Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Saturate MDCT output post-TDAC rather than pre- | Jean-Marc Valin | 2016-07-24 | 1 | -4/+4 |
| | | | | Gives us a tighter bound on the pitch postfilter input to avoid overflows | ||||
* | saturate MDCT output | Jean-Marc Valin | 2016-07-22 | 1 | -4/+4 |
| | |||||
* | silencing overflows in MDCT and FFT | Jean-Marc Valin | 2016-07-22 | 1 | -8/+8 |
| | |||||
* | armv7(float): Optimize decode usecase using NE10 library | Viswanath Puttagunta | 2015-10-07 | 1 | -2/+3 |
| | | | | | | | | | | | | Optimize opus decode (float only) use case using ARM NE10. Mainly effects opus_ifft and ctl_mdct_backward and related functions. Work based on previous Encode optimization using ARM NE10 library. See previous commit for details on how to enable this. Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org> | ||||
* | armv7(float): Optimize encode usecase using NE10 library | Viswanath Puttagunta | 2015-10-07 | 1 | -7/+8 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | Optimize opus encode (float only) usecase using ARM NE10 library. Mainly effects opus_fft and ctl_mdct_forward and related functions. This optimization can be used for ARM CPUs that have NEON VFP unit. This patch only enables optimizations for ARMv7. Official ARM NE10 library page available at http://projectne10.github.io/Ne10/ To enable this optimization, use --enable-intrinsics --with-NE10=<install_prefix> or --enable-intrinsics --with-NE10-libraries=<NE10_lib_dir> --with-NE10-includes=<NE10_includes_dir> Compile time checks made during configure process to make sure optimization option available only when compiler supports NEON instrinsics. Runtime checks made to make sure optimized functions only called on appropriate hardware. Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org> | ||||
* | MIPS optimizations | Rhishikesh Agashe | 2014-06-19 | 1 | -0/+9 |
| | | | | Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> | ||||
* | Fixes an aliasing bug in the MDCT when the frame size isn't a multiple of 4. | Marcello Caramma (mcaramma) | 2014-02-24 | 1 | -2/+2 |
| | | | | | | | In that case, the yp0 and yp1 ended up aliasing for the last element, despite being marked as restrict. Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> | ||||
* | Unifying scaling of fixed-point and float FFT | Jean-Marc Valin | 2013-12-29 | 1 | -13/+3 |
| | |||||
* | Fixes C89 issue | Jean-Marc Valin | 2013-12-29 | 1 | -2/+4 |
| | |||||
* | Getting rid of some negations | Jean-Marc Valin | 2013-12-28 | 1 | -10/+10 |
| | | | | | Since we're doing two rotations, we can invert the sign on both. Also adding a few comments for optimizing the FFT. | ||||
* | Slightly improving the accuracy of the fixed-point MDCT downscale | Jean-Marc Valin | 2013-12-28 | 1 | -13/+9 |
| | | | | Also simplifying the code | ||||
* | Commit 99968ab was causing us to allocate too much stack in the MDCT | Jean-Marc Valin | 2013-12-27 | 1 | -1/+1 |
| | |||||
* | Removes the separate 1/8N rotation in the (I)MDCT and unmerges the MDCT sizes | Jean-Marc Valin | 2013-12-27 | 1 | -62/+66 |
| | | | | Undoes commits f7547a4e and 72513f3c | ||||
* | Inverse MDCT no longer requires any scratch space | Jean-Marc Valin | 2013-12-23 | 1 | -14/+10 |
| | |||||
* | Merges the FFT scaling with the MDCT pre-rotate | Jean-Marc Valin | 2013-12-21 | 1 | -7/+14 |
| | |||||
* | Getting rid of the inverse FFT entirely | Jean-Marc Valin | 2013-12-21 | 1 | -8/+11 |
| | | | | IMDCT now uses the forward FFT. | ||||
* | Moving bitrev step to forward MDCT too | Jean-Marc Valin | 2013-12-21 | 1 | -11/+31 |
| | |||||
* | Moves the bitrev step to the IMDCT pre-rotation | Jean-Marc Valin | 2013-12-20 | 1 | -12/+15 |
| | |||||
* | Do imdct post-rotate and deshuffle in-place. | Nils Wallménius | 2012-12-03 | 1 | -15/+28 |
| | | | | | | Gets rid of one stack buffer. Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> | ||||
