, [r4,#16+8]		@ r6 = s2 = aX[2]
-	LDRB	r11,[r5],-r2		@ r11= T[0]   T -= step
+	MUL	r12,r7, r10		; r12  = s2*T[0]
+	STR	r9, [r4,#16]		; aX[0] = r9
+	MLA	r12,r6, r11,r12		; r12 -= s0*T[1]
+	STR	r12,[r4,#8+16]		; aX[2] = r12
+
+	BGE	presymmetry_loop1	; while (aX >= in+n4)
+
+presymmetry_loop2
+	LDR	r6, [r4],#-16		; r6 = s0 = aX[0]
+	LDRB	r10,[r5,#1]		; r10= T[1]
+	LDR	r7, [r4,#16+8]		; r6 = s2 = aX[2]
+	LDRB	r11,[r5],-r2		; r11= T[0]   T -= step
 	MOV	r6, r6, ASR #8
 	MOV	r7, r7, ASR #8
 
-	@ XPROD31(s0, s2, T[1], T[0], 0xaX[0], &ax[2])
-	MUL	r9, r6, r10		@ r9   = s0*T[1]
+	; XPROD31(s0, s2, T[1], T[0], &aX[0], &ax[2])
+	MUL	r9, r6, r10		; r9   = s0*T[1]
 	RSB	r6, r6, #0
-	MLA	r9, r7, r11,r9		@ r9  += s2*T[0]
+	MLA	r9, r7, r11,r9		; r9  += s2*T[0]
 	CMP	r4, r1
-	MUL	r12,r7, r10		@ r12  = s2*T[1]
-	STR	r9, [r4,#16]		@ aX[0] = r9
-	MLA	r12,r6, r11,r12		@ r12 -= s0*T[0]
-	STR	r12,[r4,#8+16]		@ aX[2] = r12
+	MUL	r12,r7, r10		; r12  = s2*T[1]
+	STR	r9, [r4,#16]		; aX[0] = r9
+	MLA	r12,r6, r11,r12		; r12 -= s0*T[0]
+	STR	r12,[r4,#8+16]		; aX[2] = r12
 
-	BGE	presymmetry_loop2	@ while (aX >= in)
+	BGE	presymmetry_loop2	; while (aX >= in)
 
-	@ r0 = n
-	@ r1 = in
-	@ r2 = step
-	@ r3 = shift
+	; r0 = n
+	; r1 = in
+	; r2 = step
+	; r3 = shift
 	STMFD	r13!,{r3}
-	LDR	r5, =sincos_lookup0	@ r5 = T=sincos_lookup0
-	ADD	r4, r1, r0, LSL #1	@ r4 = aX = in+(n>>1)
-	SUB	r4, r4, #4*4		@ r4 = aX = in+(n>>1)-4
-	LDRB	r11,[r5,#1]		@ r11= T[1]
-	LDRB	r10,[r5],r2		@ r10= T[0]    T += step
-presymmetry_loop3:
-	LDR	r8, [r1],#16 		@ r8 = ro0 = bX[0]
-	LDR	r9, [r1,#8-16]		@ r9 = ro2 = bX[2]
-	LDR	r6, [r4],#-16		@ r6 = ri0 = aX[0]
-	LDR	r7, [r4,#8+16]		@ r7 = ri2 = aX[2]
+	LDR	r5, =sincos_lookup0	; r5 = T=sincos_lookup0
+	ADD	r4, r1, r0, LSL #1	; r4 = aX = in+(n>>1)
+	SUB	r4, r4, #4*4		; r4 = aX = in+(n>>1)-4
+	LDRB	r11,[r5,#1]		; r11= T[1]
+	LDRB	r10,[r5],r2		; r10= T[0]    T += step
+presymmetry_loop3
+	LDR	r8, [r1],#16 		; r8 = ro0 = bX[0]
+	LDR	r9, [r1,#8-16]		; r9 = ro2 = bX[2]
+	LDR	r6, [r4],#-16		; r6 = ri0 = aX[0]
+	LDR	r7, [r4,#8+16]		; r7 = ri2 = aX[2]
 	MOV	r8, r8, ASR #8
 	MOV	r9, r9, ASR #8
 	MOV	r6, r6, ASR #8
 
-	@ XNPROD31( ro2, ro0, T[1], T[0], 0xaX[0], &aX[2] )
-	@ aX[0] = (ro2*T[1] - ro0*T[0])>>31 aX[2] = (ro0*T[1] + ro2*T[0])>>31
-	MUL	r12,r8, r11		@ r12  = ro0*T[1]
+	; XNPROD31( ro2, ro0, T[1], T[0], &aX[0], &aX[2] )
+	; aX[0] = (ro2*T[1] - ro0*T[0])>>31 aX[2] = (ro0*T[1] + ro2*T[0])>>31
+	MUL	r12,r8, r11		; r12  = ro0*T[1]
 	MOV	r7, r7, ASR #8
-	MLA	r12,r9, r10,r12		@ r12 += ro2*T[0]
-	RSB	r8, r8, #0		@ r8 = -ro0
-	MUL	r3, r9, r11		@ r3   = ro2*T[1]
-	LDRB	r11,[r5,#1]		@ r11= T[1]
-	MLA	r3, r8, r10,r3		@ r3  -= ro0*T[0]
-	LDRB	r10,[r5],r2		@ r10= T[0]    T += step
+	MLA	r12,r9, r10,r12		; r12 += ro2*T[0]
+	RSB	r8, r8, #0		; r8 = -ro0
+	MUL	r3, r9, r11		; r3   = ro2*T[1]
+	LDRB	r11,[r5,#1]		; r11= T[1]
+	MLA	r3, r8, r10,r3		; r3  -= ro0*T[0]
+	LDRB	r10,[r5],r2		; r10= T[0]    T += step
 	STR	r12,[r4,#16+8]
 	STR	r3, [r4,#16]
 
-	@ XNPROD31( ri2, ri0, T[0], T[1], 0xbX[0], &bX[2] )
-	@ bX[0] = (ri2*T[0] - ri0*T[1])>>31 bX[2] = (ri0*T[0] + ri2*T[1])>>31
-	MUL	r12,r6, r10		@ r12  = ri0*T[0]
-	RSB	r6, r6, #0		@ r6 = -ri0
-	MLA	r12,r7, r11,r12		@ r12 += ri2*T[1]
+	; XNPROD31( ri2, ri0, T[0], T[1], &bX[0], &bX[2] )
+	; bX[0] = (ri2*T[0] - ri0*T[1])>>31 bX[2] = (ri0*T[0] + ri2*T[1])>>31
+	MUL	r12,r6, r10		; r12  = ri0*T[0]
+	RSB	r6, r6, #0		; r6 = -ri0
+	MLA	r12,r7, r11,r12		; r12 += ri2*T[1]
 	CMP	r4, r1
-	MUL	r3, r7, r10		@ r3   = ri2*T[0]
+	MUL	r3, r7, r10		; r3   = ri2*T[0]
 	STR	r12,[r1,#8-16]
-	MLA	r3, r6, r11,r3		@ r3  -= ri0*T[1]
+	MLA	r3, r6, r11,r3		; r3  -= ri0*T[1]
 	STR	r3, [r1,#-16]
 
 	BGE	presymmetry_loop3
 
-	SUB	r1,r1,r0		@ r1 = in -= n>>2 (i.e. restore in)
+	SUB	r1,r1,r0		; r1 = in -= n>>2 (i.e. restore in)
 
 	LDR	r3,[r13]
 	STR	r2,[r13,#-4]!
 
-	@ mdct_butterflies
-	@ r0 = n  = (points * 2)
-	@ r1 = in = x
-	@ r2 = i
-	@ r3 = shift
+	; mdct_butterflies
+	; r0 = n  = (points * 2)
+	; r1 = in = x
+	; r2 = i
+	; r3 = shift
 	STMFD	r13!,{r0-r1}
-	RSBS	r4,r3,#6		@ r4 = stages = 7-shift then --stages
+	RSBS	r4,r3,#6		; r4 = stages = 7-shift then --stages
 	LDR	r5,=sincos_lookup0
 	BLE	no_generics
-	MOV	r14,#4			@ r14= 4               (i=0)
-	MOV	r6, r14,LSL r3		@ r6 = (4<<i)<<shift
-mdct_butterflies_loop1:
-	MOV	r0, r0, LSR #1		@ r0 = points>>i = POINTS
-	MOV	r2, r14,LSR #2		@ r2 = (1<<i)-j        (j=0)
+	MOV	r14,#4			; r14= 4               (i=0)
+	MOV	r6, r14,LSL r3		; r6 = (4<<i)<<shift
+mdct_butterflies_loop1
+	MOV	r0, r0, LSR #1		; r0 = points>>i = POINTS
+	MOV	r2, r14,LSR #2		; r2 = (1<<i)-j        (j=0)
 	STMFD	r13!,{r4,r14}
-mdct_butterflies_loop2:
-
-	@ mdct_butterfly_generic(x+POINTS*j, POINTS, 4<<(i+shift))
-	@ mdct_butterfly_generic(r1, r0, r6)
-	@ r0 = points
-	@ r1 = x
-	@ preserve r2 (external loop counter)
-	@ preserve r3
-	@ preserve r4 (external loop counter)
-	@ r5 = T = sincos_lookup0
-	@ r6 = step
-	@ preserve r14
-
-	STR	r2,[r13,#-4]!		@ stack r2
-	ADD	r1,r1,r0,LSL #1		@ r1 = x2+4 = x + (POINTS>>1)
-	ADD	r7,r1,r0,LSL #1		@ r7 = x1+4 = x + POINTS
-	ADD	r12,r5,#1024		@ r12= sincos_lookup0+1024
-
-mdct_bufferfly_generic_loop1:
-	LDMDB	r7!,{r2,r3,r8,r11}	@ r2 = x1[0]
-					@ r3 = x1[1]
-					@ r8 = x1[2]
-					@ r11= x1[3]    x1 -= 4
-	LDMDB	r1!,{r4,r9,r10,r14}	@ r4 = x2[0]
-					@ r9 = x2[1]
-					@ r10= x2[2]
-					@ r14= x2[3]    x2 -= 4
-
-	SUB	r2, r2, r3		@ r2 = s0 = x1[0] - x1[1]
-	ADD	r3, r2, r3, LSL #1	@ r3 =      x1[0] + x1[1] (-> x1[0])
-	SUB	r11,r11,r8		@ r11= s1 = x1[3] - x1[2]
-	ADD	r8, r11,r8, LSL #1	@ r8 =      x1[3] + x1[2] (-> x1[2])
-	SUB	r9, r9, r4		@ r9 = s2 = x2[1] - x2[0]
-	ADD	r4, r9, r4, LSL #1	@ r4 =      x2[1] + x2[0] (-> x1[1])
-	SUB	r14,r14,r10		@ r14= s3 = x2[3] - x2[2]
-	ADD	r10,r14,r10,LSL #1	@ r10=      x2[3] + x2[2] (-> x1[3])
+mdct_butterflies_loop2
+
+	; mdct_butterfly_generic(x+POINTS*j, POINTS, 4<<(i+shift))
+	; mdct_butterfly_generic(r1, r0, r6)
+	; r0 = points
+	; r1 = x
+	; preserve r2 (external loop counter)
+	; preserve r3
+	; preserve r4 (external loop counter)
+	; r5 = T = sincos_lookup0
+	; r6 = step
+	; preserve r14
+
+	STR	r2,[r13,#-4]!		; stack r2
+	ADD	r1,r1,r0,LSL #1		; r1 = x2+4 = x + (POINTS>>1)
+	ADD	r7,r1,r0,LSL #1		; r7 = x1+4 = x + POINTS
+	ADD	r12,r5,#1024		; r12= sincos_lookup0+1024
+
+mdct_bufferfly_generic_loop1
+	LDMDB	r7!,{r2,r3,r8,r11}	; r2 = x1[0]
+					; r3 = x1[1]
+					; r8 = x1[2]
+					; r11= x1[3]    x1 -= 4
+	LDMDB	r1!,{r4,r9,r10,r14}	; r4 = x2[0]
+					; r9 = x2[1]
+					; r10= x2[2]
+					; r14= x2[3]    x2 -= 4
+
+	SUB	r2, r2, r3		; r2 = s0 = x1[0] - x1[1]
+	ADD	r3, r2, r3, LSL #1	; r3 =      x1[0] + x1[1] (-> x1[0])
+	SUB	r11,r11,r8		; r11= s1 = x1[3] - x1[2]
+	ADD	r8, r11,r8, LSL #1	; r8 =      x1[3] + x1[2] (-> x1[2])
+	SUB	r9, r9, r4		; r9 = s2 = x2[1] - x2[0]
+	ADD	r4, r9, r4, LSL #1	; r4 =      x2[1] + x2[0] (-> x1[1])
+	SUB	r14,r14,r10		; r14= s3 = x2[3] - x2[2]
+	ADD	r10,r14,r10,LSL #1	; r10=      x2[3] + x2[2] (-> x1[3])
 	STMIA	r7,{r3,r4,r8,r10}
 
