summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorPaul B Mahol <onemda@gmail.com>2023-01-21 19:25:41 +0100
committerPaul B Mahol <onemda@gmail.com>2023-02-04 09:36:01 +0100
commit651da919153e385f0769238c091109c06a142ca6 (patch)
treec6fabbb173d286bac8da7c9b9cf47ec9b8480156
parent9a820ec8b1e2323b70a1cebd204bf459bf7daa1a (diff)
downloadffmpeg-651da919153e385f0769238c091109c06a142ca6.tar.gz
avcodec: add WavArc decoder
-rw-r--r--Changelog1
-rw-r--r--doc/general_contents.texi1
-rw-r--r--libavcodec/Makefile1
-rw-r--r--libavcodec/allcodecs.c1
-rw-r--r--libavcodec/codec_desc.c7
-rw-r--r--libavcodec/codec_id.h1
-rw-r--r--libavcodec/version.h2
-rw-r--r--libavcodec/wavarc.c460
8 files changed, 473 insertions, 1 deletions
diff --git a/Changelog b/Changelog
index cdbe43eac1..2d2206e1ba 100644
--- a/Changelog
+++ b/Changelog
@@ -37,6 +37,7 @@ version <next>:
- XMD ADPCM decoder and demuxer
- media100 to mjpegb bsf
- ffmpeg CLI new option: -fix_sub_duration_heartbeat
+- WavArc decoder
version 5.1:
diff --git a/doc/general_contents.texi b/doc/general_contents.texi
index 87e180c979..84df3432cf 100644
--- a/doc/general_contents.texi
+++ b/doc/general_contents.texi
@@ -1356,6 +1356,7 @@ following image formats are supported:
@item Vorbis @tab E @tab X
@tab A native but very primitive encoder exists.
@item Voxware MetaSound @tab @tab X
+@item Waveform Archiver @tab @tab X
@item WavPack @tab X @tab X
@item Westwood Audio (SND1) @tab @tab X
@item Windows Media Audio 1 @tab X @tab X
diff --git a/libavcodec/Makefile b/libavcodec/Makefile
index 1fb963f820..4971832ff4 100644
--- a/libavcodec/Makefile
+++ b/libavcodec/Makefile
@@ -781,6 +781,7 @@ OBJS-$(CONFIG_VP9_V4L2M2M_DECODER) += v4l2_m2m_dec.o
OBJS-$(CONFIG_VQA_DECODER) += vqavideo.o
OBJS-$(CONFIG_VQC_DECODER) += vqcdec.o
OBJS-$(CONFIG_WADY_DPCM_DECODER) += dpcm.o
+OBJS-$(CONFIG_WAVARC_DECODER) += wavarc.o
OBJS-$(CONFIG_WAVPACK_DECODER) += wavpack.o wavpackdata.o dsd.o
OBJS-$(CONFIG_WAVPACK_ENCODER) += wavpackdata.o wavpackenc.o
OBJS-$(CONFIG_WBMP_DECODER) += wbmpdec.o
diff --git a/libavcodec/allcodecs.c b/libavcodec/allcodecs.c
index ff82423a88..b80b6983e9 100644
--- a/libavcodec/allcodecs.c
+++ b/libavcodec/allcodecs.c
@@ -538,6 +538,7 @@ extern const FFCodec ff_twinvq_decoder;
extern const FFCodec ff_vmdaudio_decoder;
extern const FFCodec ff_vorbis_encoder;
extern const FFCodec ff_vorbis_decoder;
+extern const FFCodec ff_wavarc_decoder;
extern const FFCodec ff_wavpack_encoder;
extern const FFCodec ff_wavpack_decoder;
extern const FFCodec ff_wmalossless_decoder;
diff --git a/libavcodec/codec_desc.c b/libavcodec/codec_desc.c
index 8ab228d846..57d0f98211 100644
--- a/libavcodec/codec_desc.c
+++ b/libavcodec/codec_desc.c
@@ -3353,6 +3353,13 @@ static const AVCodecDescriptor codec_descriptors[] = {
.long_name = NULL_IF_CONFIG_SMALL("FTR Voice"),
.props = AV_CODEC_PROP_INTRA_ONLY | AV_CODEC_PROP_LOSSY,
},
+ {
+ .id = AV_CODEC_ID_WAVARC,
+ .type = AVMEDIA_TYPE_AUDIO,
+ .name = "wavarc",
+ .long_name = NULL_IF_CONFIG_SMALL("Waveform Archiver"),
+ .props = AV_CODEC_PROP_INTRA_ONLY | AV_CODEC_PROP_LOSSLESS,
+ },
/* subtitle codecs */
{
diff --git a/libavcodec/codec_id.h b/libavcodec/codec_id.h
index 0c574c9619..ad1131b464 100644
--- a/libavcodec/codec_id.h
+++ b/libavcodec/codec_id.h
@@ -536,6 +536,7 @@ enum AVCodecID {
AV_CODEC_ID_MISC4,
AV_CODEC_ID_APAC,
AV_CODEC_ID_FTR,
+ AV_CODEC_ID_WAVARC,
/* subtitle codecs */
AV_CODEC_ID_FIRST_SUBTITLE = 0x17000, ///< A dummy ID pointing at the start of subtitle codecs.
