History

Merge remote-tracking branch 'qatar/master'

* qatar/master: Remove ffmpeg. aacenc: Simplify windowing aacenc: Move saved overlap samples to the beginning of the same buffer as incoming samples. aacenc: Deinterleave input samples before processing. aacenc: Store channel count in AACEncContext. aacenc: Move Q^3/4 calculation to it's own table aacenc: Request normalized float samples instead of converting s16 samples to float. aacpsy: Replace an if with FFMAX in LAME windowing. aacenc: cosmetics, replace 'rd' with 'bits' in codebook_trellis_rate to make it more clear what is being calculated. aacpsy: cosmetics, change a FIXME to a NOTE about subshort comparisons aacenc: cosmetics: move init() and end() to the bottom of the file. aacenc: aac_encode_init() cleanup XWD encoder and decoder vc1: don't read the interpfrm and bfraction elements for interlaced frames mxfdec: fix memleak on mxf_read_close() westwood: split the AUD and VQA demuxers into separate files. Conflicts: .gitignore Changelog Makefile configure doc/ffmpeg.texi ffmpeg.c libavcodec/Makefile libavcodec/aacenc.c libavcodec/allcodecs.c libavcodec/avcodec.h libavcodec/version.h libavformat/Makefile libavformat/img2.c

Michael Niedermayer committed 14 years ago

Commit 0bb57f82 parents b955d40 0fec2cb20 files · +1027 −342Browse files

Unified · Split

20 files
  1. Changelog +1 −0
  2. doc/general.texi +2 −0
  3. libavcodec/Makefile +2 −0
  4. libavcodec/aaccoder.c +18 −17
  5. libavcodec/aacenc.c +224 −130
  6. libavcodec/aacenc.h +8 −1
  7. libavcodec/aacpsy.c +16 −18
  8. libavcodec/allcodecs.c +1 −0
  9. libavcodec/avcodec.h +1 −0
  10. libavcodec/psymodel.c +6 −11
  11. libavcodec/psymodel.h +4 −8
  12. libavcodec/vc1.c +9 −7
  13. libavcodec/version.h +2 −2
  14. libavcodec/xwd.h +41 −0
  15. libavcodec/xwddec.c +267 −0
  16. libavcodec/xwdenc.c +246 −0
  17. libavformat/Makefile +2 −2
  18. libavformat/img2.c +2 −1
  19. libavformat/westwood_aud.c +173 −0
  20. libavformat/westwood_vqa.c +2 −145
Changelog+1−0
@@ -15,6 +15,7 @@ version next: · 14 unmodified lines
15- amerge audio filter15- amerge audio filter
16- GSM audio parser16- GSM audio parser
17- SMJPEG muxer17- SMJPEG muxer
18- XWD encoder and decoder
18- Automatic thread count based on detection number of (available) CPU cores19- Automatic thread count based on detection number of (available) CPU cores
19- y41p Brooktree Uncompressed 4:1:1 12-bit encoder and decoder20- y41p Brooktree Uncompressed 4:1:1 12-bit encoder and decoder
20- ffprobe -show_error option21- ffprobe -show_error option
doc/general.texi+2−0
@@ -399,6 +399,8 @@ following image formats are supported: · 398 unmodified lines
399 @tab YUV, JPEG and some extension is not supported yet.399 @tab YUV, JPEG and some extension is not supported yet.
400@item Truevision Targa @tab X @tab X400@item Truevision Targa @tab X @tab X
401 @tab Targa (.TGA) image format401 @tab Targa (.TGA) image format
402@item XWD @tab X @tab X
403 @tab X Window Dump image format
402@end multitable404@end multitable
403 405
404@code{X} means that encoding (resp. decoding) is supported.406@code{X} means that encoding (resp. decoding) is supported.
libavcodec/Makefile+2−0
@@ -470,6 +470,8 @@ OBJS-$(CONFIG_XBIN_DECODER) += bintext.o cga_data.o · 469 unmodified lines
470OBJS-$(CONFIG_XL_DECODER) += xl.o470OBJS-$(CONFIG_XL_DECODER) += xl.o
471OBJS-$(CONFIG_XSUB_DECODER) += xsubdec.o471OBJS-$(CONFIG_XSUB_DECODER) += xsubdec.o
472OBJS-$(CONFIG_XSUB_ENCODER) += xsubenc.o472OBJS-$(CONFIG_XSUB_ENCODER) += xsubenc.o
473OBJS-$(CONFIG_XWD_DECODER) += xwddec.o
474OBJS-$(CONFIG_XWD_ENCODER) += xwdenc.o
473OBJS-$(CONFIG_Y41P_DECODER) += y41pdec.o475OBJS-$(CONFIG_Y41P_DECODER) += y41pdec.o
474OBJS-$(CONFIG_Y41P_ENCODER) += y41penc.o476OBJS-$(CONFIG_Y41P_ENCODER) += y41penc.o
475OBJS-$(CONFIG_YOP_DECODER) += yop.o477OBJS-$(CONFIG_YOP_DECODER) += yop.o
libavcodec/aaccoder.c+18−17
@@ -110,14 +110,15 @@ static av_always_inline float quantize_and_encode_band_cost_template( · 109 unmodified lines
110 int *bits, int BT_ZERO, int BT_UNSIGNED,110 int *bits, int BT_ZERO, int BT_UNSIGNED,
111 int BT_PAIR, int BT_ESC)111 int BT_PAIR, int BT_ESC)
112{112{
113 const float IQ = ff_aac_pow2sf_tab[POW_SF2_ZERO + scale_idx - SCALE_ONE_POS + SCALE_DIV_512];113 const int q_idx = POW_SF2_ZERO - scale_idx + SCALE_ONE_POS - SCALE_DIV_512;
114 const float Q = ff_aac_pow2sf_tab[POW_SF2_ZERO - scale_idx + SCALE_ONE_POS - SCALE_DIV_512];114 const float Q = ff_aac_pow2sf_tab [q_idx];
115 const float Q34 = ff_aac_pow34sf_tab[q_idx];
116 const float IQ = ff_aac_pow2sf_tab [POW_SF2_ZERO + scale_idx - SCALE_ONE_POS + SCALE_DIV_512];
115 const float CLIPPED_ESCAPE = 165140.0f*IQ;117 const float CLIPPED_ESCAPE = 165140.0f*IQ;
116 int i, j;118 int i, j;
117 float cost = 0;119 float cost = 0;
118 const int dim = BT_PAIR ? 2 : 4;120 const int dim = BT_PAIR ? 2 : 4;
119 int resbits = 0;121 int resbits = 0;
120 const float Q34 = sqrtf(Q * sqrtf(Q));
121 const int range = aac_cb_range[cb];122 const int range = aac_cb_range[cb];
122 const int maxval = aac_cb_maxval[cb];123 const int maxval = aac_cb_maxval[cb];
123 int off;124 int off;
@@ -420,7 +421,7 @@ static void codebook_trellis_rate(AACEncContext *s, SingleChannelElement *sce, · 296 unmodified lines
420 const int run_esc = (1 << run_bits) - 1;421 const int run_esc = (1 << run_bits) - 1;
421 int idx, ppos, count;422 int idx, ppos, count;