-	@ r0 = points
-	@ r1 = x2
-	@ r2 = s0
-	@ r3 free
-	@ r4 free
-	@ r5 = T
-	@ r6 = step
-	@ r7 = x1
-	@ r8 free
-	@ r9 = s2
-	@ r10 free
-	@ r11= s1
-	@ r12= limit
-	@ r14= s3
-
-	LDRB	r8, [r5,#1]		@ r8 = T[1]
-	LDRB	r10,[r5],r6		@ r10= T[0]		T += step
+	; r0 = points
+	; r1 = x2
+	; r2 = s0
+	; r3 free
+	; r4 free
+	; r5 = T
+	; r6 = step
+	; r7 = x1
+	; r8 free
+	; r9 = s2
+	; r10 free
+	; r11= s1
+	; r12= limit
+	; r14= s3
+
+	LDRB	r8, [r5,#1]		; r8 = T[1]
+	LDRB	r10,[r5],r6		; r10= T[0]		T += step
 	MOV	r2, r2, ASR #8
 	MOV	r11,r11,ASR #8
 	MOV	r9, r9, ASR #8
 	MOV	r14,r14,ASR #8
 
-	@ XPROD31(s1, s0, T[0], T[1], &x2[0], &x2[2])
-	@ x2[0] = (s1*T[0] + s0*T[1])>>31     x2[2] = (s0*T[0] - s1*T[1])>>31
-	@ stall Xscale
-	MUL	r3, r2, r8		@ r3   = s0*T[1]
-	MLA	r3, r11,r10,r3		@ r3  += s1*T[0]
+	; XPROD31(s1, s0, T[0], T[1], &x2[0], &x2[2])
+	; x2[0] = (s1*T[0] + s0*T[1])>>31     x2[2] = (s0*T[0] - s1*T[1])>>31
+	; stall Xscale
+	MUL	r3, r2, r8		; r3   = s0*T[1]
+	MLA	r3, r11,r10,r3		; r3  += s1*T[0]
 	RSB	r11,r11,#0
-	MUL	r4, r8, r11		@ r4   = -s1*T[1]
-	MLA	r4, r2, r10,r4		@ r4  += s0*T[0] = Value for x2[2]
-	MOV	r2, r3			@ r2 = r3 = Value for x2[0]
-
-	@ XPROD31(s2, s3, T[0], T[1], &x2[1], &x2[3])
-	@ x2[1] = (s2*T[0] + s3*T[1])>>31     x2[3] = (s3*T[0] - s2*T[1])>>31
-	MUL	r3, r9, r10		@ r3   = s2*T[0]
-	MLA	r3, r14,r8, r3		@ r3  += s3*T[1] = Value for x2[1]
+	MUL	r4, r8, r11		; r4   = -s1*T[1]
+	MLA	r4, r2, r10,r4		; r4  += s0*T[0] = Value for x2[2]
+	MOV	r2, r3			; r2 = r3 = Value for x2[0]
+
+	; XPROD31(s2, s3, T[0], T[1], &x2[1], &x2[3])
+	; x2[1] = (s2*T[0] + s3*T[1])>>31     x2[3] = (s3*T[0] - s2*T[1])>>31
+	MUL	r3, r9, r10		; r3   = s2*T[0]
+	MLA	r3, r14,r8, r3		; r3  += s3*T[1] = Value for x2[1]
 	RSB	r9, r9, #0
-	MUL	r11,r14,r10		@ r11  = s3*T[0]
-	MLA	r11,r9, r8, r11		@ r11 -= s2*T[1] = Value for x2[3]
+	MUL	r11,r14,r10		; r11  = s3*T[0]
+	MLA	r11,r9, r8, r11		; r11 -= s2*T[1] = Value for x2[3]
 	CMP	r5, r12
 
 	STMIA	r1,{r2,r3,r4,r11}
@@ -473,472 +473,472 @@ mdct_bufferfly_generic_loop1:
 	BLT	mdct_bufferfly_generic_loop1
 
 	SUB	r12,r12,#1024
-mdct_bufferfly_generic_loop2:
-	LDMDB	r7!,{r2,r3,r9,r10}	@ r2 = x1[0]
-					@ r3 = x1[1]
-					@ r9 = x1[2]
-					@ r10= x1[3]    x1 -= 4
-	LDMDB	r1!,{r4,r8,r11,r14}	@ r4 = x2[0]
-					@ r8 = x2[1]
-					@ r11= x2[2]
-					@ r14= x2[3]    x2 -= 4
-
-	SUB	r2, r2, r3		@ r2 = s0 = x1[0] - x1[1]
-	ADD	r3, r2, r3, LSL #1	@ r3 =      x1[0] + x1[1] (-> x1[0])
-	SUB	r9, r9,r10		@ r9 = s1 = x1[2] - x1[3]
-	ADD	r10,r9,r10, LSL #1	@ r10=      x1[2] + x1[3] (-> x1[2])
-	SUB	r4, r4, r8		@ r4 = s2 = x2[0] - x2[1]
-	ADD	r8, r4, r8, LSL #1	@ r8 =      x2[0] + x2[1] (-> x1[1])
-	SUB	r14,r14,r11		@ r14= s3 = x2[3] - x2[2]
-	ADD	r11,r14,r11,LSL #1	@ r11=      x2[3] + x2[2] (-> x1[3])
+mdct_bufferfly_generic_loop2
+	LDMDB	r7!,{r2,r3,r9,r10}	; r2 = x1[0]
+					; r3 = x1[1]
+					; r9 = x1[2]
+					; r10= x1[3]    x1 -= 4
+	LDMDB	r1!,{r4,r8,r11,r14}	; r4 = x2[0]
+					; r8 = x2[1]
+					; r11= x2[2]
+					; r14= x2[3]    x2 -= 4
+
+	SUB	r2, r2, r3		; r2 = s0 = x1[0] - x1[1]
+	ADD	r3, r2, r3, LSL #1	; r3 =      x1[0] + x1[1] (-> x1[0])
+	SUB	r9, r9,r10		; r9 = s1 = x1[2] - x1[3]
+	ADD	r10,r9,r10, LSL #1	; r10=      x1[2] + x1[3] (-> x1[2])
+	SUB	r4, r4, r8		; r4 = s2 = x2[0] - x2[1]
+	ADD	r8, r4, r8, LSL #1	; r8 =      x2[0] + x2[1] (-> x1[1])
+	SUB	r14,r14,r11		; r14= s3 = x2[3] - x2[2]
+	ADD	r11,r14,r11,LSL #1	; r11=      x2[3] + x2[2] (-> x1[3])
 	STMIA	r7,{r3,r8,r10,r11}
 
-	@ r0 = points
-	@ r1 = x2
-	@ r2 = s0
-	@ r3 free
-	@ r4 = s2
-	@ r5 = T
-	@ r6 = step
-	@ r7 = x1
-	@ r8 free
-	@ r9 = s1
-	@ r10 free
-	@ r11 free
-	@ r12= limit
-	@ r14= s3
-
-	LDRB	r8, [r5,#1]		@ r8 = T[1]
-	LDRB	r10,[r5],-r6		@ r10= T[0]		T -= step
+	; r0 = points
+	; r1 = x2
+	; r2 = s0
+	; r3 free
+	; r4 = s2
+	; r5 = T
+	; r6 = step
+	; r7 = x1
+	; r8 free
+	; r9 = s1
+	; r10 free
+	; r11 free
+	; r12= limit
+	; r14= s3
+
+	LDRB	r8, [r5,#1]		; r8 = T[1]
+	LDRB	r10,[r5],-r6		; r10= T[0]		T -= step
 	MOV	r2, r2, ASR #8
 	MOV	r9, r9, ASR #8
 	MOV	r4, r4, ASR #8
 	MOV	r14,r14,ASR #8
 
