2 * Sierra VMD audio decoder
4 * This file is part of Libav.
6 * Libav is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Lesser General Public
8 * License as published by the Free Software Foundation; either
9 * version 2.1 of the License, or (at your option) any later version.
11 * Libav is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Lesser General Public License for more details.
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with Libav; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
23 * Sierra VMD audio decoder
24 * by Vladimir "VAG" Gneushev (vagsoft at mail.ru)
25 * for more information on the Sierra VMD format, visit:
26 * http://www.pcisys.net/~melanson/codecs/
28 * The audio decoder, expects each encoded data
29 * chunk to be prepended with the appropriate 16-byte frame information
30 * record from the VMD file. It does not require the 0x330-byte VMD file
31 * header, but it does need the audio setup parameters passed in through
32 * normal libavcodec API means.
37 #include "libavutil/channel_layout.h"
38 #include "libavutil/common.h"
39 #include "libavutil/intreadwrite.h"
44 #define BLOCK_TYPE_AUDIO 1
45 #define BLOCK_TYPE_INITIAL 2
46 #define BLOCK_TYPE_SILENCE 3
48 typedef struct VmdAudioContext {
53 static const uint16_t vmdaudio_table[128] = {
54 0x000, 0x008, 0x010, 0x020, 0x030, 0x040, 0x050, 0x060, 0x070, 0x080,
55 0x090, 0x0A0, 0x0B0, 0x0C0, 0x0D0, 0x0E0, 0x0F0, 0x100, 0x110, 0x120,
56 0x130, 0x140, 0x150, 0x160, 0x170, 0x180, 0x190, 0x1A0, 0x1B0, 0x1C0,
57 0x1D0, 0x1E0, 0x1F0, 0x200, 0x208, 0x210, 0x218, 0x220, 0x228, 0x230,
58 0x238, 0x240, 0x248, 0x250, 0x258, 0x260, 0x268, 0x270, 0x278, 0x280,
59 0x288, 0x290, 0x298, 0x2A0, 0x2A8, 0x2B0, 0x2B8, 0x2C0, 0x2C8, 0x2D0,
60 0x2D8, 0x2E0, 0x2E8, 0x2F0, 0x2F8, 0x300, 0x308, 0x310, 0x318, 0x320,
61 0x328, 0x330, 0x338, 0x340, 0x348, 0x350, 0x358, 0x360, 0x368, 0x370,
62 0x378, 0x380, 0x388, 0x390, 0x398, 0x3A0, 0x3A8, 0x3B0, 0x3B8, 0x3C0,
63 0x3C8, 0x3D0, 0x3D8, 0x3E0, 0x3E8, 0x3F0, 0x3F8, 0x400, 0x440, 0x480,
64 0x4C0, 0x500, 0x540, 0x580, 0x5C0, 0x600, 0x640, 0x680, 0x6C0, 0x700,
65 0x740, 0x780, 0x7C0, 0x800, 0x900, 0xA00, 0xB00, 0xC00, 0xD00, 0xE00,
66 0xF00, 0x1000, 0x1400, 0x1800, 0x1C00, 0x2000, 0x3000, 0x4000
69 static av_cold int vmdaudio_decode_init(AVCodecContext *avctx)
71 VmdAudioContext *s = avctx->priv_data;
73 if (avctx->channels < 1 || avctx->channels > 2) {
74 av_log(avctx, AV_LOG_ERROR, "invalid number of channels\n");
75 return AVERROR(EINVAL);
77 if (avctx->block_align < 1) {
78 av_log(avctx, AV_LOG_ERROR, "invalid block align\n");
79 return AVERROR(EINVAL);
82 avctx->channel_layout = avctx->channels == 1 ? AV_CH_LAYOUT_MONO :
85 if (avctx->bits_per_coded_sample == 16)
86 avctx->sample_fmt = AV_SAMPLE_FMT_S16;
88 avctx->sample_fmt = AV_SAMPLE_FMT_U8;
89 s->out_bps = av_get_bytes_per_sample(avctx->sample_fmt);
91 s->chunk_size = avctx->block_align + avctx->channels * (s->out_bps == 2);
93 av_log(avctx, AV_LOG_DEBUG, "%d channels, %d bits/sample, "
94 "block align = %d, sample rate = %d\n",
95 avctx->channels, avctx->bits_per_coded_sample, avctx->block_align,