* | Fixes forward MDCT for overlaps that aren't a multiple of 4 | Jean-Marc Valin | 2012-12-03 | 1 | -2/+2 |
| | | | | Adds more sizes to test_unit_mdct.c | ||||
* | Place output of imdct post-rotate and deshuffle in out buffer to avoid ↵ | Nils Wallménius | 2012-12-01 | 1 | -28/+8 |
| | | | | | | copying and simplify the code. Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> | ||||
* | Window both sides of overlap when adding them. | Nils Wallménius | 2012-12-01 | 1 | -12/+5 |
| | | | | Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> | ||||
* | Merge inverse mdct post-rotate and de-shuffle loops | Nils Wallménius | 2012-10-21 | 1 | -18/+9 |
| | | | | | | Saves some memory access and gives a tiny speedup Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org> | ||||
* | Merge branch 'exp_analysis7' | Jean-Marc Valin | 2012-10-09 | 1 | -2/+4 |
|\ | | | | | | | | | | | | | | | Conflicts: celt/celt.c celt/mdct.c include/opus_defines.h src/opus_encoder.c | ||||
| * | Adds a 3 Hz high-pass filter and boost allocation on leakage | Jean-Marc Valin | 2012-07-13 | 1 | -2/+4 |
| | | | | | | | | | | | | | | Fixes two leakage problems on the wood blocks sample - Removes DC which causes leakage with no masking - Detect leakage by comparing short-MDCT energy to long-MDCT energy and boost allocation for bands with leakage | ||||
* | | Replace C99 restrict keyword with OPUS_RESTRICT. | Gregory Maxwell | 2012-07-18 | 1 | -29/+29 |
|/ | | | | | | | We had previously advised people to -Drestrict on non-C99 compilers, but this creates problems for some of the MSVC headers. Instead this just uses a macro and defines it sanely. | ||||
* | s/FOUNDATION/COPYRIGHT OWNER/ in CELT code and "glue code" | Jean-Marc Valin | 2012-04-20 | 1 | -2/+2 |
| | | | | Also added 3rd clause to "master" COPYING file | ||||
* | renames the libcelt/ directory to celt/ | Jean-Marc Valin | 2011-09-13 | 1 | -0/+332 |
| | |||||
* | Moved all SILK source code to the silk/ directory | Jean-Marc Valin | 2011-04-28 | 1 | -6/+0 |
| | |||||
* | Loss rate adaptation for the CELT layer | Jean-Marc Valin | 2011-04-25 | 1 | -5/+5 |
| | |||||
* | draft update | Jean-Marc Valin | 2011-03-31 | 1 | -5/+5 |
| | |||||
* | Build fixes | Jean-Marc Valin | 2011-03-23 | 1 | -5/+5 |
| | |||||
* | Making mode switching use the same window as CELT (squared) | Jean-Marc Valin | 2011-03-21 | 1 | -5/+5 |
| | |||||
* | CELT update | Jean-Marc Valin | 2011-03-18 | 1 | -5/+5 |
| | | | | With minor fixes | ||||
* | More project files update | Jean-Marc Valin | 2011-03-18 | 1 | -5/+5 |
| | |||||
* | Fixes a few PLC/DTX bugs due to the recent decode API change | Jean-Marc Valin | 2011-03-16 | 1 | -5/+5 |
| | |||||
* | Adding constrained VBR mode | Jean-Marc Valin | 2011-03-14 | 1 | -5/+5 |
| | |||||
* | Draft update (allocation | Gregory Maxwell | 2011-03-14 | 1 | -5/+5 |
| | |||||
* | Disable newly introduced CELT signalling | Jean-Marc Valin | 2011-03-11 | 1 | -5/+5 |
| | |||||
* | Nothing to see here | Jean-Marc Valin | 2011-03-10 | 1 | -5/+5 |
| | |||||
* | CELT update | Jean-Marc Valin | 2011-03-09 | 1 | -18/+5 |
| | |||||
* | Support for glitchles mode switching | Jean-Marc Valin | 2011-03-07 | 1 | -5/+18 |
| | | | | | | Uses a 5ms redundant CELT frame embedded into the SILK or hybrid packet to handle the switching. It's still possible to use the PLC-based method when no redundant packet is included. | ||||
* | SILK/CELT update | Jean-Marc Valin | 2011-02-28 | 1 | -5/+5 |
| | |||||
* | preparing the next version | Jean-Marc Valin | 2011-02-14 | 1 | -5/+5 |
| | |||||
* | missing files | Jean-Marc Valin | 2011-02-14 | 1 | -10/+5 |
| | |||||
* | SILK and CELT updates | Jean-Marc Valin | 2011-02-14 | 1 | -5/+10 |
| | |||||
* | Update for in-band FEC | Koen Vos | 2011-02-14 | 1 | -5/+5 |
| | |||||
* | draft date | Jean-Marc Valin | 2011-02-04 | 1 | -0/+6 |
| | |||||
* | CELT update | Jean-Marc Valin | 2011-02-04 | 1 | -6/+0 |
| | |||||
* | Tuning the hybrid bit-rate split | Koen Vos | 2011-02-04 | 1 | -24/+5 |
| |