-	@ XNPROD31(s0, s1, T[0], T[1], &x2[0], &x2[2])
-	@ x2[0] = (s0*T[0] - s1*T[1])>>31     x2[2] = (s1*T[0] + s0*T[1])>>31
-	@ stall Xscale
-	MUL	r11,r2, r8		@ r11  = s0*T[1]
-	MLA	r11,r9, r10,r11		@ r11 += s1*T[0]
+	; XNPROD31(s0, s1, T[0], T[1], &x2[0], &x2[2])
+	; x2[0] = (s0*T[0] - s1*T[1])>>31     x2[2] = (s1*T[0] + s0*T[1])>>31
+	; stall Xscale
+	MUL	r11,r2, r8		; r11  = s0*T[1]
+	MLA	r11,r9, r10,r11		; r11 += s1*T[0]
 	RSB	r9, r9, #0
-	MUL	r2, r10,r2		@ r2   = s0*T[0]
-	MLA	r2, r9, r8, r2		@ r2  += -s1*T[1] = Value for x2[0]
-	MOV	r9, r11			@ r9 = r11 = Value for x2[2]
-
-	@ XNPROD31(s3, s2, T[0], T[1], &x2[1], &x2[3])
-	@ x2[1] = (s3*T[0] - s2*T[1])>>31     x2[3] = (s2*T[0] + s3*T[1])>>31
-	MUL	r11,r4, r10		@ r11   = s2*T[0]
-	MLA	r11,r14,r8, r11		@ r11  += s3*T[1] = Value for x2[3]
+	MUL	r2, r10,r2		; r2   = s0*T[0]
+	MLA	r2, r9, r8, r2		; r2  += -s1*T[1] = Value for x2[0]
+	MOV	r9, r11			; r9 = r11 = Value for x2[2]
+
+	; XNPROD31(s3, s2, T[0], T[1], &x2[1], &x2[3])
+	; x2[1] = (s3*T[0] - s2*T[1])>>31     x2[3] = (s2*T[0] + s3*T[1])>>31
+	MUL	r11,r4, r10		; r11   = s2*T[0]
+	MLA	r11,r14,r8, r11		; r11  += s3*T[1] = Value for x2[3]
 	RSB	r4, r4, #0
-	MUL	r3, r14,r10		@ r3   = s3*T[0]
-	MLA	r3, r4, r8, r3		@ r3  -= s2*T[1] = Value for x2[1]
+	MUL	r3, r14,r10		; r3   = s3*T[0]
+	MLA	r3, r4, r8, r3		; r3  -= s2*T[1] = Value for x2[1]
 	CMP	r5, r12
 
 	STMIA	r1,{r2,r3,r9,r11}
 
 	BGT	mdct_bufferfly_generic_loop2
 
-	LDR	r2,[r13],#4		@ unstack r2
-	ADD	r1, r1, r0, LSL #2	@ r1 = x+POINTS*j
-	@ stall Xscale
-	SUBS	r2, r2, #1		@ r2--                 (j++)
+	LDR	r2,[r13],#4		; unstack r2
+	ADD	r1, r1, r0, LSL #2	; r1 = x+POINTS*j
+	; stall Xscale
+	SUBS	r2, r2, #1		; r2--                 (j++)
 	BGT	mdct_butterflies_loop2
 
 	LDMFD	r13!,{r4,r14}
 
 	LDR	r1,[r13,#4]
 
-	SUBS	r4, r4, #1		@ stages--
-	MOV	r14,r14,LSL #1		@ r14= 4<<i            (i++)
-	MOV	r6, r6, LSL #1		@ r6 = step <<= 1      (i++)
+	SUBS	r4, r4, #1		; stages--
+	MOV	r14,r14,LSL #1		; r14= 4<<i            (i++)
+	MOV	r6, r6, LSL #1		; r6 = step <<= 1      (i++)
 	BGE	mdct_butterflies_loop1
 	LDMFD	r13,{r0-r1}
 
-no_generics:
-	@ mdct_butterflies part2 (loop around mdct_bufferfly_32)
-	@ r0 = points
-	@ r1 = in
-	@ r2 = step
-	@ r3 = shift
-
-mdct_bufferflies_loop3:
-	@ mdct_bufferfly_32
-
-	@ block1
-	ADD	r4, r1, #16*4		@ r4 = &in[16]
-	LDMIA	r4,{r5,r6,r9,r10}	@ r5 = x[16]
-					@ r6 = x[17]
-					@ r9 = x[18]
-					@ r10= x[19]
-	LDMIA	r1,{r7,r8,r11,r12}	@ r7 = x[0]
-					@ r8 = x[1]
-					@ r11= x[2]
-					@ r12= x[3]
-	SUB	r5, r5, r6		@ r5 = s0 = x[16] - x[17]
-	ADD	r6, r5, r6, LSL #1	@ r6 =      x[16] + x[17]  -> x[16]
-	SUB	r9, r9, r10		@ r9 = s1 = x[18] - x[19]
-	ADD	r10,r9, r10,LSL #1	@ r10=      x[18] + x[19]  -> x[18]
-	SUB	r8, r8, r7		@ r8 = s2 = x[ 1] - x[ 0]
-	ADD	r7, r8, r7, LSL #1	@ r7 =      x[ 1] + x[ 0]  -> x[17]
-	SUB	r12,r12,r11		@ r12= s3 = x[ 3] - x[ 2]
-	ADD	r11,r12,r11, LSL #1	@ r11=      x[ 3] + x[ 2]  -> x[19]
+no_generics
+	; mdct_butterflies part2 (loop around mdct_bufferfly_32)
+	; r0 = points
+	; r1 = in
+	; r2 = step
+	; r3 = shift
+
+mdct_bufferflies_loop3
+	; mdct_bufferfly_32
+
+	; block1
+	ADD	r4, r1, #16*4		; r4 = &in[16]
+	LDMIA	r4,{r5,r6,r9,r10}	; r5 = x[16]
+					; r6 = x[17]
+					; r9 = x[18]
+					; r10= x[19]
+	LDMIA	r1,{r7,r8,r11,r12}	; r7 = x[0]
+					; r8 = x[1]
+					; r11= x[2]
+					; r12= x[3]
+	SUB	r5, r5, r6		; r5 = s0 = x[16] - x[17]
+	ADD	r6, r5, r6, LSL #1	; r6 =      x[16] + x[17]  -> x[16]
+	SUB	r9, r9, r10		; r9 = s1 = x[18] - x[19]
+	ADD	r10,r9, r10,LSL #1	; r10=      x[18] + x[19]  -> x[18]
+	SUB	r8, r8, r7		; r8 = s2 = x[ 1] - x[ 0]
+	ADD	r7, r8, r7, LSL #1	; r7 =      x[ 1] + x[ 0]  -> x[17]
+	SUB	r12,r12,r11		; r12= s3 = x[ 3] - x[ 2]
+	ADD	r11,r12,r11, LSL #1	; r11=      x[ 3] + x[ 2]  -> x[19]
 	STMIA	r4!,{r6,r7,r10,r11}
 
-	MOV	r6,#0xed		@ r6 =cPI1_8
-	MOV	r7,#0x62		@ r7 =cPI3_8
+	MOV	r6,#0xed		; r6 =cPI1_8
+	MOV	r7,#0x62		; r7 =cPI3_8
 
 	MOV	r5, r5, ASR #8
 	MOV	r9, r9, ASR #8
 	MOV	r8, r8, ASR #8
 	MOV	r12,r12,ASR #8
 
-	@ XNPROD31( s0, s1, cPI3_8, cPI1_8, &x[ 0], &x[ 2] )
-	@ x[0] = s0*cPI3_8 - s1*cPI1_8     x[2] = s1*cPI3_8 + s0*cPI1_8
-	@ stall Xscale
-	MUL	r11,r5, r6		@ r11  = s0*cPI1_8
-	MLA	r11,r9, r7, r11		@ r11 += s1*cPI3_8
+	; XNPROD31( s0, s1, cPI3_8, cPI1_8, &x[ 0], &x[ 2] )
+	; x[0] = s0*cPI3_8 - s1*cPI1_8     x[2] = s1*cPI3_8 + s0*cPI1_8
+	; stall Xscale
+	MUL	r11,r5, r6		; r11  = s0*cPI1_8
+	MLA	r11,r9, r7, r11		; r11 += s1*cPI3_8
 	RSB	r9, r9, #0
-	MUL	r5, r7, r5		@ r5   = s0*cPI3_8
-	MLA	r5, r9, r6, r5		@ r5  -= s1*cPI1_8
+	MUL	r5, r7, r5		; r5   = s0*cPI3_8
+	MLA	r5, r9, r6, r5		; r5  -= s1*cPI1_8
 
-	@ XPROD31 ( s2, s3, cPI1_8, cPI3_8, &x[ 1], &x[ 3] )
-	@ x[1] = s2*cPI1_8 + s3*cPI3_8     x[3] = s3*cPI1_8 - s2*cPI3_8
-	MUL	r9, r8, r6		@ r9   = s2*cPI1_8
-	MLA	r9, r12,r7, r9		@ r9  += s3*cPI3_8
+	; XPROD31 ( s2, s3, cPI1_8, cPI3_8, &x[ 1], &x[ 3] )
+	; x[1] = s2*cPI1_8 + s3*cPI3_8     x[3] = s3*cPI1_8 - s2*cPI3_8
+	MUL	r9, r8, r6		; r9   = s2*cPI1_8
+	MLA	r9, r12,r7, r9		; r9  += s3*cPI3_8
 	RSB	r8,r8,#0
-	MUL	r12,r6, r12		@ r12  = s3*cPI1_8
-	MLA	r12,r8, r7, r12		@ r12 -= s2*cPI3_8
+	MUL	r12,r6, r12		; r12  = s3*cPI1_8
+	MLA	r12,r8, r7, r12		; r12 -= s2*cPI3_8
 	STMIA	r1!,{r5,r9,r11,r12}
 
-	@ block2
-	LDMIA	r4,{r5,r6,r9,r10}	@ r5 = x[20]
-					@ r6 = x[21]
-					@ r9 = x[22]
-					@ r10= x[23]
-	LDMIA	r1,{r7,r8,r11,r12}	@ r7 = x[4]
-					@ r8 = x[5]
-					@ r11= x[6]
-					@ r12= x[7]
-	SUB	r5, r5, r6		@ r5 = s0 = x[20] - x[21]
-	ADD	r6, r5, r6, LSL #1	@ r6 =      x[20] + x[21]  -> x[20]
-	SUB	r9, r9, r10		@ r9 = s1 = x[22] - x[23]
-	ADD	r10,r9, r10,LSL #1	@ r10=      x[22] + x[23]  -> x[22]
-	SUB	r8, r8, r7		@ r8 = s2 = x[ 5] - x[ 4]
-	ADD	r7, r8, r7, LSL #1	@ r7 =      x[ 5] + x[ 4]  -> x[21]
-	SUB	r12,r12,r11		@ r12= s3 = x[ 7] - x[ 6]
-	ADD	r11,r12,r11, LSL #1	@ r11=      x[ 7] + x[ 6]  -> x[23]
-	MOV	r14,#0xb5		@ cPI2_8
+	; block2
+	LDMIA	r4,{r5,r6,r9,r10}	; r5 = x[20]
+					; r6 = x[21]
+					; r9 = x[22]
+					; r10= x[23]
+	LDMIA	r1,{r7,r8,r11,r12}	; r7 = x[4]
+					; r8 = x[5]
+					; r11= x[6]
+					; r12= x[7]
+	SUB	r5, r5, r6		; r5 = s0 = x[20] - x[21]
+	ADD	r6, r5, r6, LSL #1	; r6 =      x[20] + x[21]  -> x[20]
+	SUB	r9, r9, r10		; r9 = s1 = x[22] - x[23]
+	ADD	r10,r9, r10,LSL #1	; r10=      x[22] + x[23]  -> x[22]
+	SUB	r8, r8, r7		; r8 = s2 = x[ 5] - x[ 4]
+	ADD	r7, r8, r7, LSL #1	; r7 =      x[ 5] + x[ 4]  -> x[21]
+	SUB	r12,r12,r11		; r12= s3 = x[ 7] - x[ 6]
+	ADD	r11,r12,r11, LSL #1	; r11=      x[ 7] + x[ 6]  -> x[23]
+	MOV	r14,#0xb5		; cPI2_8
 	STMIA	r4!,{r6,r7,r10,r11}
 