422 int stackrun[120], stackcb[120], stack_len;423 int stackrun[120], stackcb[120], stack_len;
423 float next_minrd = INFINITY;424 float next_minbits = INFINITY;
424 int next_mincb = 0;425 int next_mincb = 0;
425 426
426 abs_pow34_v(s->scoefs, sce->coeffs, 1024);427 abs_pow34_v(s->scoefs, sce->coeffs, 1024);
@@ -434,7 +435,7 @@ static void codebook_trellis_rate(AACEncContext *s, SingleChannelElement *sce, · 7 unmodified lines
434 size = sce->ics.swb_sizes[swb];435 size = sce->ics.swb_sizes[swb];
435 if (sce->zeroes[win*16 + swb]) {436 if (sce->zeroes[win*16 + swb]) {
436 float cost_stay_here = path[swb][0].cost;437 float cost_stay_here = path[swb][0].cost;
437 float cost_get_here = next_minrd + run_bits + 4;438 float cost_get_here = next_minbits + run_bits + 4;
438 if ( run_value_bits[sce->ics.num_windows == 8][path[swb][0].run]439 if ( run_value_bits[sce->ics.num_windows == 8][path[swb][0].run]
439 != run_value_bits[sce->ics.num_windows == 8][path[swb][0].run+1])440 != run_value_bits[sce->ics.num_windows == 8][path[swb][0].run+1])
440 cost_stay_here += run_bits;441 cost_stay_here += run_bits;
@@ -447,7 +448,7 @@ static void codebook_trellis_rate(AACEncContext *s, SingleChannelElement *sce, · 6 unmodified lines
447 path[swb+1][0].cost = cost_stay_here;448 path[swb+1][0].cost = cost_stay_here;
448 path[swb+1][0].run = path[swb][0].run + 1;449 path[swb+1][0].run = path[swb][0].run + 1;
449 }450 }
450 next_minrd = path[swb+1][0].cost;451 next_minbits = path[swb+1][0].cost;
451 next_mincb = 0;452 next_mincb = 0;
452 for (cb = 1; cb < 12; cb++) {453 for (cb = 1; cb < 12; cb++) {
453 path[swb+1][cb].cost = 61450;454 path[swb+1][cb].cost = 61450;
@@ -455,10 +456,10 @@ static void codebook_trellis_rate(AACEncContext *s, SingleChannelElement *sce, · 1 unmodified line
455 path[swb+1][cb].run = 0;456 path[swb+1][cb].run = 0;
456 }457 }
457 } else {458 } else {
458 float minrd = next_minrd;459 float minbits = next_minbits;
459 int mincb = next_mincb;460 int mincb = next_mincb;
460 int startcb = sce->band_type[win*16+swb];461 int startcb = sce->band_type[win*16+swb];
461 next_minrd = INFINITY;462 next_minbits = INFINITY;
462 next_mincb = 0;463 next_mincb = 0;
463 for (cb = 0; cb < startcb; cb++) {464 for (cb = 0; cb < startcb; cb++) {
464 path[swb+1][cb].cost = 61450;465 path[swb+1][cb].cost = 61450;
@@ -467,15 +468,15 @@ static void codebook_trellis_rate(AACEncContext *s, SingleChannelElement *sce, · 2 unmodified lines
467 }468 }
468 for (cb = startcb; cb < 12; cb++) {469 for (cb = startcb; cb < 12; cb++) {
469 float cost_stay_here, cost_get_here;470 float cost_stay_here, cost_get_here;
470 float rd = 0.0f;471 float bits = 0.0f;
471 for (w = 0; w < group_len; w++) {472 for (w = 0; w < group_len; w++) {
472 rd += quantize_band_cost(s, sce->coeffs + start + w*128,473 bits += quantize_band_cost(s, sce->coeffs + start + w*128,
473 s->scoefs + start + w*128, size,474 s->scoefs + start + w*128, size,
474 sce->sf_idx[(win+w)*16+swb], cb,475 sce->sf_idx[(win+w)*16+swb], cb,
475 0, INFINITY, NULL);476 0, INFINITY, NULL);
476 }477 }
477 cost_stay_here = path[swb][cb].cost + rd;478 cost_stay_here = path[swb][cb].cost + bits;
478 cost_get_here = minrd + rd + run_bits + 4;479 cost_get_here = minbits + bits + run_bits + 4;
479 if ( run_value_bits[sce->ics.num_windows == 8][path[swb][cb].run]480 if ( run_value_bits[sce->ics.num_windows == 8][path[swb][cb].run]
480 != run_value_bits[sce->ics.num_windows == 8][path[swb][cb].run+1])481 != run_value_bits[sce->ics.num_windows == 8][path[swb][cb].run+1])
481 cost_stay_here += run_bits;482 cost_stay_here += run_bits;
@@ -488,8 +489,8 @@ static void codebook_trellis_rate(AACEncContext *s, SingleChannelElement *sce, · 6 unmodified lines
488 path[swb+1][cb].cost = cost_stay_here;489 path[swb+1][cb].cost = cost_stay_here;
489 path[swb+1][cb].run = path[swb][cb].run + 1;490 path[swb+1][cb].run = path[swb][cb].run + 1;
490 }491 }
491 if (path[swb+1][cb].cost < next_minrd) {492 if (path[swb+1][cb].cost < next_minbits) {
492 next_minrd = path[swb+1][cb].cost;493 next_minbits = path[swb+1][cb].cost;
493 next_mincb = cb;494 next_mincb = cb;
494 }495 }
495 }496 }
libavcodec/aacenc.c+224−130
@@ -46,6 +46,14 @@ · 45 unmodified lines
46 46
47#define AAC_MAX_CHANNELS 647#define AAC_MAX_CHANNELS 6
48 48
49#define ERROR_IF(cond, ...) \
50 if (cond) { \
51 av_log(avctx, AV_LOG_ERROR, __VA_ARGS__); \
52 return AVERROR(EINVAL); \
53 }
54
55float ff_aac_pow34sf_tab[428];
56
49static const uint8_t swb_size_1024_96[] = {57static const uint8_t swb_size_1024_96[] = {
50 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 8, 8, 8, 8, 8,58 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 8, 8, 8, 8, 8,
51 12, 12, 12, 12, 12, 16, 16, 24, 28, 36, 44,59 12, 12, 12, 12, 12, 16, 16, 24, 28, 36, 44,
@@ -135,6 +143,18 @@ static const uint8_t aac_chan_configs[6][5] = { · 83 unmodified lines
135 {4, TYPE_SCE, TYPE_CPE, TYPE_CPE, TYPE_LFE}, // 6 channels - front center + stereo + back stereo + LFE143 {4, TYPE_SCE, TYPE_CPE, TYPE_CPE, TYPE_LFE}, // 6 channels - front center + stereo + back stereo + LFE
136};144};
137 145
146/**
147 * Table to remap channels from Libav's default order to AAC order.
148 */
149static const uint8_t aac_chan_maps[AAC_MAX_CHANNELS][AAC_MAX_CHANNELS] = {
150 { 0 },
151 { 0, 1 },
152 { 2, 0, 1 },
153 { 2, 0, 1, 3 },
154 { 2, 0, 1, 3, 4 },
155 { 2, 0, 1, 4, 5, 3 },
156};
157
138/**158/**
139 * Make AAC audio config object.159 * Make AAC audio config object.