101 static void decode_audio_s16(int16_t *out, const uint8_t *buf, int buf_size,
105 const uint8_t *buf_end = buf + buf_size;
107 int st = channels - 1;
109 /* decode initial raw sample */
110 for (ch = 0; ch < channels; ch++) {
111 predictor[ch] = (int16_t)AV_RL16(buf);
113 *out++ = predictor[ch];
116 /* decode DPCM samples */
118 while (buf < buf_end) {
121 predictor[ch] -= vmdaudio_table[b & 0x7F];
123 predictor[ch] += vmdaudio_table[b];
124 predictor[ch] = av_clip_int16(predictor[ch]);
125 *out++ = predictor[ch];
130 static int vmdaudio_decode_frame(AVCodecContext *avctx, void *data,
131 int *got_frame_ptr, AVPacket *avpkt)
133 AVFrame *frame = data;
134 const uint8_t *buf = avpkt->data;
135 const uint8_t *buf_end;
136 int buf_size = avpkt->size;
137 VmdAudioContext *s = avctx->priv_data;
138 int block_type, silent_chunks, audio_chunks;
140 uint8_t *output_samples_u8;
141 int16_t *output_samples_s16;
144 av_log(avctx, AV_LOG_WARNING, "skipping small junk packet\n");
150 if (block_type < BLOCK_TYPE_AUDIO || block_type > BLOCK_TYPE_SILENCE) {
151 av_log(avctx, AV_LOG_ERROR, "unknown block type: %d\n", block_type);
152 return AVERROR(EINVAL);
157 /* get number of silent chunks */
159 if (block_type == BLOCK_TYPE_INITIAL) {
162 av_log(avctx, AV_LOG_ERROR, "packet is too small\n");
163 return AVERROR(EINVAL);
165 flags = AV_RB32(buf);
166 silent_chunks = av_popcount(flags);
169 } else if (block_type == BLOCK_TYPE_SILENCE) {
171 buf_size = 0; // should already be zero but set it just to be sure
174 /* ensure output buffer is large enough */
175 audio_chunks = buf_size / s->chunk_size;
177 /* drop incomplete chunks */
178 buf_size = audio_chunks * s->chunk_size;
180 /* get output buffer */
181 frame->nb_samples = ((silent_chunks + audio_chunks) * avctx->block_align) /
183 if ((ret = ff_get_buffer(avctx, frame, 0)) < 0) {
184 av_log(avctx, AV_LOG_ERROR, "get_buffer() failed\n");
187 output_samples_u8 = frame->data[0];
188 output_samples_s16 = (int16_t *)frame->data[0];
190 /* decode silent chunks */
191 if (silent_chunks > 0) {
192 int silent_size = FFMIN(avctx->block_align * silent_chunks,
193 frame->nb_samples * avctx->channels);
194 if (s->out_bps == 2) {
195 memset(output_samples_s16, 0x00, silent_size * 2);
196 output_samples_s16 += silent_size;
198 memset(output_samples_u8, 0x80, silent_size);
199 output_samples_u8 += silent_size;
203 /* decode audio chunks */
204 if (audio_chunks > 0) {
205 buf_end = buf + (buf_size & ~(avctx->channels > 1));
206 while (buf + s->chunk_size <= buf_end) {
207 if (s->out_bps == 2) {
208 decode_audio_s16(output_samples_s16, buf, s->chunk_size,
210 output_samples_s16 += avctx->block_align;
212 memcpy(output_samples_u8, buf, s->chunk_size);
213 output_samples_u8 += avctx->block_align;
215 buf += s->chunk_size;
224 AVCodec ff_vmdaudio_decoder = {
226 .long_name = NULL_IF_CONFIG_SMALL("Sierra VMD audio"),
227 .type = AVMEDIA_TYPE_AUDIO,
228 .id = AV_CODEC_ID_VMDAUDIO,
229 .priv_data_size = sizeof(VmdAudioContext),
230 .init = vmdaudio_decode_init,
231 .decode = vmdaudio_decode_frame,
232 .capabilities = CODEC_CAP_DR1,