-	SUB	r5, r5, r9		@ r5 = s0 - s1
-	ADD	r9, r5, r9, LSL #1	@ r9 = s0 + s1
+	SUB	r5, r5, r9		; r5 = s0 - s1
+	ADD	r9, r5, r9, LSL #1	; r9 = s0 + s1
 	MOV	r5, r5, ASR #8
-	MUL	r5, r14,r5		@ r5 = (s0-s1)*cPI2_8
-	SUB	r12,r12,r8		@ r12= s3 - s2
-	ADD	r8, r12,r8, LSL #1	@ r8 = s3 + s2
+	MUL	r5, r14,r5		; r5 = (s0-s1)*cPI2_8
+	SUB	r12,r12,r8		; r12= s3 - s2
+	ADD	r8, r12,r8, LSL #1	; r8 = s3 + s2
 
 	MOV	r8, r8, ASR #8
-	MUL	r8, r14,r8		@ r8  = (s3+s2)*cPI2_8
+	MUL	r8, r14,r8		; r8  = (s3+s2)*cPI2_8
 	MOV	r9, r9, ASR #8
-	MUL	r9, r14,r9		@ r9  = (s0+s1)*cPI2_8
+	MUL	r9, r14,r9		; r9  = (s0+s1)*cPI2_8
 	MOV	r12,r12,ASR #8
-	MUL	r12,r14,r12		@ r12 = (s3-s2)*cPI2_8
+	MUL	r12,r14,r12		; r12 = (s3-s2)*cPI2_8
 	STMIA	r1!,{r5,r8,r9,r12}
 
-	@ block3
-	LDMIA	r4,{r5,r6,r9,r10}	@ r5 = x[24]
-					@ r6 = x[25]
-					@ r9 = x[25]
-					@ r10= x[26]
-	LDMIA	r1,{r7,r8,r11,r12}	@ r7 = x[8]
-					@ r8 = x[9]
-					@ r11= x[10]
-					@ r12= x[11]
-	SUB	r5, r5, r6		@ r5 = s0 = x[24] - x[25]
-	ADD	r6, r5, r6, LSL #1	@ r6 =      x[24] + x[25]  -> x[25]
-	SUB	r9, r9, r10		@ r9 = s1 = x[26] - x[27]
-	ADD	r10,r9, r10,LSL #1	@ r10=      x[26] + x[27]  -> x[26]
-	SUB	r8, r8, r7		@ r8 = s2 = x[ 9] - x[ 8]
-	ADD	r7, r8, r7, LSL #1	@ r7 =      x[ 9] + x[ 8]  -> x[25]
-	SUB	r12,r12,r11		@ r12= s3 = x[11] - x[10]
-	ADD	r11,r12,r11, LSL #1	@ r11=      x[11] + x[10]  -> x[27]
+	; block3
+	LDMIA	r4,{r5,r6,r9,r10}	; r5 = x[24]
+					; r6 = x[25]
+					; r9 = x[25]
+					; r10= x[26]
+	LDMIA	r1,{r7,r8,r11,r12}	; r7 = x[8]
+					; r8 = x[9]
+					; r11= x[10]
+					; r12= x[11]
+	SUB	r5, r5, r6		; r5 = s0 = x[24] - x[25]
+	ADD	r6, r5, r6, LSL #1	; r6 =      x[24] + x[25]  -> x[25]
+	SUB	r9, r9, r10		; r9 = s1 = x[26] - x[27]
+	ADD	r10,r9, r10,LSL #1	; r10=      x[26] + x[27]  -> x[26]
+	SUB	r8, r8, r7		; r8 = s2 = x[ 9] - x[ 8]
+	ADD	r7, r8, r7, LSL #1	; r7 =      x[ 9] + x[ 8]  -> x[25]
+	SUB	r12,r12,r11		; r12= s3 = x[11] - x[10]
+	ADD	r11,r12,r11, LSL #1	; r11=      x[11] + x[10]  -> x[27]
 	STMIA	r4!,{r6,r7,r10,r11}
 
-	MOV	r6,#0x62		@ r6 = cPI3_8
-	MOV	r7,#0xED		@ r7 = cPI1_8
+	MOV	r6,#0x62		; r6 = cPI3_8
+	MOV	r7,#0xED		; r7 = cPI1_8
 
-	@ XNPROD31( s0, s1, cPI1_8, cPI3_8, &x[ 8], &x[10] )
-	@ x[8] = s0*cPI1_8 - s1*cPI3_8     x[10] = s1*cPI1_8 + s0*cPI3_8
-	@ stall Xscale
+	; XNPROD31( s0, s1, cPI1_8, cPI3_8, &x[ 8], &x[10] )
+	; x[8] = s0*cPI1_8 - s1*cPI3_8     x[10] = s1*cPI1_8 + s0*cPI3_8
+	; stall Xscale
 	MOV	r5, r5, ASR #8
-	MUL	r11,r5, r6		@ r11  = s0*cPI3_8
+	MUL	r11,r5, r6		; r11  = s0*cPI3_8
 	MOV	r9, r9, ASR #8
-	MLA	r11,r9, r7, r11		@ r11 += s1*cPI1_8
+	MLA	r11,r9, r7, r11		; r11 += s1*cPI1_8
 	RSB	r9, r9, #0
-	MUL	r5, r7, r5		@ r5   = s0*cPI1_8
-	MLA	r5, r9, r6, r5		@ r5  -= s1*cPI3_8
+	MUL	r5, r7, r5		; r5   = s0*cPI1_8
+	MLA	r5, r9, r6, r5		; r5  -= s1*cPI3_8
 
-	@ XPROD31 ( s2, s3, cPI3_8, cPI1_8, &x[ 9], &x[11] )
-	@ x[9] = s2*cPI3_8 + s3*cPI1_8     x[11] = s3*cPI3_8 - s2*cPI1_8
+	; XPROD31 ( s2, s3, cPI3_8, cPI1_8, &x[ 9], &x[11] )
+	; x[9] = s2*cPI3_8 + s3*cPI1_8     x[11] = s3*cPI3_8 - s2*cPI1_8
 	MOV	r8, r8, ASR #8
-	MUL	r9, r8, r6		@ r9   = s2*cPI3_8
+	MUL	r9, r8, r6		; r9   = s2*cPI3_8
 	MOV	r12,r12,ASR #8
-	MLA	r9, r12,r7, r9		@ r9  += s3*cPI1_8
+	MLA	r9, r12,r7, r9		; r9  += s3*cPI1_8
 	RSB	r8,r8,#0
-	MUL	r12,r6, r12		@ r12  = s3*cPI3_8
-	MLA	r12,r8, r7, r12		@ r12 -= s2*cPI1_8
+	MUL	r12,r6, r12		; r12  = s3*cPI3_8
+	MLA	r12,r8, r7, r12		; r12 -= s2*cPI1_8
 	STMIA	r1!,{r5,r9,r11,r12}
 
-	@ block4
-	LDMIA	r4,{r5,r6,r10,r11}	@ r5 = x[28]
-					@ r6 = x[29]
-					@ r10= x[30]
-					@ r11= x[31]
-	LDMIA	r1,{r8,r9,r12,r14}	@ r8 = x[12]
-					@ r9 = x[13]
-					@ r12= x[14]
-					@ r14= x[15]
-	SUB	r5, r5, r6		@ r5 = s0 = x[28] - x[29]
-	ADD	r6, r5, r6, LSL #1	@ r6 =      x[28] + x[29]  -> x[28]
-	SUB	r7, r14,r12		@ r7 = s3 = x[15] - x[14]
-	ADD	r12,r7, r12, LSL #1	@ r12=      x[15] + x[14]  -> x[31]
-	SUB	r10,r10,r11		@ r10= s1 = x[30] - x[31]
-	ADD	r11,r10,r11,LSL #1	@ r11=      x[30] + x[31]  -> x[30]
-	SUB	r14, r8, r9		@ r14= s2 = x[12] - x[13]
-	ADD	r9, r14, r9, LSL #1	@ r9 =      x[12] + x[13]  -> x[29]
+	; block4
+	LDMIA	r4,{r5,r6,r10,r11}	; r5 = x[28]
+					; r6 = x[29]
+					; r10= x[30]
+					; r11= x[31]
+	LDMIA	r1,{r8,r9,r12,r14}	; r8 = x[12]
+					; r9 = x[13]
+					; r12= x[14]
+					; r14= x[15]
+	SUB	r5, r5, r6		; r5 = s0 = x[28] - x[29]
+	ADD	r6, r5, r6, LSL #1	; r6 =      x[28] + x[29]  -> x[28]
+	SUB	r7, r14,r12		; r7 = s3 = x[15] - x[14]
+	ADD	r12,r7, r12, LSL #1	; r12=      x[15] + x[14]  -> x[31]
+	SUB	r10,r10,r11		; r10= s1 = x[30] - x[31]
+	ADD	r11,r10,r11,LSL #1	; r11=      x[30] + x[31]  -> x[30]
+	SUB	r14, r8, r9		; r14= s2 = x[12] - x[13]
+	ADD	r9, r14, r9, LSL #1	; r9 =      x[12] + x[13]  -> x[29]
 	STMIA	r4!,{r6,r9,r11,r12}
 	STMIA	r1!,{r5,r7,r10,r14}
 
-	@ mdct_butterfly16 (1st version)
-	@ block 1
+	; mdct_butterfly16 (1st version)
+	; block 1
 	SUB	r1,r1,#16*4
 	ADD	r4,r1,#8*4
-	LDMIA	r4,{r5,r6,r9,r10}	@ r5 = x[ 8]
-					@ r6 = x[ 9]
-					@ r9 = x[10]
-					@ r10= x[11]
-	LDMIA	r1,{r7,r8,r11,r12}	@ r7 = x[0]
-					@ r8 = x[1]
-					@ r11= x[2]
-					@ r12= x[3]
-	SUB	r5, r5, r6		@ r5 = s0 = x[ 8] - x[ 9]
-	ADD	r6, r5, r6, LSL #1	@ r6 =      x[ 8] + x[ 9]  -> x[ 8]
-	SUB	r9, r9, r10		@ r9 = s1 = x[10] - x[11]
-	ADD	r10,r9, r10,LSL #1	@ r10=      x[10] + x[11]  -> x[10]
-	SUB	r8, r8, r7		@ r8 = s2 = x[ 1] - x[ 0]
-	ADD	r7, r8, r7, LSL #1	@ r7 =      x[ 1] + x[ 0]  -> x[ 9]
-	SUB	r12,r12,r11		@ r12= s3 = x[ 3] - x[ 2]
-	ADD	r11,r12,r11, LSL #1	@ r11=      x[ 3] + x[ 2]  -> x[11]
-	MOV	r14,#0xB5		@ r14= cPI2_8
+	LDMIA	r4,{r5,r6,r9,r10}	; r5 = x[ 8]
+					; r6 = x[ 9]
+					; r9 = x[10]
+					; r10= x[11]
+	LDMIA	r1,{r7,r8,r11,r12}	; r7 = x[0]
+					; r8 = x[1]
+					; r11= x[2]
+					; r12= x[3]
+	SUB	r5, r5, r6		; r5 = s0 = x[ 8] - x[ 9]
+	ADD	r6, r5, r6, LSL #1	; r6 =      x[ 8] + x[ 9]  -> x[ 8]
+	SUB	r9, r9, r10		; r9 = s1 = x[10] - x[11]
+	ADD	r10,r9, r10,LSL #1	; r10=      x[10] + x[11]  -> x[10]
+	SUB	r8, r8, r7		; r8 = s2 = x[ 1] - x[ 0]
+	ADD	r7, r8, r7, LSL #1	; r7 =      x[ 1] + x[ 0]  -> x[ 9]
+	SUB	r12,r12,r11		; r12= s3 = x[ 3] - x[ 2]
+	ADD	r11,r12,r11, LSL #1	; r11=      x[ 3] + x[ 2]  -> x[11]
+	MOV	r14,#0xB5		; r14= cPI2_8
 	STMIA	r4!,{r6,r7,r10,r11}
 