140 * @see 1.6.2.1 "Syntax - AudioSpecificConfig"160 * @see 1.6.2.1 "Syntax - AudioSpecificConfig"
@@ -147,7 +167,7 @@ static void put_audio_specific_config(AVCodecContext *avctx) · 6 unmodified lines
147 init_put_bits(&pb, avctx->extradata, avctx->extradata_size*8);167 init_put_bits(&pb, avctx->extradata, avctx->extradata_size*8);
148 put_bits(&pb, 5, 2); //object type - AAC-LC168 put_bits(&pb, 5, 2); //object type - AAC-LC
149 put_bits(&pb, 4, s->samplerate_index); //sample rate index169 put_bits(&pb, 4, s->samplerate_index); //sample rate index
150 put_bits(&pb, 4, avctx->channels);170 put_bits(&pb, 4, s->channels);
151 //GASpecificConfig171 //GASpecificConfig
152 put_bits(&pb, 1, 0); //frame length - 1024 samples172 put_bits(&pb, 1, 0); //frame length - 1024 samples
153 put_bits(&pb, 1, 0); //does not depend on core coder173 put_bits(&pb, 1, 0); //does not depend on core coder
@@ -160,117 +180,80 @@ static void put_audio_specific_config(AVCodecContext *avctx) · 6 unmodified lines
160 flush_put_bits(&pb);180 flush_put_bits(&pb);
161}181}
162 182
163static av_cold int aac_encode_init(AVCodecContext *avctx)183#define WINDOW_FUNC(type) \
164{184static void apply_ ##type ##_window(DSPContext *dsp, SingleChannelElement *sce, const float *audio)
165 AACEncContext *s = avctx->priv_data;
166 int i;
167 const uint8_t *sizes[2];
168 uint8_t grouping[AAC_MAX_CHANNELS];
169 int lengths[2];
170
171 avctx->frame_size = 1024;
172 185
173 for (i = 0; i < 16; i++)186WINDOW_FUNC(only_long)
174 if (avctx->sample_rate == avpriv_mpeg4audio_sample_rates[i])187{
175 break;188 const float *lwindow = sce->ics.use_kb_window[0] ? ff_aac_kbd_long_1024 : ff_sine_1024;
176 if (i == 16) {189 const float *pwindow = sce->ics.use_kb_window[1] ? ff_aac_kbd_long_1024 : ff_sine_1024;
177 av_log(avctx, AV_LOG_ERROR, "Unsupported sample rate %d\n", avctx->sample_rate);190 float *out = sce->ret;
178 return -1;
179 }
180 if (avctx->channels > AAC_MAX_CHANNELS) {
181 av_log(avctx, AV_LOG_ERROR, "Unsupported number of channels: %d\n", avctx->channels);
182 return -1;
183 }
184 if (avctx->profile != FF_PROFILE_UNKNOWN && avctx->profile != FF_PROFILE_AAC_LOW) {
185 av_log(avctx, AV_LOG_ERROR, "Unsupported profile %d\n", avctx->profile);
186 return -1;
187 }
188 if (1024.0 * avctx->bit_rate / avctx->sample_rate > 6144 * avctx->channels) {
189 av_log(avctx, AV_LOG_ERROR, "Too many bits per frame requested\n");
190 return -1;
191 }
192 s->samplerate_index = i;
193
194 dsputil_init(&s->dsp, avctx);
195 ff_mdct_init(&s->mdct1024, 11, 0, 1.0);
196 ff_mdct_init(&s->mdct128, 8, 0, 1.0);
197 // window init
198 ff_kbd_window_init(ff_aac_kbd_long_1024, 4.0, 1024);
199 ff_kbd_window_init(ff_aac_kbd_short_128, 6.0, 128);
200 ff_init_ff_sine_windows(10);
201 ff_init_ff_sine_windows(7);
202
203 s->chan_map = aac_chan_configs[avctx->channels-1];
204 s->samples = av_malloc(2 * 1024 * avctx->channels * sizeof(s->samples[0]));
205 s->cpe = av_mallocz(sizeof(ChannelElement) * s->chan_map[0]);
206 avctx->extradata = av_mallocz(5 + FF_INPUT_BUFFER_PADDING_SIZE);
207 avctx->extradata_size = 5;
208 put_audio_specific_config(avctx);
209 191
210 sizes[0] = swb_size_1024[i];192 dsp->vector_fmul (out, audio, lwindow, 1024);
211 sizes[1] = swb_size_128[i];193 dsp->vector_fmul_reverse(out + 1024, audio + 1024, pwindow, 1024);
212 lengths[0] = ff_aac_num_swb_1024[i];194}
213 lengths[1] = ff_aac_num_swb_128[i];
214 for (i = 0; i < s->chan_map[0]; i++)
215 grouping[i] = s->chan_map[i + 1] == TYPE_CPE;
216 ff_psy_init(&s->psy, avctx, 2, sizes, lengths, s->chan_map[0], grouping);
217 s->psypp = ff_psy_preprocess_init(avctx);
218 s->coder = &ff_aac_coders[s->options.aac_coder];
219 195
220 s->lambda = avctx->global_quality ? avctx->global_quality : 120;196WINDOW_FUNC(long_start)
197{
198 const float *lwindow = sce->ics.use_kb_window[1] ? ff_aac_kbd_long_1024 : ff_sine_1024;
199 const float *swindow = sce->ics.use_kb_window[0] ? ff_aac_kbd_short_128 : ff_sine_128;
200 float *out = sce->ret;
201
202 dsp->vector_fmul(out, audio, lwindow, 1024);
203 memcpy(out + 1024, audio, sizeof(out[0]) * 448);
204 dsp->vector_fmul_reverse(out + 1024 + 448, audio, swindow, 128);
205 memset(out + 1024 + 576, 0, sizeof(out[0]) * 448);
206}
221 207
222 ff_aac_tableinit();208WINDOW_FUNC(long_stop)
209{
210 const float *lwindow = sce->ics.use_kb_window[0] ? ff_aac_kbd_long_1024 : ff_sine_1024;
211 const float *swindow = sce->ics.use_kb_window[1] ? ff_aac_kbd_short_128 : ff_sine_128;
212 float *out = sce->ret;
213
214 memset(out, 0, sizeof(out[0]) * 448);
215 dsp->vector_fmul(out + 448, audio + 448, swindow, 128);
216 memcpy(out + 576, audio + 576, sizeof(out[0]) * 448);
217 dsp->vector_fmul_reverse(out + 1024, audio + 1024, lwindow, 1024);
218}
223 219
224 return 0;220WINDOW_FUNC(eight_short)
221{
222 const float *swindow = sce->ics.use_kb_window[0] ? ff_aac_kbd_short_128 : ff_sine_128;
223 const float *pwindow = sce->ics.use_kb_window[1] ? ff_aac_kbd_short_128 : ff_sine_128;
224 const float *in = audio + 448;
225 float *out = sce->ret;
226
227 for (int w = 0; w < 8; w++) {
228 dsp->vector_fmul (out, in, w ? pwindow : swindow, 128);
229 out += 128;
230 in += 128;
231 dsp->vector_fmul_reverse(out, in, swindow, 128);
232 out += 128;
233 }
225}234}
226 235
227static void apply_window_and_mdct(AVCodecContext *avctx, AACEncContext *s,236static void (*const apply_window[4])(DSPContext *dsp, SingleChannelElement *sce, const float *audio) = {
228 SingleChannelElement *sce, short *audio)237 [ONLY_LONG_SEQUENCE] = apply_only_long_window,
238 [LONG_START_SEQUENCE] = apply_long_start_window,
239 [EIGHT_SHORT_SEQUENCE] = apply_eight_short_window,
240 [LONG_STOP_SEQUENCE] = apply_long_stop_window
241};
242
243static void apply_window_and_mdct(AACEncContext *s, SingleChannelElement *sce,