diff --git a/libavcodec/version.h b/libavcodec/version.h
index 499c6bb175..310c80eeef 100644
--- a/libavcodec/version.h
+++ b/libavcodec/version.h
@@ -29,7 +29,7 @@
#include "version_major.h"
-#define LIBAVCODEC_VERSION_MINOR 61
+#define LIBAVCODEC_VERSION_MINOR 62
#define LIBAVCODEC_VERSION_MICRO 100
#define LIBAVCODEC_VERSION_INT AV_VERSION_INT(LIBAVCODEC_VERSION_MAJOR, \
diff --git a/libavcodec/wavarc.c b/libavcodec/wavarc.c
new file mode 100644
index 0000000000..898c3c2055
--- /dev/null
+++ b/libavcodec/wavarc.c
@@ -0,0 +1,460 @@
+/*
+ * WavArc audio decoder
+ * Copyright (c) 2023 Paul B Mahol
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "libavutil/internal.h"
+#include "libavutil/intreadwrite.h"
+#include "avcodec.h"
+#include "codec_internal.h"
+#include "decode.h"
+#include "get_bits.h"
+#include "bytestream.h"
+#include "mathops.h"
+#include "unary.h"
+
+typedef struct WavArcContext {
+ GetBitContext gb;
+
+ int shift;
+ int nb_samples;
+ int offset;
+
+ int eof;
+ int skip;
+ uint8_t *bitstream;
+ int64_t max_framesize;
+ int bitstream_size;
+ int bitstream_index;
+
+ int pred[2][70];
+ int filter[2][70];
+ int samples[2][640];
+} WavArcContext;
+
+static av_cold int wavarc_init(AVCodecContext *avctx)
+{
+ WavArcContext *s = avctx->priv_data;
+
+ if (avctx->extradata_size < 44)
+ return AVERROR_INVALIDDATA;
+ if (AV_RL32(avctx->extradata + 16) != MKTAG('R','I','F','F'))
+ return AVERROR_INVALIDDATA;
+ if (AV_RL32(avctx->extradata + 24) != MKTAG('W','A','V','E'))
+ return AVERROR_INVALIDDATA;
+ if (AV_RL32(avctx->extradata + 28) != MKTAG('f','m','t',' '))
+ return AVERROR_INVALIDDATA;
+ if (AV_RL16(avctx->extradata + 38) != 1 &&
+ AV_RL16(avctx->extradata + 38) != 2)
+ return AVERROR_INVALIDDATA;
+
+ av_channel_layout_uninit(&avctx->ch_layout);
+ av_channel_layout_default(&avctx->ch_layout, AV_RL16(avctx->extradata + 38));
+ avctx->sample_rate = AV_RL32(avctx->extradata + 40);
+
+ switch (avctx->extradata[36]) {
+ case 0: avctx->sample_fmt = AV_SAMPLE_FMT_U8P; break;
+ case 1: avctx->sample_fmt = AV_SAMPLE_FMT_S16P; break;
+ }
+
+ s->shift = 0;
+ switch (avctx->codec_tag) {
+ case MKTAG('1','D','I','F'):
+ s->nb_samples = 256;
+ s->offset = 4;
+ break;
+ case MKTAG('2','S','L','P'):
+ case MKTAG('3','N','L','P'):
+ case MKTAG('4','A','L','P'):
+ s->nb_samples = 570;
+ s->offset = 70;
+ break;
+ default:
+ return AVERROR_INVALIDDATA;
+ }
+
+ s->max_framesize = s->nb_samples * 16;
+ s->bitstream = av_calloc(s->max_framesize, sizeof(*s->bitstream));
+ if (!s->bitstream)
+ return AVERROR(ENOMEM);
+
+ return 0;
+}
+
+static unsigned get_urice(GetBitContext *gb, int k)
+{
+ unsigned x = get_unary(gb, 1, get_bits_left(gb));
+ unsigned y = get_bits_long(gb, k);
+ unsigned z = (x << k) | y;
+
+ return z;
+}
+