-	SUB	r5, r5, r9		@ r5 = s0 - s1
-	ADD	r9, r5, r9, LSL #1	@ r9 = s0 + s1
+	SUB	r5, r5, r9		; r5 = s0 - s1
+	ADD	r9, r5, r9, LSL #1	; r9 = s0 + s1
 	MOV	r5, r5, ASR #8
-	MUL	r5, r14,r5		@ r5  = (s0-s1)*cPI2_8
-	SUB	r12,r12,r8		@ r12= s3 - s2
-	ADD	r8, r12,r8, LSL #1	@ r8 = s3 + s2
+	MUL	r5, r14,r5		; r5  = (s0-s1)*cPI2_8
+	SUB	r12,r12,r8		; r12= s3 - s2
+	ADD	r8, r12,r8, LSL #1	; r8 = s3 + s2
 
 	MOV	r8, r8, ASR #8
-	MUL	r8, r14,r8		@ r8  = (s3+s2)*cPI2_8
+	MUL	r8, r14,r8		; r8  = (s3+s2)*cPI2_8
 	MOV	r9, r9, ASR #8
-	MUL	r9, r14,r9		@ r9  = (s0+s1)*cPI2_8
+	MUL	r9, r14,r9		; r9  = (s0+s1)*cPI2_8
 	MOV	r12,r12,ASR #8
-	MUL	r12,r14,r12		@ r12 = (s3-s2)*cPI2_8
+	MUL	r12,r14,r12		; r12 = (s3-s2)*cPI2_8
 	STMIA	r1!,{r5,r8,r9,r12}
 
-	@ block2
-	LDMIA	r4,{r5,r6,r9,r10}	@ r5 = x[12]
-					@ r6 = x[13]
-					@ r9 = x[14]
-					@ r10= x[15]
-	LDMIA	r1,{r7,r8,r11,r12}	@ r7 = x[ 4]
-					@ r8 = x[ 5]
-					@ r11= x[ 6]
-					@ r12= x[ 7]
-	SUB	r14,r7, r8		@ r14= s0 = x[ 4] - x[ 5]
-	ADD	r8, r14,r8, LSL #1	@ r8 =      x[ 4] + x[ 5]  -> x[13]
-	SUB	r7, r12,r11		@ r7 = s1 = x[ 7] - x[ 6]
-	ADD	r11,r7, r11, LSL #1	@ r11=      x[ 7] + x[ 6]  -> x[15]
-	SUB	r5, r5, r6		@ r5 = s2 = x[12] - x[13]
-	ADD	r6, r5, r6, LSL #1	@ r6 =      x[12] + x[13]  -> x[12]
-	SUB	r12,r9, r10		@ r12= s3 = x[14] - x[15]
-	ADD	r10,r12,r10,LSL #1	@ r10=      x[14] + x[15]  -> x[14]
+	; block2
+	LDMIA	r4,{r5,r6,r9,r10}	; r5 = x[12]
+					; r6 = x[13]
+					; r9 = x[14]
+					; r10= x[15]
+	LDMIA	r1,{r7,r8,r11,r12}	; r7 = x[ 4]
+					; r8 = x[ 5]
+					; r11= x[ 6]
+					; r12= x[ 7]
+	SUB	r14,r7, r8		; r14= s0 = x[ 4] - x[ 5]
+	ADD	r8, r14,r8, LSL #1	; r8 =      x[ 4] + x[ 5]  -> x[13]
+	SUB	r7, r12,r11		; r7 = s1 = x[ 7] - x[ 6]
+	ADD	r11,r7, r11, LSL #1	; r11=      x[ 7] + x[ 6]  -> x[15]
+	SUB	r5, r5, r6		; r5 = s2 = x[12] - x[13]
+	ADD	r6, r5, r6, LSL #1	; r6 =      x[12] + x[13]  -> x[12]
+	SUB	r12,r9, r10		; r12= s3 = x[14] - x[15]
+	ADD	r10,r12,r10,LSL #1	; r10=      x[14] + x[15]  -> x[14]
 	STMIA	r4!,{r6,r8,r10,r11}
 	STMIA	r1!,{r5,r7,r12,r14}
 
-	@ mdct_butterfly_8
+	; mdct_butterfly_8
 	LDMDB	r1,{r6,r7,r8,r9,r10,r11,r12,r14}
-					@ r6 = x[0]
-					@ r7 = x[1]
-					@ r8 = x[2]
-					@ r9 = x[3]
-					@ r10= x[4]
-					@ r11= x[5]
-					@ r12= x[6]
-					@ r14= x[7]
-	ADD	r6, r6, r7		@ r6 = s0 = x[0] + x[1]
-	SUB	r7, r6, r7, LSL #1	@ r7 = s1 = x[0] - x[1]
-	ADD	r8, r8, r9		@ r8 = s2 = x[2] + x[3]
-	SUB	r9, r8, r9, LSL #1	@ r9 = s3 = x[2] - x[3]
-	ADD	r10,r10,r11		@ r10= s4 = x[4] + x[5]
-	SUB	r11,r10,r11,LSL #1	@ r11= s5 = x[4] - x[5]
-	ADD	r12,r12,r14		@ r12= s6 = x[6] + x[7]
-	SUB	r14,r12,r14,LSL #1	@ r14= s7 = x[6] - x[7]
-
-	ADD	r2, r11,r9		@ r2 = x[0] = s5 + s3
-	SUB	r4, r2, r9, LSL #1	@ r4 = x[2] = s5 - s3
-	SUB	r3, r14,r7		@ r3 = x[1] = s7 - s1
-	ADD	r5, r3, r7, LSL #1	@ r5 = x[3] = s7 + s1
-	SUB	r10,r10,r6		@ r10= x[4] = s4 - s0
-	SUB	r11,r12,r8		@ r11= x[5] = s6 - s2
-	ADD	r12,r10,r6, LSL #1	@ r12= x[6] = s4 + s0
-	ADD	r14,r11,r8, LSL #1	@ r14= x[7] = s6 + s2
+					; r6 = x[0]
+					; r7 = x[1]
+					; r8 = x[2]
+					; r9 = x[3]
+					; r10= x[4]
+					; r11= x[5]
+					; r12= x[6]
+					; r14= x[7]
+	ADD	r6, r6, r7		; r6 = s0 = x[0] + x[1]
+	SUB	r7, r6, r7, LSL #1	; r7 = s1 = x[0] - x[1]
+	ADD	r8, r8, r9		; r8 = s2 = x[2] + x[3]
+	SUB	r9, r8, r9, LSL #1	; r9 = s3 = x[2] - x[3]
+	ADD	r10,r10,r11		; r10= s4 = x[4] + x[5]
+	SUB	r11,r10,r11,LSL #1	; r11= s5 = x[4] - x[5]
+	ADD	r12,r12,r14		; r12= s6 = x[6] + x[7]
+	SUB	r14,r12,r14,LSL #1	; r14= s7 = x[6] - x[7]
+
+	ADD	r2, r11,r9		; r2 = x[0] = s5 + s3
+	SUB	r4, r2, r9, LSL #1	; r4 = x[2] = s5 - s3
+	SUB	r3, r14,r7		; r3 = x[1] = s7 - s1
+	ADD	r5, r3, r7, LSL #1	; r5 = x[3] = s7 + s1
+	SUB	r10,r10,r6		; r10= x[4] = s4 - s0
+	SUB	r11,r12,r8		; r11= x[5] = s6 - s2
+	ADD	r12,r10,r6, LSL #1	; r12= x[6] = s4 + s0
+	ADD	r14,r11,r8, LSL #1	; r14= x[7] = s6 + s2
 	STMDB	r1,{r2,r3,r4,r5,r10,r11,r12,r14}
 
-	@ mdct_butterfly_8
+	; mdct_butterfly_8
 	LDMIA	r1,{r6,r7,r8,r9,r10,r11,r12,r14}
-					@ r6 = x[0]
-					@ r7 = x[1]
-					@ r8 = x[2]
-					@ r9 = x[3]
-					@ r10= x[4]
-					@ r11= x[5]
-					@ r12= x[6]
-					@ r14= x[7]
-	ADD	r6, r6, r7		@ r6 = s0 = x[0] + x[1]
-	SUB	r7, r6, r7, LSL #1	@ r7 = s1 = x[0] - x[1]
-	ADD	r8, r8, r9		@ r8 = s2 = x[2] + x[3]
-	SUB	r9, r8, r9, LSL #1	@ r9 = s3 = x[2] - x[3]
-	ADD	r10,r10,r11		@ r10= s4 = x[4] + x[5]
-	SUB	r11,r10,r11,LSL #1	@ r11= s5 = x[4] - x[5]
-	ADD	r12,r12,r14		@ r12= s6 = x[6] + x[7]
-	SUB	r14,r12,r14,LSL #1	@ r14= s7 = x[6] - x[7]
-
-	ADD	r2, r11,r9		@ r2 = x[0] = s5 + s3
-	SUB	r4, r2, r9, LSL #1	@ r4 = x[2] = s5 - s3
-	SUB	r3, r14,r7		@ r3 = x[1] = s7 - s1
-	ADD	r5, r3, r7, LSL #1	@ r5 = x[3] = s7 + s1
-	SUB	r10,r10,r6		@ r10= x[4] = s4 - s0
-	SUB	r11,r12,r8		@ r11= x[5] = s6 - s2
-	ADD	r12,r10,r6, LSL #1	@ r12= x[6] = s4 + s0
-	ADD	r14,r11,r8, LSL #1	@ r14= x[7] = s6 + s2
+					; r6 = x[0]
+					; r7 = x[1]
+					; r8 = x[2]
+					; r9 = x[3]
+					; r10= x[4]
+					; r11= x[5]
+					; r12= x[6]
+					; r14= x[7]
+	ADD	r6, r6, r7		; r6 = s0 = x[0] + x[1]
+	SUB	r7, r6, r7, LSL #1	; r7 = s1 = x[0] - x[1]
+	ADD	r8, r8, r9		; r8 = s2 = x[2] + x[3]
+	SUB	r9, r8, r9, LSL #1	; r9 = s3 = x[2] - x[3]
+	ADD	r10,r10,r11		; r10= s4 = x[4] + x[5]
+	SUB	r11,r10,r11,LSL #1	; r11= s5 = x[4] - x[5]
+	ADD	r12,r12,r14		; r12= s6 = x[6] + x[7]
+	SUB	r14,r12,r14,LSL #1	; r14= s7 = x[6] - x[7]
+
+	ADD	r2, r11,r9		; r2 = x[0] = s5 + s3
+	SUB	r4, r2, r9, LSL #1	; r4 = x[2] = s5 - s3
+	SUB	r3, r14,r7		; r3 = x[1] = s7 - s1
+	ADD	r5, r3, r7, LSL #1	; r5 = x[3] = s7 + s1
+	SUB	r10,r10,r6		; r10= x[4] = s4 - s0
+	SUB	r11,r12,r8		; r11= x[5] = s6 - s2
+	ADD	r12,r10,r6, LSL #1	; r12= x[6] = s4 + s0
+	ADD	r14,r11,r8, LSL #1	; r14= x[7] = s6 + s2
 	STMIA	r1,{r2,r3,r4,r5,r10,r11,r12,r14}
 