244 float *audio)
229{245{
230 int i, k;246 int i;
231 const int chans = avctx->channels;
232 const float * lwindow = sce->ics.use_kb_window[0] ? ff_aac_kbd_long_1024 : ff_sine_1024;
233 const float * swindow = sce->ics.use_kb_window[0] ? ff_aac_kbd_short_128 : ff_sine_128;
234 const float * pwindow = sce->ics.use_kb_window[1] ? ff_aac_kbd_short_128 : ff_sine_128;
235 float *output = sce->ret;247 float *output = sce->ret;
236 248
237 if (sce->ics.window_sequence[0] != EIGHT_SHORT_SEQUENCE) {249 apply_window[sce->ics.window_sequence[0]](&s->dsp, sce, audio);
238 memcpy(output, sce->saved, sizeof(float)*1024);250
239 if (sce->ics.window_sequence[0] == LONG_STOP_SEQUENCE) {251 if (sce->ics.window_sequence[0] != EIGHT_SHORT_SEQUENCE)
240 memset(output, 0, sizeof(output[0]) * 448);
241 for (i = 448; i < 576; i++)
242 output[i] = sce->saved[i] * pwindow[i - 448];
243 for (i = 576; i < 704; i++)
244 output[i] = sce->saved[i];
245 }
246 if (sce->ics.window_sequence[0] != LONG_START_SEQUENCE) {
247 for (i = 0; i < 1024; i++) {
248 output[i+1024] = audio[i * chans] * lwindow[1024 - i - 1];
249 sce->saved[i] = audio[i * chans] * lwindow[i];
250 }
251 } else {
252 for (i = 0; i < 448; i++)
253 output[i+1024] = audio[i * chans];
254 for (; i < 576; i++)
255 output[i+1024] = audio[i * chans] * swindow[576 - i - 1];
256 memset(output+1024+576, 0, sizeof(output[0]) * 448);
257 for (i = 0; i < 1024; i++)
258 sce->saved[i] = audio[i * chans];
259 }
260 s->mdct1024.mdct_calc(&s->mdct1024, sce->coeffs, output);252 s->mdct1024.mdct_calc(&s->mdct1024, sce->coeffs, output);
261 } else {253 else
262 for (k = 0; k < 1024; k += 128) {254 for (i = 0; i < 1024; i += 128)
263 for (i = 448 + k; i < 448 + k + 256; i++)255 s->mdct128.mdct_calc(&s->mdct128, sce->coeffs + i, output + i*2);
264 output[i - 448 - k] = (i < 1024)256 memcpy(audio, audio + 1024, sizeof(audio[0]) * 1024);
265 ? sce->saved[i]
266 : audio[(i-1024)*chans];
267 s->dsp.vector_fmul (output, output, k ? swindow : pwindow, 128);
268 s->dsp.vector_fmul_reverse(output+128, output+128, swindow, 128);
269 s->mdct128.mdct_calc(&s->mdct128, sce->coeffs + k, output);
270 }
271 for (i = 0; i < 1024; i++)
272 sce->saved[i] = audio[i * chans];
273 }
274}257}
275 258
276/**259/**
@@ -488,11 +471,37 @@ static void put_bitstream_info(AVCodecContext *avctx, AACEncContext *s, · 211 unmodified lines
488 put_bits(&s->pb, 12 - padbits, 0);471 put_bits(&s->pb, 12 - padbits, 0);
489}472}
490 473
474/*
475 * Deinterleave input samples.
476 * Channels are reordered from Libav's default order to AAC order.
477 */
478static void deinterleave_input_samples(AACEncContext *s,
479 const float *samples)
480{
481 int ch, i;
482 const int sinc = s->channels;
483 const uint8_t *channel_map = aac_chan_maps[sinc - 1];
484
485 /* deinterleave and remap input samples */
486 for (ch = 0; ch < sinc; ch++) {
487 const float *sptr = samples + channel_map[ch];
488
489 /* copy last 1024 samples of previous frame to the start of the current frame */
490 memcpy(&s->planar_samples[ch][0], &s->planar_samples[ch][1024], 1024 * sizeof(s->planar_samples[0][0]));
491
492 /* deinterleave */
493 for (i = 1024; i < 1024 * 2; i++) {
494 s->planar_samples[ch][i] = *sptr;
495 sptr += sinc;
496 }
497 }
498}
499
491static int aac_encode_frame(AVCodecContext *avctx,500static int aac_encode_frame(AVCodecContext *avctx,
492 uint8_t *frame, int buf_size, void *data)501 uint8_t *frame, int buf_size, void *data)
493{502{
494 AACEncContext *s = avctx->priv_data;503 AACEncContext *s = avctx->priv_data;
495 int16_t *samples = s->samples, *samples2, *la;504 float **samples = s->planar_samples, *samples2, *la, *overlap;
496 ChannelElement *cpe;505 ChannelElement *cpe;
497 int i, ch, w, g, chans, tag, start_ch;506 int i, ch, w, g, chans, tag, start_ch;
498 int chan_el_counter[4];507 int chan_el_counter[4];
@@ -500,27 +509,15 @@ static int aac_encode_frame(AVCodecContext *avctx, · 1 unmodified line
500 509
501 if (s->last_frame)510 if (s->last_frame)
502 return 0;511 return 0;
512
503 if (data) {513 if (data) {
504 if (!s->psypp) {514 deinterleave_input_samples(s, data);
505 memcpy(s->samples + 1024 * avctx->channels, data,515 if (s->psypp)
506 1024 * avctx->channels * sizeof(s->samples[0]));516 ff_psy_preprocess(s->psypp, s->planar_samples, s->channels);
507 } else {
508 start_ch = 0;
509 samples2 = s->samples + 1024 * avctx->channels;
510 for (i = 0; i < s->chan_map[0]; i++) {
511 tag = s->chan_map[i+1];
512 chans = tag == TYPE_CPE ? 2 : 1;
513 ff_psy_preprocess(s->psypp, (uint16_t*)data + start_ch,
514 samples2 + start_ch, start_ch, chans);
515 start_ch += chans;
516 }
517 }
518 }517 }
519 if (!avctx->frame_number) {518
520 memcpy(s->samples, s->samples + 1024 * avctx->channels,519 if (!avctx->frame_number)
521 1024 * avctx->channels * sizeof(s->samples[0]));
522 return 0;520 return 0;
523 }
524 521
525 start_ch = 0;522 start_ch = 0;
526 for (i = 0; i < s->chan_map[0]; i++) {523 for (i = 0; i < s->chan_map[0]; i++) {
@@ -531,8 +528,9 @@ static int aac_encode_frame(AVCodecContext *avctx, · 4 unmodified lines
531 for (ch = 0; ch < chans; ch++) {528 for (ch = 0; ch < chans; ch++) {
532 IndividualChannelStream *ics = &cpe->ch[ch].ics;529 IndividualChannelStream *ics = &cpe->ch[ch].ics;
533 int cur_channel = start_ch + ch;530 int cur_channel = start_ch + ch;
534 samples2 = samples + cur_channel;531 overlap = &samples[cur_channel][0];
535 la = samples2 + (448+64) * avctx->channels;532 samples2 = overlap + 1024;
533 la = samples2 + (448+64);
536 if (!data)534 if (!data)
537 la = NULL;535 la = NULL;
538 if (tag == TYPE_LFE) {536 if (tag == TYPE_LFE) {
@@ -560,7 +558,7 @@ static int aac_encode_frame(AVCodecContext *avctx, · 21 unmodified lines