+static int get_srice(GetBitContext *gb, int k)
+{
+ unsigned z = get_urice(gb, k);
+
+ return (z & 1) ? ~((int)(z >> 1)) : z >> 1;
+}
+
+static void do_stereo(WavArcContext *s, int ch, int correlated, int len)
+{
+ const int nb_samples = s->nb_samples;
+ const int shift = s->shift;
+
+ if (ch == 0) {
+ if (correlated) {
+ for (int n = 0; n < len; n++) {
+ s->samples[0][n] = s->samples[0][nb_samples + n] >> shift;
+ s->samples[1][n] = s->pred[1][n] >> shift;
+ }
+ } else {
+ for (int n = 0; n < len; n++) {
+ s->samples[0][n] = s->samples[0][nb_samples + n] >> shift;
+ s->samples[1][n] = s->pred[0][n] >> shift;
+ }
+ }
+ } else {
+ if (correlated) {
+ for (int n = 0; n < nb_samples; n++)
+ s->samples[1][n + len] += s->samples[0][n + len];
+ }
+ for (int n = 0; n < len; n++) {
+ s->pred[0][n] = s->samples[1][nb_samples + n];
+ s->pred[1][n] = s->pred[0][n] - s->samples[0][nb_samples + n];
+ }
+ }
+}
+
+static int decode_1dif(AVCodecContext *avctx,
+ WavArcContext *s, GetBitContext *gb)
+{
+ int ch, finished, fill, correlated;
+
+ ch = 0;
+ finished = 0;
+ while (!finished) {
+ int *samples = s->samples[ch];
+ int k, block_type;
+
+ if (get_bits_left(gb) <= 0)
+ return AVERROR_INVALIDDATA;
+
+ block_type = get_urice(gb, 1);
+ if (block_type < 4 && block_type >= 0) {
+ k = 1 + (avctx->sample_fmt == AV_SAMPLE_FMT_S16P);
+ k = get_urice(gb, k) + 1;
+ }
+
+ switch (block_type) {
+ case 8:
+ s->eof = 1;
+ return AVERROR_EOF;
+ case 7:
+ s->nb_samples = get_bits(gb, 8);
+ continue;
+ case 6:
+ s->shift = get_urice(gb, 2);
+ continue;
+ case 5:
+ if (avctx->sample_fmt == AV_SAMPLE_FMT_U8P) {
+ fill = (int8_t)get_bits(gb, 8);
+ fill -= 0x80;
+ } else {
+ fill = (int16_t)get_bits(gb, 16);
+ fill -= 0x8000;
+ }
+
+ for (int n = 0; n < s->nb_samples; n++)
+ samples[n + 4] = fill;
+ finished = 1;
+ break;
+ case 4:
+ for (int n = 0; n < s->nb_samples; n++)
+ samples[n + 4] = 0;
+ finished = 1;
+ break;
+ case 3:
+ for (int n = 0; n < s->nb_samples; n++)
+ samples[n + 4] = get_srice(gb, k) + (samples[n + 3] - samples[n + 2]) * 3 +
+ samples[n + 1];
+ finished = 1;
+ break;
+ case 2:
+ for (int n = 0; n < s->nb_samples; n++)
+ samples[n + 4] = get_srice(gb, k) + (samples[n + 3] * 2 - samples[n + 2]);
+ finished = 1;
+ break;
+ case 1:
+ for (int n = 0; n < s->nb_samples; n++)
+ samples[n + 4] = get_srice(gb, k) + samples[n + 3];
+ finished = 1;
+ break;
+ case 0:
+ for (int n = 0; n < s->nb_samples; n++)
+ samples[n + 4] = get_srice(gb, k);
+ finished = 1;
+ break;
+ default:
+ return AVERROR_INVALIDDATA;
+ }
+
+ if (finished == 1 && avctx->ch_layout.nb_channels == 2) {
+ if (ch == 0)
+ correlated = get_bits1(gb);
+ finished = ch != 0;
+ do_stereo(s, ch, correlated, 4);
+ ch = 1;
+ }
+ }
+
+ if (avctx->ch_layout.nb_channels == 1) {
+ for (int n = 0; n < 4; n++)