-	@ mdct_butterfly16 (2nd version)
-	@ block 1
+	; mdct_butterfly16 (2nd version)
+	; block 1
 	ADD	r1,r1,#16*4-8*4
 	ADD	r4,r1,#8*4
-	LDMIA	r4,{r5,r6,r9,r10}	@ r5 = x[ 8]
-					@ r6 = x[ 9]
-					@ r9 = x[10]
-					@ r10= x[11]
-	LDMIA	r1,{r7,r8,r11,r12}	@ r7 = x[0]
-					@ r8 = x[1]
-					@ r11= x[2]
-					@ r12= x[3]
-	SUB	r5, r5, r6		@ r5 = s0 = x[ 8] - x[ 9]
-	ADD	r6, r5, r6, LSL #1	@ r6 =      x[ 8] + x[ 9]  -> x[ 8]
-	SUB	r9, r9, r10		@ r9 = s1 = x[10] - x[11]
-	ADD	r10,r9, r10,LSL #1	@ r10=      x[10] + x[11]  -> x[10]
-	SUB	r8, r8, r7		@ r8 = s2 = x[ 1] - x[ 0]
-	ADD	r7, r8, r7, LSL #1	@ r7 =      x[ 1] + x[ 0]  -> x[ 9]
-	SUB	r12,r12,r11		@ r12= s3 = x[ 3] - x[ 2]
-	ADD	r11,r12,r11, LSL #1	@ r11=      x[ 3] + x[ 2]  -> x[11]
-	MOV	r14,#0xb5		@ r14= cPI2_8
+	LDMIA	r4,{r5,r6,r9,r10}	; r5 = x[ 8]
+					; r6 = x[ 9]
+					; r9 = x[10]
+					; r10= x[11]
+	LDMIA	r1,{r7,r8,r11,r12}	; r7 = x[0]
+					; r8 = x[1]
+					; r11= x[2]
+					; r12= x[3]
+	SUB	r5, r5, r6		; r5 = s0 = x[ 8] - x[ 9]
+	ADD	r6, r5, r6, LSL #1	; r6 =      x[ 8] + x[ 9]  -> x[ 8]
+	SUB	r9, r9, r10		; r9 = s1 = x[10] - x[11]
+	ADD	r10,r9, r10,LSL #1	; r10=      x[10] + x[11]  -> x[10]
+	SUB	r8, r8, r7		; r8 = s2 = x[ 1] - x[ 0]
+	ADD	r7, r8, r7, LSL #1	; r7 =      x[ 1] + x[ 0]  -> x[ 9]
+	SUB	r12,r12,r11		; r12= s3 = x[ 3] - x[ 2]
+	ADD	r11,r12,r11, LSL #1	; r11=      x[ 3] + x[ 2]  -> x[11]
+	MOV	r14,#0xb5		; r14= cPI2_8
 	STMIA	r4!,{r6,r7,r10,r11}
 
-	SUB	r5, r5, r9		@ r5 = s0 - s1
-	ADD	r9, r5, r9, LSL #1	@ r9 = s0 + s1
+	SUB	r5, r5, r9		; r5 = s0 - s1
+	ADD	r9, r5, r9, LSL #1	; r9 = s0 + s1
 	MOV	r5, r5, ASR #8
-	MUL	r5, r14,r5		@ r5  = (s0-s1)*cPI2_8
-	SUB	r12,r12,r8		@ r12= s3 - s2
-	ADD	r8, r12,r8, LSL #1	@ r8 = s3 + s2
+	MUL	r5, r14,r5		; r5  = (s0-s1)*cPI2_8
+	SUB	r12,r12,r8		; r12= s3 - s2
+	ADD	r8, r12,r8, LSL #1	; r8 = s3 + s2
 
 	MOV	r8, r8, ASR #8
-	MUL	r8, r14,r8		@ r8  = (s3+s2)*cPI2_8
+	MUL	r8, r14,r8		; r8  = (s3+s2)*cPI2_8
 	MOV	r9, r9, ASR #8
-	MUL	r9, r14,r9		@ r9  = (s0+s1)*cPI2_8
+	MUL	r9, r14,r9		; r9  = (s0+s1)*cPI2_8
 	MOV	r12,r12,ASR #8
-	MUL	r12,r14,r12		@ r12 = (s3-s2)*cPI2_8
+	MUL	r12,r14,r12		; r12 = (s3-s2)*cPI2_8
 	STMIA	r1!,{r5,r8,r9,r12}
 
-	@ block2
-	LDMIA	r4,{r5,r6,r9,r10}	@ r5 = x[12]
-					@ r6 = x[13]
-					@ r9 = x[14]
-					@ r10= x[15]
-	LDMIA	r1,{r7,r8,r11,r12}	@ r7 = x[ 4]
-					@ r8 = x[ 5]
-					@ r11= x[ 6]
-					@ r12= x[ 7]
-	SUB	r5, r5, r6		@ r5 = s2 = x[12] - x[13]
-	ADD	r6, r5, r6, LSL #1	@ r6 =      x[12] + x[13]  -> x[12]
-	SUB	r9, r9, r10		@ r9 = s3 = x[14] - x[15]
-	ADD	r10,r9, r10,LSL #1	@ r10=      x[14] + x[15]  -> x[14]
-	SUB	r14,r7, r8		@ r14= s0 = x[ 4] - x[ 5]
-	ADD	r8, r14,r8, LSL #1	@ r8 =      x[ 4] + x[ 5]  -> x[13]
-	SUB	r7, r12,r11		@ r7 = s1 = x[ 7] - x[ 6]
-	ADD	r11,r7, r11, LSL #1	@ r11=      x[ 7] + x[ 6]  -> x[15]
+	; block2
+	LDMIA	r4,{r5,r6,r9,r10}	; r5 = x[12]
+					; r6 = x[13]
+					; r9 = x[14]
+					; r10= x[15]
+	LDMIA	r1,{r7,r8,r11,r12}	; r7 = x[ 4]
+					; r8 = x[ 5]
+					; r11= x[ 6]
+					; r12= x[ 7]
+	SUB	r5, r5, r6		; r5 = s2 = x[12] - x[13]
+	ADD	r6, r5, r6, LSL #1	; r6 =      x[12] + x[13]  -> x[12]
+	SUB	r9, r9, r10		; r9 = s3 = x[14] - x[15]
+	ADD	r10,r9, r10,LSL #1	; r10=      x[14] + x[15]  -> x[14]
+	SUB	r14,r7, r8		; r14= s0 = x[ 4] - x[ 5]
+	ADD	r8, r14,r8, LSL #1	; r8 =      x[ 4] + x[ 5]  -> x[13]
+	SUB	r7, r12,r11		; r7 = s1 = x[ 7] - x[ 6]
+	ADD	r11,r7, r11, LSL #1	; r11=      x[ 7] + x[ 6]  -> x[15]
 	STMIA	r4!,{r6,r8,r10,r11}
 	STMIA	r1!,{r5,r7,r9,r14}
 
-	@ mdct_butterfly_8
+	; mdct_butterfly_8
 	LDMDB	r1,{r6,r7,r8,r9,r10,r11,r12,r14}
-					@ r6 = x[0]
-					@ r7 = x[1]
-					@ r8 = x[2]
-					@ r9 = x[3]
-					@ r10= x[4]
-					@ r11= x[5]
-					@ r12= x[6]
-					@ r14= x[7]
-	ADD	r6, r6, r7		@ r6 = s0 = x[0] + x[1]
-	SUB	r7, r6, r7, LSL #1	@ r7 = s1 = x[0] - x[1]
-	ADD	r8, r8, r9		@ r8 = s2 = x[2] + x[3]
-	SUB	r9, r8, r9, LSL #1	@ r9 = s3 = x[2] - x[3]
-	ADD	r10,r10,r11		@ r10= s4 = x[4] + x[5]
-	SUB	r11,r10,r11,LSL #1	@ r11= s5 = x[4] - x[5]
-	ADD	r12,r12,r14		@ r12= s6 = x[6] + x[7]
-	SUB	r14,r12,r14,LSL #1	@ r14= s7 = x[6] - x[7]
-
-	ADD	r2, r11,r9		@ r2 = x[0] = s5 + s3
-	SUB	r4, r2, r9, LSL #1	@ r4 = x[2] = s5 - s3
-	SUB	r3, r14,r7		@ r3 = x[1] = s7 - s1
-	ADD	r5, r3, r7, LSL #1	@ r5 = x[3] = s7 + s1
-	SUB	r10,r10,r6		@ r10= x[4] = s4 - s0
-	SUB	r11,r12,r8		@ r11= x[5] = s6 - s2
-	ADD	r12,r10,r6, LSL #1	@ r12= x[6] = s4 + s0
-	ADD	r14,r11,r8, LSL #1	@ r14= x[7] = s6 + s2
+					; r6 = x[0]
+					; r7 = x[1]
+					; r8 = x[2]
+					; r9 = x[3]
+					; r10= x[4]
+					; r11= x[5]
+					; r12= x[6]
+					; r14= x[7]
+	ADD	r6, r6, r7		; r6 = s0 = x[0] + x[1]
+	SUB	r7, r6, r7, LSL #1	; r7 = s1 = x[0] - x[1]
+	ADD	r8, r8, r9		; r8 = s2 = x[2] + x[3]
+	SUB	r9, r8, r9, LSL #1	; r9 = s3 = x[2] - x[3]
+	ADD	r10,r10,r11		; r10= s4 = x[4] + x[5]
+	SUB	r11,r10,r11,LSL #1	; r11= s5 = x[4] - x[5]
+	ADD	r12,r12,r14		; r12= s6 = x[6] + x[7]
+	SUB	r14,r12,r14,LSL #1	; r14= s7 = x[6] - x[7]
+
+	ADD	r2, r11,r9		; r2 = x[0] = s5 + s3
+	SUB	r4, r2, r9, LSL #1	; r4 = x[2] = s5 - s3
+	SUB	r3, r14,r7		; r3 = x[1] = s7 - s1
+	ADD	r5, r3, r7, LSL #1	; r5 = x[3] = s7 + s1
+	SUB	r10,r10,r6		; r10= x[4] = s4 - s0
+	SUB	r11,r12,r8		; r11= x[5] = s6 - s2
+	ADD	r12,r10,r6, LSL #1	; r12= x[6] = s4 + s0
+	ADD	r14,r11,r8, LSL #1	; r14= x[7] = s6 + s2
 	STMDB	r1,{r2,r3,r4,r5,r10,r11,r12,r14}
 