560 for (w = 0; w < ics->num_windows; w++)558 for (w = 0; w < ics->num_windows; w++)
561 ics->group_len[w] = wi[ch].grouping[w];559 ics->group_len[w] = wi[ch].grouping[w];
562 560
563 apply_window_and_mdct(avctx, s, &cpe->ch[ch], samples2);561 apply_window_and_mdct(s, &cpe->ch[ch], overlap);
564 }562 }
565 start_ch += chans;563 start_ch += chans;
566 }564 }
@@ -626,8 +624,8 @@ static int aac_encode_frame(AVCodecContext *avctx, · 59 unmodified lines
626 }624 }
627 625
628 frame_bits = put_bits_count(&s->pb);626 frame_bits = put_bits_count(&s->pb);
629 if (frame_bits <= 6144 * avctx->channels - 3) {627 if (frame_bits <= 6144 * s->channels - 3) {
630 s->psy.bitres.bits = frame_bits / avctx->channels;628 s->psy.bitres.bits = frame_bits / s->channels;
631 break;629 break;
632 }630 }
633 631
@@ -648,8 +646,7 @@ static int aac_encode_frame(AVCodecContext *avctx, · 14 unmodified lines
648 646
649 if (!data)647 if (!data)
650 s->last_frame = 1;648 s->last_frame = 1;
651 memcpy(s->samples, s->samples + 1024 * avctx->channels,649
652 1024 * avctx->channels * sizeof(s->samples[0]));
653 return put_bits_count(&s->pb)>>3;650 return put_bits_count(&s->pb)>>3;
654}651}
655 652
@@ -660,12 +657,109 @@ static av_cold int aac_encode_end(AVCodecContext *avctx) · 4 unmodified lines
660 ff_mdct_end(&s->mdct1024);657 ff_mdct_end(&s->mdct1024);
661 ff_mdct_end(&s->mdct128);658 ff_mdct_end(&s->mdct128);
662 ff_psy_end(&s->psy);659 ff_psy_end(&s->psy);
663 ff_psy_preprocess_end(s->psypp);660 if (s->psypp)
664 av_freep(&s->samples);661 ff_psy_preprocess_end(s->psypp);
662 av_freep(&s->buffer.samples);
665 av_freep(&s->cpe);663 av_freep(&s->cpe);
666 return 0;664 return 0;
667}665}
668 666
667static av_cold int dsp_init(AVCodecContext *avctx, AACEncContext *s)
668{
669 int ret = 0;
670
671 dsputil_init(&s->dsp, avctx);
672
673 // window init
674 ff_kbd_window_init(ff_aac_kbd_long_1024, 4.0, 1024);
675 ff_kbd_window_init(ff_aac_kbd_short_128, 6.0, 128);
676 ff_init_ff_sine_windows(10);
677 ff_init_ff_sine_windows(7);
678
679 if (ret = ff_mdct_init(&s->mdct1024, 11, 0, 32768.0))
680 return ret;
681 if (ret = ff_mdct_init(&s->mdct128, 8, 0, 32768.0))
682 return ret;
683
684 return 0;
685}
686
687static av_cold int alloc_buffers(AVCodecContext *avctx, AACEncContext *s)
688{
689 FF_ALLOCZ_OR_GOTO(avctx, s->buffer.samples, 3 * 1024 * s->channels * sizeof(s->buffer.samples[0]), alloc_fail);
690 FF_ALLOCZ_OR_GOTO(avctx, s->cpe, sizeof(ChannelElement) * s->chan_map[0], alloc_fail);
691 FF_ALLOCZ_OR_GOTO(avctx, avctx->extradata, 5 + FF_INPUT_BUFFER_PADDING_SIZE, alloc_fail);
692
693 for(int ch = 0; ch < s->channels; ch++)
694 s->planar_samples[ch] = s->buffer.samples + 3 * 1024 * ch;
695
696 return 0;
697alloc_fail:
698 return AVERROR(ENOMEM);
699}
700
701static av_cold int aac_encode_init(AVCodecContext *avctx)
702{
703 AACEncContext *s = avctx->priv_data;
704 int i, ret = 0;
705 const uint8_t *sizes[2];
706 uint8_t grouping[AAC_MAX_CHANNELS];
707 int lengths[2];
708
709 avctx->frame_size = 1024;
710
711 for (i = 0; i < 16; i++)
712 if (avctx->sample_rate == avpriv_mpeg4audio_sample_rates[i])
713 break;
714
715 s->channels = avctx->channels;
716
717 ERROR_IF(i == 16,
718 "Unsupported sample rate %d\n", avctx->sample_rate);
719 ERROR_IF(s->channels > AAC_MAX_CHANNELS,
720 "Unsupported number of channels: %d\n", s->channels);
721 ERROR_IF(avctx->profile != FF_PROFILE_UNKNOWN && avctx->profile != FF_PROFILE_AAC_LOW,
722 "Unsupported profile %d\n", avctx->profile);
723 ERROR_IF(1024.0 * avctx->bit_rate / avctx->sample_rate > 6144 * s->channels,
724 "Too many bits per frame requested\n");
725
726 s->samplerate_index = i;
727
728 s->chan_map = aac_chan_configs[s->channels-1];
729
730 if (ret = dsp_init(avctx, s))
731 goto fail;
732
733 if (ret = alloc_buffers(avctx, s))
734 goto fail;
735
736 avctx->extradata_size = 5;
737 put_audio_specific_config(avctx);
738
739 sizes[0] = swb_size_1024[i];
740 sizes[1] = swb_size_128[i];
741 lengths[0] = ff_aac_num_swb_1024[i];
742 lengths[1] = ff_aac_num_swb_128[i];
743 for (i = 0; i < s->chan_map[0]; i++)
744 grouping[i] = s->chan_map[i + 1] == TYPE_CPE;
745 if (ret = ff_psy_init(&s->psy, avctx, 2, sizes, lengths, s->chan_map[0], grouping))
746 goto fail;
747 s->psypp = ff_psy_preprocess_init(avctx);
748 s->coder = &ff_aac_coders[s->options.aac_coder];
749
750 s->lambda = avctx->global_quality ? avctx->global_quality : 120;
751
752 ff_aac_tableinit();
753
754 for (i = 0; i < 428; i++)
755 ff_aac_pow34sf_tab[i] = sqrt(ff_aac_pow2sf_tab[i] * sqrt(ff_aac_pow2sf_tab[i]));
756
757 return 0;
758fail:
759 aac_encode_end(avctx);
760 return ret;
761}
762
669#define AACENC_FLAGS AV_OPT_FLAG_ENCODING_PARAM | AV_OPT_FLAG_AUDIO_PARAM763#define AACENC_FLAGS AV_OPT_FLAG_ENCODING_PARAM | AV_OPT_FLAG_AUDIO_PARAM
670static const AVOption aacenc_options[] = {764static const AVOption aacenc_options[] = {
671 {"stereo_mode", "Stereo coding method", offsetof(AACEncContext, options.stereo_mode), AV_OPT_TYPE_INT, {.dbl = 0}, -1, 1, AACENC_FLAGS, "stereo_mode"},765 {"stereo_mode", "Stereo coding method", offsetof(AACEncContext, options.stereo_mode), AV_OPT_TYPE_INT, {.dbl = 0}, -1, 1, AACENC_FLAGS, "stereo_mode"},
@@ -692,7 +786,7 @@ AVCodec ff_aac_encoder = { · 20 unmodified lines
692 .encode = aac_encode_frame,786 .encode = aac_encode_frame,
693 .close = aac_encode_end,787 .close = aac_encode_end,
694 .capabilities = CODEC_CAP_SMALL_LAST_FRAME | CODEC_CAP_DELAY | CODEC_CAP_EXPERIMENTAL,788 .capabilities = CODEC_CAP_SMALL_LAST_FRAME | CODEC_CAP_DELAY | CODEC_CAP_EXPERIMENTAL,
695 .sample_fmts = (const enum AVSampleFormat[]){AV_SAMPLE_FMT_S16,AV_SAMPLE_FMT_NONE},789 .sample_fmts = (const enum AVSampleFormat[]){AV_SAMPLE_FMT_FLT,AV_SAMPLE_FMT_NONE},