+ s->samples[0][n] = s->samples[0][s->nb_samples + n] >> s->shift;
+ }
+
+ return 0;
+}
+
+static int decode_2slp(AVCodecContext *avctx,
+ WavArcContext *s, GetBitContext *gb)
+{
+ int ch, finished, fill, correlated, order;
+
+ ch = 0;
+ finished = 0;
+ while (!finished) {
+ int *samples = s->samples[ch];
+ int k, block_type;
+
+ if (get_bits_left(gb) <= 0)
+ return AVERROR_INVALIDDATA;
+
+ block_type = get_urice(gb, 1);
+ if (block_type < 5 && block_type >= 0) {
+ k = 1 + (avctx->sample_fmt == AV_SAMPLE_FMT_S16P);
+ k = get_urice(gb, k) + 1;
+ }
+
+ switch (block_type) {
+ case 9:
+ s->eof = 1;
+ return AVERROR_EOF;
+ case 8:
+ s->nb_samples = get_urice(gb, 8);
+ continue;
+ case 7:
+ s->shift = get_urice(gb, 2);
+ continue;
+ case 6:
+ if (avctx->sample_fmt == AV_SAMPLE_FMT_U8P) {
+ fill = (int8_t)get_bits(gb, 8);
+ fill -= 0x80;
+ } else {
+ fill = (int16_t)get_bits(gb, 16);
+ fill -= 0x8000;
+ }
+
+ for (int n = 0; n < s->nb_samples; n++)
+ samples[n + 70] = fill;
+ finished = 1;
+ break;
+ case 5:
+ for (int n = 0; n < s->nb_samples; n++)
+ samples[n + 70] = 0;
+ finished = 1;
+ break;
+ case 4:
+ for (int n = 0; n < s->nb_samples; n++)
+ samples[n + 70] = get_srice(gb, k) + (samples[n + 69] - samples[n + 68]) * 3 +
+ samples[n + 67];
+ finished = 1;
+ break;
+ case 3:
+ for (int n = 0; n < s->nb_samples; n++)
+ samples[n + 70] = get_srice(gb, k) + (samples[n + 69] * 2 - samples[n + 68]);
+ finished = 1;
+ break;
+ case 2:
+ for (int n = 0; n < s->nb_samples; n++)
+ samples[n + 70] = get_srice(gb, k);
+ finished = 1;
+ break;
+ case 1:
+ for (int n = 0; n < s->nb_samples; n++)
+ samples[n + 70] = get_srice(gb, k) + samples[n + 69];
+ finished = 1;
+ break;
+ case 0:
+ order = get_urice(gb, 2);
+ for (int o = 0; o < order; o++)
+ s->filter[ch][o] = get_srice(gb, 2);
+ for (int n = 0; n < s->nb_samples; n++) {
+ int sum = 15;
+
+ for (int o = 0; o < order; o++)
+ sum += s->filter[ch][o] * samples[n + 70 - o - 1];
+
+ samples[n + 70] = get_srice(gb, k) + (sum >> 4);
+ }
+ finished = 1;
+ break;
+ default:
+ return AVERROR_INVALIDDATA;
+ }
+
+ if (finished == 1 && avctx->ch_layout.nb_channels == 2) {
+ if (ch == 0)
+ correlated = get_bits1(gb);
+ finished = ch != 0;
+ do_stereo(s, ch, correlated, 70);
+ ch = 1;
+ }
+ }
+
+ if (avctx->ch_layout.nb_channels == 1) {
+ for (int n = 0; n < 70; n++)
+ s->samples[0][n] = s->samples[0][s->nb_samples + n] >> s->shift;
+ }
+
+ return 0;
+}
+
+static int wavarc_decode(AVCodecContext *avctx, AVFrame *frame,
+ int *got_frame_ptr, AVPacket *pkt)
+{
+ WavArcContext *s = avctx->priv_data;
+ GetBitContext *gb = &s->gb;
+ int buf_size, input_buf_size;
+ const uint8_t *buf;
+ int ret, n;
+
+ if ((!pkt->size && !s->bitstream_size) || s->nb_samples == 0 || s->eof) {
+ *got_frame_ptr = 0;
+ return pkt->size;