-	@ mdct_butterfly_8
+	; mdct_butterfly_8
 	LDMIA	r1,{r6,r7,r8,r9,r10,r11,r12,r14}
-					@ r6 = x[0]
-					@ r7 = x[1]
-					@ r8 = x[2]
-					@ r9 = x[3]
-					@ r10= x[4]
-					@ r11= x[5]
-					@ r12= x[6]
-					@ r14= x[7]
-	ADD	r6, r6, r7		@ r6 = s0 = x[0] + x[1]
-	SUB	r7, r6, r7, LSL #1	@ r7 = s1 = x[0] - x[1]
-	ADD	r8, r8, r9		@ r8 = s2 = x[2] + x[3]
-	SUB	r9, r8, r9, LSL #1	@ r9 = s3 = x[2] - x[3]
-	ADD	r10,r10,r11		@ r10= s4 = x[4] + x[5]
-	SUB	r11,r10,r11,LSL #1	@ r11= s5 = x[4] - x[5]
-	ADD	r12,r12,r14		@ r12= s6 = x[6] + x[7]
-	SUB	r14,r12,r14,LSL #1	@ r14= s7 = x[6] - x[7]
-
-	ADD	r2, r11,r9		@ r2 = x[0] = s5 + s3
-	SUB	r4, r2, r9, LSL #1	@ r4 = x[2] = s5 - s3
-	SUB	r3, r14,r7		@ r3 = x[1] = s7 - s1
-	ADD	r5, r3, r7, LSL #1	@ r5 = x[3] = s7 + s1
-	SUB	r10,r10,r6		@ r10= x[4] = s4 - s0
-	SUB	r11,r12,r8		@ r11= x[5] = s6 - s2
-	ADD	r12,r10,r6, LSL #1	@ r12= x[6] = s4 + s0
-	ADD	r14,r11,r8, LSL #1	@ r14= x[7] = s6 + s2
+					; r6 = x[0]
+					; r7 = x[1]
+					; r8 = x[2]
+					; r9 = x[3]
+					; r10= x[4]
+					; r11= x[5]
+					; r12= x[6]
+					; r14= x[7]
+	ADD	r6, r6, r7		; r6 = s0 = x[0] + x[1]
+	SUB	r7, r6, r7, LSL #1	; r7 = s1 = x[0] - x[1]
+	ADD	r8, r8, r9		; r8 = s2 = x[2] + x[3]
+	SUB	r9, r8, r9, LSL #1	; r9 = s3 = x[2] - x[3]
+	ADD	r10,r10,r11		; r10= s4 = x[4] + x[5]
+	SUB	r11,r10,r11,LSL #1	; r11= s5 = x[4] - x[5]
+	ADD	r12,r12,r14		; r12= s6 = x[6] + x[7]
+	SUB	r14,r12,r14,LSL #1	; r14= s7 = x[6] - x[7]
+
+	ADD	r2, r11,r9		; r2 = x[0] = s5 + s3
+	SUB	r4, r2, r9, LSL #1	; r4 = x[2] = s5 - s3
+	SUB	r3, r14,r7		; r3 = x[1] = s7 - s1
+	ADD	r5, r3, r7, LSL #1	; r5 = x[3] = s7 + s1
+	SUB	r10,r10,r6		; r10= x[4] = s4 - s0
+	SUB	r11,r12,r8		; r11= x[5] = s6 - s2
+	ADD	r12,r10,r6, LSL #1	; r12= x[6] = s4 + s0
+	ADD	r14,r11,r8, LSL #1	; r14= x[7] = s6 + s2
 	STMIA	r1,{r2,r3,r4,r5,r10,r11,r12,r14}
 
 	ADD	r1,r1,#8*4
@@ -947,117 +947,117 @@ mdct_bufferflies_loop3:
 
 	LDMFD	r13,{r0-r3}
 
-mdct_bitreverse_arm_low:
-	@ r0 = points
-	@ r1 = in
-	@ r2 = step
-	@ r3 = shift
+mdct_bitreverse_arm_low
+	; r0 = points
+	; r1 = in
+	; r2 = step
+	; r3 = shift
 
-	MOV	r4, #0			@ r4 = bit = 0
-	ADD	r5, r1, r0, LSL #1	@ r5 = w = x + (n>>1)
+	MOV	r4, #0			; r4 = bit = 0
+	ADD	r5, r1, r0, LSL #1	; r5 = w = x + (n>>1)
 	ADR	r6, bitrev
 	SUB	r5, r5, #8
-brev_lp:
+brev_lp
 	LDRB	r7, [r6, r4, LSR #6]
 	AND	r8, r4, #0x3f
 	LDRB	r8, [r6, r8]
-	ADD	r4, r4, #1		@ bit++
-	@ stall XScale
-	ORR	r7, r7, r8, LSL #6	@ r7 = bitrev[bit]
+	ADD	r4, r4, #1		; bit++
+	; stall XScale
+	ORR	r7, r7, r8, LSL #6	; r7 = bitrev[bit]
 	MOV	r7, r7, LSR r3
-	ADD	r9, r1, r7, LSL #2	@ r9 = xx = x + (b>>shift)
-	CMP	r5, r9			@ if (w > xx)
-	LDR	r10,[r5],#-8		@   r10 = w[0]		w -= 2
-	LDRGT	r11,[r5,#12]		@   r11 = w[1]
-	LDRGT	r12,[r9]		@   r12 = xx[0]
-	LDRGT	r14,[r9,#4]		@   r14 = xx[1]
-	STRGT	r10,[r9]		@   xx[0]= w[0]
-	STRGT	r11,[r9,#4]		@   xx[1]= w[1]
-	STRGT	r12,[r5,#8]		@   w[0] = xx[0]
-	STRGT	r14,[r5,#12]		@   w[1] = xx[1]
+	ADD	r9, r1, r7, LSL #2	; r9 = xx = x + (b>>shift)
+	CMP	r5, r9			; if (w > xx)
+	LDR	r10,[r5],#-8		;   r10 = w[0]		w -= 2
+	LDRGT	r11,[r5,#12]		;   r11 = w[1]
+	LDRGT	r12,[r9]		;   r12 = xx[0]
+	LDRGT	r14,[r9,#4]		;   r14 = xx[1]
+	STRGT	r10,[r9]		;   xx[0]= w[0]
+	STRGT	r11,[r9,#4]		;   xx[1]= w[1]
+	STRGT	r12,[r5,#8]		;   w[0] = xx[0]
+	STRGT	r14,[r5,#12]		;   w[1] = xx[1]
 	CMP	r5,r1
 	BGT	brev_lp
 
-	@ mdct_step7
-	@ r0 = points
-	@ r1 = in
-	@ r2 = step
-	@ r3 = shift
-
-	CMP	r2, #4			@ r5 = T = (step>=4) ?
-	LDRGE	r5, =sincos_lookup0	@          sincos_lookup0 +
-	LDRLT	r5, =sincos_lookup1	@          sincos_lookup0 +
-	ADD	r7, r1, r0, LSL #1	@ r7 = w1 = x + (n>>1)
-	ADDGE	r5, r5, r2, LSR #1	@		            (step>>1)
-	ADD	r8, r5, #1024		@ r8 = Ttop
-step7_loop1:
-	LDR	r6, [r1]		@ r6 = w0[0]
-	LDR	r9, [r1,#4]		@ r9 = w0[1]
-	LDR	r10,[r7,#-8]!		@ r10= w1[0]	w1 -= 2
-	LDR	r11,[r7,#4]		@ r11= w1[1]
-	LDRB	r14,[r5,#1]		@ r14= T[1]
-	LDRB	r12,[r5],r2		@ r12= T[0]	T += step
-
-	ADD	r6, r6, r10		@ r6 = s0 = w0[0] + w1[0]
-	SUB	r10,r6, r10,LSL #1	@ r10= s1b= w0[0] - w1[0]
-	SUB	r11,r11,r9		@ r11= s1 = w1[1] - w0[1]
-	ADD	r9, r11,r9, LSL #1	@ r9 = s0b= w1[1] + w0[1]
+	; mdct_step7
+	; r0 = points
+	; r1 = in
+	; r2 = step
+	; r3 = shift
+
+	CMP	r2, #4			; r5 = T = (step>=4) ?
+	LDRGE	r5, =sincos_lookup0	;          sincos_lookup0 +
+	LDRLT	r5, =sincos_lookup1	;          sincos_lookup0 +
+	ADD	r7, r1, r0, LSL #1	; r7 = w1 = x + (n>>1)
+	ADDGE	r5, r5, r2, LSR #1	;		            (step>>1)
+	ADD	r8, r5, #1024		; r8 = Ttop
+step7_loop1
+	LDR	r6, [r1]		; r6 = w0[0]
+	LDR	r9, [r1,#4]		; r9 = w0[1]
+	LDR	r10,[r7,#-8]!		; r10= w1[0]	w1 -= 2
+	LDR	r11,[r7,#4]		; r11= w1[1]
+	LDRB	r14,[r5,#1]		; r14= T[1]
+	LDRB	r12,[r5],r2		; r12= T[0]	T += step
+
+	ADD	r6, r6, r10		; r6 = s0 = w0[0] + w1[0]
+	SUB	r10,r6, r10,LSL #1	; r10= s1b= w0[0] - w1[0]
+	SUB	r11,r11,r9		; r11= s1 = w1[1] - w0[1]
+	ADD	r9, r11,r9, LSL #1	; r9 = s0b= w1[1] + w0[1]
 
 	MOV	r6, r6, ASR #9
-	MUL	r3, r6, r14		@ r3   = s0*T[1]
+	MUL	r3, r6, r14		; r3   = s0*T[1]
 	MOV	r11,r11,ASR #9
-	MUL	r4, r11,r12		@ r4  += s1*T[0] = s2
+	MUL	r4, r11,r12		; r4  += s1*T[0] = s2
 	ADD	r3, r3, r4
-	MUL	r14,r11,r14		@ r14  = s1*T[1]
-	MUL	r12,r6, r12		@ r12 += s0*T[0] = s3
+	MUL	r14,r11,r14		; r14  = s1*T[1]
+	MUL	r12,r6, r12		; r12 += s0*T[0] = s3
 	SUB	r14,r14,r12
 