696 .long_name = NULL_IF_CONFIG_SMALL("Advanced Audio Coding"),790 .long_name = NULL_IF_CONFIG_SMALL("Advanced Audio Coding"),
697 .priv_class = &aacenc_class,791 .priv_class = &aacenc_class,
698};792};
libavcodec/aacenc.h+8−1
@@ -61,9 +61,10 @@ typedef struct AACEncContext { · 60 unmodified lines
61 FFTContext mdct1024; ///< long (1024 samples) frame transform context61 FFTContext mdct1024; ///< long (1024 samples) frame transform context
62 FFTContext mdct128; ///< short (128 samples) frame transform context62 FFTContext mdct128; ///< short (128 samples) frame transform context
63 DSPContext dsp;63 DSPContext dsp;
64 int16_t *samples; ///< saved preprocessed input64 float *planar_samples[6]; ///< saved preprocessed input
65 65
66 int samplerate_index; ///< MPEG-4 samplerate index66 int samplerate_index; ///< MPEG-4 samplerate index
67 int channels; ///< channel count
67 const uint8_t *chan_map; ///< channel configuration map68 const uint8_t *chan_map; ///< channel configuration map
68 69
69 ChannelElement *cpe; ///< channel elements70 ChannelElement *cpe; ///< channel elements
@@ -75,6 +76,12 @@ typedef struct AACEncContext { · 5 unmodified lines
75 float lambda;76 float lambda;
76 DECLARE_ALIGNED(16, int, qcoefs)[96]; ///< quantized coefficients77 DECLARE_ALIGNED(16, int, qcoefs)[96]; ///< quantized coefficients
77 DECLARE_ALIGNED(32, float, scoefs)[1024]; ///< scaled coefficients78 DECLARE_ALIGNED(32, float, scoefs)[1024]; ///< scaled coefficients
79
80 struct {
81 float *samples;
82 } buffer;
78} AACEncContext;83} AACEncContext;
79 84
85extern float ff_aac_pow34sf_tab[428];
86
80#endif /* AVCODEC_AACENC_H */87#endif /* AVCODEC_AACENC_H */
libavcodec/aacpsy.c+16−18
@@ -400,7 +400,7 @@ static av_unused FFPsyWindowInfo psy_3gpp_window(FFPsyContext *ctx, · 399 unmodified lines
400 int stay_short = 0;400 int stay_short = 0;
401 for (i = 0; i < 8; i++) {401 for (i = 0; i < 8; i++) {
402 for (j = 0; j < 128; j++) {402 for (j = 0; j < 128; j++) {
403 v = iir_filter(la[(i*128+j)*ctx->avctx->channels], pch->iir_state);403 v = iir_filter(la[i*128+j], pch->iir_state);
404 sum += v*v;404 sum += v*v;
405 }405 }
406 s[i] = sum;406 s[i] = sum;
@@ -776,9 +776,8 @@ static void lame_apply_block_type(AacPsyChannel *ctx, FFPsyWindowInfo *wi, int u · 369 unmodified lines
776 ctx->next_window_seq = blocktype;776 ctx->next_window_seq = blocktype;
777}777}
778 778
779static FFPsyWindowInfo psy_lame_window(FFPsyContext *ctx,779static FFPsyWindowInfo psy_lame_window(FFPsyContext *ctx, const float *audio,
780 const int16_t *audio, const int16_t *la,780 const float *la, int channel, int prev_type)
781 int channel, int prev_type)
782{781{
783 AacPsyContext *pctx = (AacPsyContext*) ctx->model_priv_data;782 AacPsyContext *pctx = (AacPsyContext*) ctx->model_priv_data;
784 AacPsyChannel *pch = &pctx->ch[channel];783 AacPsyChannel *pch = &pctx->ch[channel];
@@ -795,20 +794,20 @@ static FFPsyWindowInfo psy_lame_window(FFPsyContext *ctx, · 10 unmodified lines
795 float attack_intensity[(AAC_NUM_BLOCKS_SHORT + 1) * PSY_LAME_NUM_SUBBLOCKS];794 float attack_intensity[(AAC_NUM_BLOCKS_SHORT + 1) * PSY_LAME_NUM_SUBBLOCKS];
796 float energy_subshort[(AAC_NUM_BLOCKS_SHORT + 1) * PSY_LAME_NUM_SUBBLOCKS];795 float energy_subshort[(AAC_NUM_BLOCKS_SHORT + 1) * PSY_LAME_NUM_SUBBLOCKS];
797 float energy_short[AAC_NUM_BLOCKS_SHORT + 1] = { 0 };796 float energy_short[AAC_NUM_BLOCKS_SHORT + 1] = { 0 };
798 int chans = ctx->avctx->channels;797 const float *firbuf = la + (AAC_BLOCK_SIZE_SHORT/4 - PSY_LAME_FIR_LEN);
799 const int16_t *firbuf = la + (AAC_BLOCK_SIZE_SHORT/4 - PSY_LAME_FIR_LEN) * chans;
800 int j, att_sum = 0;798 int j, att_sum = 0;
801 799
802 /* LAME comment: apply high pass filter of fs/4 */800 /* LAME comment: apply high pass filter of fs/4 */
803 for (i = 0; i < AAC_BLOCK_SIZE_LONG; i++) {801 for (i = 0; i < AAC_BLOCK_SIZE_LONG; i++) {
804 float sum1, sum2;802 float sum1, sum2;
805 sum1 = firbuf[(i + ((PSY_LAME_FIR_LEN - 1) / 2)) * chans];803 sum1 = firbuf[i + (PSY_LAME_FIR_LEN - 1) / 2];
806 sum2 = 0.0;804 sum2 = 0.0;
807 for (j = 0; j < ((PSY_LAME_FIR_LEN - 1) / 2) - 1; j += 2) {805 for (j = 0; j < ((PSY_LAME_FIR_LEN - 1) / 2) - 1; j += 2) {
808 sum1 += psy_fir_coeffs[j] * (firbuf[(i + j) * chans] + firbuf[(i + PSY_LAME_FIR_LEN - j) * chans]);806 sum1 += psy_fir_coeffs[j] * (firbuf[i + j] + firbuf[i + PSY_LAME_FIR_LEN - j]);
809 sum2 += psy_fir_coeffs[j + 1] * (firbuf[(i + j + 1) * chans] + firbuf[(i + PSY_LAME_FIR_LEN - j - 1) * chans]);807 sum2 += psy_fir_coeffs[j + 1] * (firbuf[i + j + 1] + firbuf[i + PSY_LAME_FIR_LEN - j - 1]);
810 }808 }
811 hpfsmpl[i] = sum1 + sum2;809 /* NOTE: The LAME psymodel expects it's input in the range -32768 to 32768. Tuning this for normalized floats would be difficult. */
810 hpfsmpl[i] = (sum1 + sum2) * 32768.0f;
812 }811 }
813 812
814 /* Calculate the energies of each sub-shortblock */813 /* Calculate the energies of each sub-shortblock */
@@ -823,16 +822,15 @@ static FFPsyWindowInfo psy_lame_window(FFPsyContext *ctx, · 8 unmodified lines
823 float const *const pfe = pf + AAC_BLOCK_SIZE_LONG / (AAC_NUM_BLOCKS_SHORT * PSY_LAME_NUM_SUBBLOCKS);822 float const *const pfe = pf + AAC_BLOCK_SIZE_LONG / (AAC_NUM_BLOCKS_SHORT * PSY_LAME_NUM_SUBBLOCKS);
824 float p = 1.0f;823 float p = 1.0f;
825 for (; pf < pfe; pf++)824 for (; pf < pfe; pf++)
826 if (p < fabsf(*pf))825 p = FFMAX(p, fabsf(*pf));
827 p = fabsf(*pf);
828 pch->prev_energy_subshort[i] = energy_subshort[i + PSY_LAME_NUM_SUBBLOCKS] = p;826 pch->prev_energy_subshort[i] = energy_subshort[i + PSY_LAME_NUM_SUBBLOCKS] = p;
829 energy_short[1 + i / PSY_LAME_NUM_SUBBLOCKS] += p;827 energy_short[1 + i / PSY_LAME_NUM_SUBBLOCKS] += p;
830 /* FIXME: The indexes below are [i + 3 - 2] in the LAME source.828 /* NOTE: The indexes below are [i + 3 - 2] in the LAME source.