+ }
+
+ buf_size = FFMIN(pkt->size, s->max_framesize - s->bitstream_size);
+ input_buf_size = buf_size;
+ if (s->bitstream_index + s->bitstream_size + buf_size + AV_INPUT_BUFFER_PADDING_SIZE > s->max_framesize) {
+ memmove(s->bitstream, &s->bitstream[s->bitstream_index], s->bitstream_size);
+ s->bitstream_index = 0;
+ }
+ if (pkt->data)
+ memcpy(&s->bitstream[s->bitstream_index + s->bitstream_size], pkt->data, buf_size);
+ buf = &s->bitstream[s->bitstream_index];
+ buf_size += s->bitstream_size;
+ s->bitstream_size = buf_size;
+ if (buf_size < s->max_framesize && pkt->data) {
+ *got_frame_ptr = 0;
+ return input_buf_size;
+ }
+
+ if ((ret = init_get_bits8(gb, buf, buf_size)) < 0)
+ return ret;
+ skip_bits(gb, s->skip);
+
+ switch (avctx->codec_tag) {
+ case MKTAG('1','D','I','F'):
+ ret = decode_1dif(avctx, s, gb);
+ break;
+ case MKTAG('2','S','L','P'):
+ case MKTAG('3','N','L','P'):
+ case MKTAG('4','A','L','P'):
+ ret = decode_2slp(avctx, s, gb);
+ break;
+ default:
+ ret = AVERROR_INVALIDDATA;
+ }
+
+ if (ret < 0)
+ goto fail;
+
+ s->skip = get_bits_count(gb) - 8 * (get_bits_count(gb) / 8);
+ n = get_bits_count(gb) / 8;
+
+ if (n > buf_size) {
+fail:
+ s->bitstream_size = 0;
+ s->bitstream_index = 0;
+ return ret;
+ }
+
+ frame->nb_samples = s->nb_samples;
+ if ((ret = ff_get_buffer(avctx, frame, 0)) < 0)
+ return ret;
+
+ switch (avctx->sample_fmt) {
+ case AV_SAMPLE_FMT_U8P:
+ for (int ch = 0; ch < avctx->ch_layout.nb_channels; ch++) {
+ uint8_t *dst = (uint8_t *)frame->extended_data[ch];
+ const int *src = s->samples[ch] + s->offset;
+
+ for (int n = 0; n < frame->nb_samples; n++)
+ dst[n] = src[n] * (1 << s->shift);
+ }
+ break;
+ case AV_SAMPLE_FMT_S16P:
+ for (int ch = 0; ch < avctx->ch_layout.nb_channels; ch++) {
+ int16_t *dst = (int16_t *)frame->extended_data[ch];
+ const int *src = s->samples[ch] + s->offset;
+
+ for (int n = 0; n < frame->nb_samples; n++)
+ dst[n] = src[n] * (1 << s->shift);
+ }
+ break;
+ }
+
+ *got_frame_ptr = 1;
+
+ if (s->bitstream_size) {
+ s->bitstream_index += n;
+ s->bitstream_size -= n;
+ return input_buf_size;
+ }
+
+ return n;
+}
+
+static av_cold int wavarc_close(AVCodecContext *avctx)
+{
+ WavArcContext *s = avctx->priv_data;
+
+ av_freep(&s->bitstream);
+ s->bitstream_size = 0;
+
+ return 0;
+}
+
+const FFCodec ff_wavarc_decoder = {
+ .p.name = "wavarc",
+ CODEC_LONG_NAME("Waveform Archiver"),
+ .p.type = AVMEDIA_TYPE_AUDIO,
+ .p.id = AV_CODEC_ID_WAVARC,
+ .priv_data_size = sizeof(WavArcContext),
+ .init = wavarc_init,
+ FF_CODEC_DECODE_CB(wavarc_decode),
+ .close = wavarc_close,
+ .p.capabilities = AV_CODEC_CAP_DR1 |
+ AV_CODEC_CAP_SUBFRAMES |
+ AV_CODEC_CAP_DELAY,
+ .p.sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_U8P,
+ AV_SAMPLE_FMT_S16P,
+ AV_SAMPLE_FMT_NONE },
+};