-	@ r9 = s0b<<1
-	@ r10= s1b<<1
-	ADD	r9, r3, r9, ASR #1	@ r9 = s0b + s2
-	SUB	r3, r9, r3, LSL #1	@ r3 = s0b - s2
+	; r9 = s0b<<1
+	; r10= s1b<<1
+	ADD	r9, r3, r9, ASR #1	; r9 = s0b + s2
+	SUB	r3, r9, r3, LSL #1	; r3 = s0b - s2
 
-	SUB	r12,r14,r10,ASR #1	@ r12= s3  - s1b
-	ADD	r10,r14,r10,ASR #1	@ r10= s3  + s1b
+	SUB	r12,r14,r10,ASR #1	; r12= s3  - s1b
+	ADD	r10,r14,r10,ASR #1	; r10= s3  + s1b
 	STR	r9, [r1],#4
-	STR	r10,[r1],#4		@ w0 += 2
+	STR	r10,[r1],#4		; w0 += 2
 	STR	r3, [r7]
 	STR	r12,[r7,#4]
 
 	CMP	r5,r8
 	BLT	step7_loop1
 
-step7_loop2:
-	LDR	r6, [r1]		@ r6 = w0[0]
-	LDR	r9, [r1,#4]		@ r9 = w0[1]
-	LDR	r10,[r7,#-8]!		@ r10= w1[0]	w1 -= 2
-	LDR	r11,[r7,#4]		@ r11= w1[1]
-	LDRB	r14,[r5,-r2]!		@ r12= T[1]	T -= step
-	LDRB	r12,[r5,#1]		@ r14= T[0]
+step7_loop2
+	LDR	r6, [r1]		; r6 = w0[0]
+	LDR	r9, [r1,#4]		; r9 = w0[1]
+	LDR	r10,[r7,#-8]!		; r10= w1[0]	w1 -= 2
+	LDR	r11,[r7,#4]		; r11= w1[1]
+	LDRB	r14,[r5,-r2]!		; r12= T[1]	T -= step
+	LDRB	r12,[r5,#1]		; r14= T[0]
 
-	ADD	r6, r6, r10		@ r6 = s0 = w0[0] + w1[0]
-	SUB	r10,r6, r10,LSL #1	@ r10= s1b= w0[0] - w1[0]
-	SUB	r11,r11,r9		@ r11= s1 = w1[1] - w0[1]
-	ADD	r9, r11,r9, LSL #1	@ r9 = s0b= w1[1] + w0[1]
+	ADD	r6, r6, r10		; r6 = s0 = w0[0] + w1[0]
+	SUB	r10,r6, r10,LSL #1	; r10= s1b= w0[0] - w1[0]
+	SUB	r11,r11,r9		; r11= s1 = w1[1] - w0[1]
+	ADD	r9, r11,r9, LSL #1	; r9 = s0b= w1[1] + w0[1]
 
 	MOV	r6, r6, ASR #9
-	MUL	r3, r6, r14		@ r3   = s0*T[0]
+	MUL	r3, r6, r14		; r3   = s0*T[0]
 	MOV	r11,r11,ASR #9
-	MUL	r4, r11,r12		@ r4  += s1*T[1] = s2
+	MUL	r4, r11,r12		; r4  += s1*T[1] = s2
 	ADD	r3, r3, r4
-	MUL	r14,r11,r14		@ r14  = s1*T[0]
-	MUL	r12,r6, r12		@ r12 += s0*T[1] = s3
+	MUL	r14,r11,r14		; r14  = s1*T[0]
+	MUL	r12,r6, r12		; r12 += s0*T[1] = s3
 	SUB	r14,r14,r12
 
-	@ r9 = s0b<<1
-	@ r10= s1b<<1
-	ADD	r9, r3, r9, ASR #1	@ r9 = s0b + s2
-	SUB	r3, r9, r3, LSL #1	@ r3 = s0b - s2
+	; r9 = s0b<<1
+	; r10= s1b<<1
+	ADD	r9, r3, r9, ASR #1	; r9 = s0b + s2
+	SUB	r3, r9, r3, LSL #1	; r3 = s0b - s2
 
-	SUB	r12,r14,r10,ASR #1	@ r12= s3  - s1b
-	ADD	r10,r14,r10,ASR #1	@ r10= s3  + s1b
+	SUB	r12,r14,r10,ASR #1	; r12= s3  - s1b
+	ADD	r10,r14,r10,ASR #1	; r10= s3  + s1b
 	STR	r9, [r1],#4
-	STR	r10,[r1],#4		@ w0 += 2
+	STR	r10,[r1],#4		; w0 += 2
 	STR	r3, [r7]
 	STR	r12,[r7,#4]
 
@@ -1066,110 +1066,110 @@ step7_loop2:
 
 	LDMFD	r13!,{r0-r3}
 
-	@ r0 = points
-	@ r1 = in
-	@ r2 = step
-	@ r3 = shift
-	MOV	r2, r2, ASR #2		@ r2 = step >>= 2
+	; r0 = points
+	; r1 = in
+	; r2 = step
+	; r3 = shift
+	MOV	r2, r2, ASR #2		; r2 = step >>= 2
 	CMP	r2, #0
 	CMPNE	r2, #1
 	BEQ	mdct_end
 
-	@ step > 1 (default case)
-	CMP	r2, #4			@ r5 = T = (step>=4) ?
-	LDRGE	r5, =sincos_lookup0	@          sincos_lookup0 +
-	LDRLT	r5, =sincos_lookup1	@          sincos_lookup1
-	ADD	r7, r1, r0, LSL #1	@ r7 = iX = x + (n>>1)
-	ADDGE	r5, r5, r2, LSR #1	@		            (step>>1)
-mdct_step8_default:
-	LDR	r6, [r1],#4		@ r6 =  s0 = x[0]
-	LDR	r8, [r1],#4		@ r8 = -s1 = x[1]
-	LDRB	r12,[r5,#1]       	@ r12= T[1]
-	LDRB	r14,[r5],r2		@ r14= T[0]	T += step
-	RSB	r8, r8, #0		@ r8 = s1
-
-	@ XPROD31(s0, s1, T[0], T[1], x, x+1)
-	@ x[0] = s0 * T[0] + s1 * T[1]      x[1] = s1 * T[0] - s0 * T[1]
+	; step > 1 (default case)
+	CMP	r2, #4			; r5 = T = (step>=4) ?
+	LDRGE	r5, =sincos_lookup0	;          sincos_lookup0 +
+	LDRLT	r5, =sincos_lookup1	;          sincos_lookup1
+	ADD	r7, r1, r0, LSL #1	; r7 = iX = x + (n>>1)
+	ADDGE	r5, r5, r2, LSR #1	;		            (step>>1)
+mdct_step8_default
+	LDR	r6, [r1],#4		; r6 =  s0 = x[0]
+	LDR	r8, [r1],#4		; r8 = -s1 = x[1]
+	LDRB	r12,[r5,#1]       	; r12= T[1]
+	LDRB	r14,[r5],r2		; r14= T[0]	T += step
+	RSB	r8, r8, #0		; r8 = s1
+
+	; XPROD31(s0, s1, T[0], T[1], x, x+1)
+	; x[0] = s0 * T[0] + s1 * T[1]      x[1] = s1 * T[0] - s0 * T[1]
 	MOV	r6, r6, ASR #8
 	MOV	r8, r8, ASR #8
-	MUL	r10,r8, r12		@ r10  = s1 * T[1]
+	MUL	r10,r8, r12		; r10  = s1 * T[1]
 	CMP	r1, r7
-	MLA	r10,r6, r14,r10	@ r10 += s0 * T[0]
-	RSB	r6, r6, #0		@ r6 = -s0
-	MUL	r11,r8, r14		@ r11  = s1 * T[0]
-	MLA	r11,r6, r12,r11	@ r11 -= s0 * T[1]
+	MLA	r10,r6, r14,r10	; r10 += s0 * T[0]
+	RSB	r6, r6, #0		; r6 = -s0
+	MUL	r11,r8, r14		; r11  = s1 * T[0]
+	MLA	r11,r6, r12,r11	; r11 -= s0 * T[1]
 	STR	r10,[r1,#-8]
 	STR	r11,[r1,#-4]
 	BLT	mdct_step8_default
 
-mdct_end:
+mdct_end
 	MOV	r0, r2
 	LDMFD	r13!,{r4-r11,PC}
 
-bitrev:
-	.byte	0
-	.byte	32
-	.byte	16
-	.byte	48
-	.byte	8
-	.byte	40
-	.byte	24
-	.byte	56
-	.byte	4
-	.byte	36
-	.byte	20
-	.byte	52
-	.byte	12
-	.byte	44
-	.byte	28
-	.byte	60
-	.byte	2
-	.byte	34
-	.byte	18
-	.byte	50
-	.byte	10
-	.byte	42
-	.byte	26
-	.byte	58
-	.byte	6
-	.byte	38
-	.byte	22
-	.byte	54
-	.byte	14
-	.byte	46
-	.byte	30
-	.byte	62
-	.byte	1
-	.byte	33
-	.byte	17
-	.byte	49
-	.byte	9
-	.byte	41
-	.byte	25
-	.byte	57
-	.byte	5
-	.byte	37
-	.byte	21
-	.byte	53
-	.byte	13
-	.byte	45
-	.byte	29
-	.byte	61
-	.byte	3
-	.byte	35
-	.byte	19
-	.byte	51
-	.byte	11
-	.byte	43
-	.byte	27
-	.byte	59
-	.byte	7
-	.byte	39
-	.byte	23
-	.byte	55
-	.byte	15
-	.byte	47
-	.byte	31
-	.byte	63
-
-	@ END
+bitrev
+	DCB	0
+	DCB	32
+	DCB	16
+	DCB	48
+	DCB	8
+	DCB	40
+	DCB	24
+	DCB	56
+	DCB	4
+	DCB	36
+	DCB	20
+	DCB	52
+	DCB	12
+	DCB	44
+	DCB	28
+	DCB	60
+	DCB	2
+	DCB	34
+	DCB	18
+	DCB	50
+	DCB	10
+	DCB	42
+	DCB	26
+	DCB	58
+	DCB	6
+	DCB	38
+	DCB	22
+	DCB	54
+	DCB	14
+	DCB	46
+	DCB	30
+	DCB	62
+	DCB	1
+	DCB	33
+	DCB	17
+	DCB	49
+	DCB	9
+	DCB	41
+	DCB	25
+	DCB	57
+	DCB	5
+	DCB	37
+	DCB	21
+	DCB	53
+	DCB	13
+	DCB	45
+	DCB	29
+	DCB	61
+	DCB	3
+	DCB	35
+	DCB	19
+	DCB	51
+	DCB	11
+	DCB	43
+	DCB	27
+	DCB	59
+	DCB	7
+	DCB	39
+	DCB	23
+	DCB	55
+	DCB	15
+	DCB	47
+	DCB	31
+	DCB	63
+
+	END