831 * Obviously the 3 and 2 have some significance, or this would be just [i + 1]829 * Obviously the 3 and 2 have some significance, or this would be just [i + 1]
832 * (which is what we use here). What the 3 stands for is ambigious, as it is both830 * (which is what we use here). What the 3 stands for is ambiguous, as it is both
833 * number of short blocks, and the number of sub-short blocks.831 * number of short blocks, and the number of sub-short blocks.
834 * It seems that LAME is comparing each sub-block to sub-block + 1 in the832 * It seems that LAME is comparing each sub-block to sub-block + 1 in the
835 * previous block.833 * previous block.
836 */834 */
837 if (p > energy_subshort[i + 1])835 if (p > energy_subshort[i + 1])
838 p = p / energy_subshort[i + 1];836 p = p / energy_subshort[i + 1];
libavcodec/allcodecs.c+1−0
@@ -245,6 +245,7 @@ void avcodec_register_all(void) · 244 unmodified lines
245 REGISTER_DECODER (XAN_WC3, xan_wc3);245 REGISTER_DECODER (XAN_WC3, xan_wc3);
246 REGISTER_DECODER (XAN_WC4, xan_wc4);246 REGISTER_DECODER (XAN_WC4, xan_wc4);
247 REGISTER_DECODER (XL, xl);247 REGISTER_DECODER (XL, xl);
248 REGISTER_ENCDEC (XWD, xwd);
248 REGISTER_ENCDEC (Y41P, y41p);249 REGISTER_ENCDEC (Y41P, y41p);
249 REGISTER_DECODER (YOP, yop);250 REGISTER_DECODER (YOP, yop);
250 REGISTER_ENCDEC (YUV4, yuv4);251 REGISTER_ENCDEC (YUV4, yuv4);
libavcodec/avcodec.h+1−0
@@ -255,6 +255,7 @@ enum CodecID { · 254 unmodified lines
255 CODEC_ID_VBLE,255 CODEC_ID_VBLE,
256 CODEC_ID_DXTORY,256 CODEC_ID_DXTORY,
257 CODEC_ID_V410,257 CODEC_ID_V410,
258 CODEC_ID_XWD,
258 CODEC_ID_Y41P = MKBETAG('Y','4','1','P'),259 CODEC_ID_Y41P = MKBETAG('Y','4','1','P'),
259 CODEC_ID_UTVIDEO = 0x800,260 CODEC_ID_UTVIDEO = 0x800,
260 CODEC_ID_ESCAPE130 = MKBETAG('E','1','3','0'),261 CODEC_ID_ESCAPE130 = MKBETAG('E','1','3','0'),
libavcodec/psymodel.c+6−11
@@ -112,20 +112,15 @@ av_cold struct FFPsyPreprocessContext* ff_psy_preprocess_init(AVCodecContext *av · 111 unmodified lines
112 return ctx;112 return ctx;
113}113}
114 114
115void ff_psy_preprocess(struct FFPsyPreprocessContext *ctx,115void ff_psy_preprocess(struct FFPsyPreprocessContext *ctx, float **audio, int channels)
116 const int16_t *audio, int16_t *dest,
117 int tag, int channels)
118{116{
119 int ch, i;117 int ch;
118 int frame_size = ctx->avctx->frame_size;
119
120 if (ctx->fstate) {120 if (ctx->fstate) {
121 for (ch = 0; ch < channels; ch++)121 for (ch = 0; ch < channels; ch++)
122 ff_iir_filter(ctx->fcoeffs, ctx->fstate[tag+ch], ctx->avctx->frame_size,122 ff_iir_filter_flt(ctx->fcoeffs, ctx->fstate[ch], frame_size,
123 audio + ch, ctx->avctx->channels,123 &audio[ch][frame_size], 1, &audio[ch][frame_size], 1);
124 dest + ch, ctx->avctx->channels);
125 } else {
126 for (ch = 0; ch < channels; ch++)
127 for (i = 0; i < ctx->avctx->frame_size; i++)
128 dest[i*ctx->avctx->channels + ch] = audio[i*ctx->avctx->channels + ch];
129 }124 }
130}125}
131 126
libavcodec/psymodel.h+4−8
@@ -109,7 +109,7 @@ typedef struct FFPsyModel { · 108 unmodified lines
109 *109 *
110 * @return suggested window information in a structure110 * @return suggested window information in a structure
111 */111 */
112 FFPsyWindowInfo (*window)(FFPsyContext *ctx, const int16_t *audio, const int16_t *la, int channel, int prev_type);112 FFPsyWindowInfo (*window)(FFPsyContext *ctx, const float *audio, const float *la, int channel, int prev_type);
113 113
114 /**114 /**
115 * Perform psychoacoustic analysis and set band info (threshold, energy) for a group of channels.115 * Perform psychoacoustic analysis and set band info (threshold, energy) for a group of channels.
@@ -174,14 +174,10 @@ av_cold struct FFPsyPreprocessContext* ff_psy_preprocess_init(AVCodecContext *av · 58 unmodified lines
174 * Preprocess several channel in audio frame in order to compress it better.174 * Preprocess several channel in audio frame in order to compress it better.
175 *175 *
176 * @param ctx preprocessing context176 * @param ctx preprocessing context
177 * @param audio samples to preprocess177 * @param audio samples to be filtered (in place)
178 * @param dest place to put filtered samples178 * @param channels number of channel to preprocess
179 * @param tag channel number
180 * @param channels number of channel to preprocess (some additional work may be done on stereo pair)
181 */179 */
182void ff_psy_preprocess(struct FFPsyPreprocessContext *ctx,180void ff_psy_preprocess(struct FFPsyPreprocessContext *ctx, float **audio, int channels);
183 const int16_t *audio, int16_t *dest,
184 int tag, int channels);
185 181
186/**182/**
187 * Cleanup audio preprocessing module.183 * Cleanup audio preprocessing module.
libavcodec/vc1.c+9−7
@@ -918,13 +918,15 @@ int vc1_parse_frame_header_adv(VC1Context *v, GetBitContext* gb) · 917 unmodified lines
918 }918 }
919 goto parse_common_info;919 goto parse_common_info;
920 }920 }
921 if (v->finterpflag)921 if (v->fcm == PROGRESSIVE) {
922 v->interpfrm = get_bits1(gb);922 if (v->finterpflag)
923 if (v->s.pict_type == AV_PICTURE_TYPE_B) {923 v->interpfrm = get_bits1(gb);
924 v->bfraction_lut_index = get_vlc2(gb, ff_vc1_bfraction_vlc.table, VC1_BFRACTION_VLC_BITS, 1);924 if (v->s.pict_type == AV_PICTURE_TYPE_B) {
925 v->bfraction = ff_vc1_bfraction_lut[v->bfraction_lut_index];925 v->bfraction_lut_index = get_vlc2(gb, ff_vc1_bfraction_vlc.table, VC1_BFRACTION_VLC_BITS, 1);
926 if (v->bfraction == 0) {926 v->bfraction = ff_vc1_bfraction_lut[v->bfraction_lut_index];
927 v->s.pict_type = AV_PICTURE_TYPE_BI; /* XXX: should not happen here */927 if (v->bfraction == 0) {
928 v->s.pict_type = AV_PICTURE_TYPE_BI; /* XXX: should not happen here */
929 }
928 }930 }
929 }931 }
930 932
libavcodec/version.h+2−2
@@ -21,8 +21,8 @@ · 20 unmodified lines
21#define AVCODEC_VERSION_H21#define AVCODEC_VERSION_H
22 22
23#define LIBAVCODEC_VERSION_MAJOR 5323#define LIBAVCODEC_VERSION_MAJOR 53
24#define LIBAVCODEC_VERSION_MINOR 5724#define LIBAVCODEC_VERSION_MINOR 58
25#define LIBAVCODEC_VERSION_MICRO 10525#define LIBAVCODEC_VERSION_MICRO 100
26 26
27#define LIBAVCODEC_VERSION_INT AV_VERSION_INT(LIBAVCODEC_VERSION_MAJOR, \27#define LIBAVCODEC_VERSION_INT AV_VERSION_INT(LIBAVCODEC_VERSION_MAJOR, \
28 LIBAVCODEC_VERSION_MINOR, \28 LIBAVCODEC_VERSION_MINOR, \
libavcodec/xwd.h+41−0
@@ -0,0 +1,41 @@
1/*
2 * XWD image format
3 *
4 * Copyright (c) 2012 Paul B Mahol
5 *
6 * This file is part of Libav.
7 *
8 * Libav is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2.1 of the License, or (at your option) any later version.
12 *
13 * Libav is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
17 *
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with Libav; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21 */
22
23#ifndef AVCODEC_XWD_H
24#define AVCODEC_XWD_H
25
26#define XWD_VERSION 7
27#define XWD_HEADER_SIZE 100
28#define XWD_CMAP_SIZE 12
29
30#define XWD_XY_BITMAP 0
31#define XWD_XY_PIXMAP 1
32#define XWD_Z_PIXMAP 2
33
34#define XWD_STATIC_GRAY 0
35#define XWD_GRAY_SCALE 1
36#define XWD_STATIC_COLOR 2
37#define XWD_PSEUDO_COLOR 3
38#define XWD_TRUE_COLOR 4
39#define XWD_DIRECT_COLOR 5
40
41#endif /* AVCODEC_XWD_H */
libavcodec/xwddec.c+267−0
@@ -0,0 +1,267 @@
1/*
2 * XWD image format
3 *
4 * Copyright (c) 2012 Paul B Mahol
5 *
6 * This file is part of Libav.
7 *
8 * Libav is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2.1 of the License, or (at your option) any later version.
12 *
13 * Libav is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
17 *
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with Libav; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21 */
22
23#include "libavutil/imgutils.h"
24#include "avcodec.h"
25#include "bytestream.h"
26#include "xwd.h"
27
28static av_cold int xwd_decode_init(AVCodecContext *avctx)
29{
30 avctx->coded_frame = avcodec_alloc_frame();
31 if (!avctx->coded_frame)
32 return AVERROR(ENOMEM);
33
34 return 0;
35}
36
37static int xwd_decode_frame(AVCodecContext *avctx, void *data,
38 int *data_size, AVPacket *avpkt)
39{
40 AVFrame *p = avctx->coded_frame;
41 const uint8_t *buf = avpkt->data;
42 int i, ret, buf_size = avpkt->size;
43 uint32_t version, header_size, vclass, ncolors;
44 uint32_t xoffset, be, bpp, lsize, rsize;
45 uint32_t pixformat, pixdepth, bunit, bitorder, bpad;
46 uint32_t rgb[3];
47 uint8_t *ptr;
48
49 if (buf_size < XWD_HEADER_SIZE)
50 return AVERROR_INVALIDDATA;
51
52 header_size = bytestream_get_be32(&buf);
53 if (buf_size < header_size)
54 return AVERROR_INVALIDDATA;
55
56 version = bytestream_get_be32(&buf);
57 if (version != XWD_VERSION) {
58 av_log(avctx, AV_LOG_ERROR, "unsupported version\n");
59 return AVERROR_INVALIDDATA;
60 }
61
62 if (header_size < XWD_HEADER_SIZE) {
63 av_log(avctx, AV_LOG_ERROR, "invalid header size\n");
64 return AVERROR_INVALIDDATA;
65 }
66
67 pixformat = bytestream_get_be32(&buf);
68 pixdepth = bytestream_get_be32(&buf);
69 avctx->width = bytestream_get_be32(&buf);
70 avctx->height = bytestream_get_be32(&buf);
71 xoffset = bytestream_get_be32(&buf);
72 be = bytestream_get_be32(&buf);
73 bunit = bytestream_get_be32(&buf);
74 bitorder = bytestream_get_be32(&buf);
75 bpad = bytestream_get_be32(&buf);
76 bpp = bytestream_get_be32(&buf);
77 lsize = bytestream_get_be32(&buf);
78 vclass = bytestream_get_be32(&buf);
79 rgb[0] = bytestream_get_be32(&buf);
80 rgb[1] = bytestream_get_be32(&buf);
81 rgb[2] = bytestream_get_be32(&buf);
82 buf += 8;
83 ncolors = bytestream_get_be32(&buf);
84 buf += header_size - (XWD_HEADER_SIZE - 20);
85
86 av_log(avctx, AV_LOG_DEBUG, "pixformat %d, pixdepth %d, bunit %d, bitorder %d, bpad %d\n",
87 pixformat, pixdepth, bunit, bitorder, bpad);
88 av_log(avctx, AV_LOG_DEBUG, "vclass %d, ncolors %d, bpp %d, be %d, lsize %d, xoffset %d\n",
89 vclass, ncolors, bpp, be, lsize, xoffset);
90 av_log(avctx, AV_LOG_DEBUG, "red %0x, green %0x, blue %0x\n", rgb[0], rgb[1], rgb[2]);
91
92 if (pixformat > XWD_Z_PIXMAP) {
93 av_log(avctx, AV_LOG_ERROR, "invalid pixmap format\n");
94 return AVERROR_INVALIDDATA;
95 }
96
97 if (pixdepth == 0 || pixdepth > 32) {
98 av_log(avctx, AV_LOG_ERROR, "invalid pixmap depth\n");
99 return AVERROR_INVALIDDATA;
100 }
101
102 if (xoffset) {
103 av_log_ask_for_sample(avctx, "unsupported xoffset %d\n", xoffset);
104 return AVERROR_PATCHWELCOME;
105 }
106
107 if (be > 1) {
108 av_log(avctx, AV_LOG_ERROR, "invalid byte order\n");
109 return AVERROR_INVALIDDATA;
110 }
111
112 if (bitorder > 1) {
113 av_log(avctx, AV_LOG_ERROR, "invalid bitmap bit order\n");
114 return AVERROR_INVALIDDATA;
115 }
116
117 if (bunit != 8 && bunit != 16 && bunit != 32) {
118 av_log(avctx, AV_LOG_ERROR, "invalid bitmap unit\n");
119 return AVERROR_INVALIDDATA;
120 }
121
122 if (bpad != 8 && bpad != 16 && bpad != 32) {
123 av_log(avctx, AV_LOG_ERROR, "invalid bitmap scan-line pad\n");
124 return AVERROR_INVALIDDATA;
125 }
126
127 if (bpp == 0 || bpp > 32) {
128 av_log(avctx, AV_LOG_ERROR, "invalid bits per pixel\n");
129 return AVERROR_INVALIDDATA;
130 }
131
132 if (ncolors > 256) {
133 av_log(avctx, AV_LOG_ERROR, "invalid number of entries in colormap\n");
134 return AVERROR_INVALIDDATA;
135 }
136
137 if ((ret = av_image_check_size(avctx->width, avctx->height, 0, NULL)) < 0)
138 return ret;
139
140 rsize = FFALIGN(avctx->width * bpp, bpad) / 8;
141 if (lsize < rsize) {
142 av_log(avctx, AV_LOG_ERROR, "invalid bytes per scan-line\n");
143 return AVERROR_INVALIDDATA;
144 }
145
146 if (buf_size < header_size + ncolors * XWD_CMAP_SIZE + avctx->height * lsize) {
147 av_log(avctx, AV_LOG_ERROR, "input buffer too small\n");
148 return AVERROR_INVALIDDATA;
149 }
150
151 if (pixformat != XWD_Z_PIXMAP) {
152 av_log(avctx, AV_LOG_ERROR, "pixmap format %d unsupported\n", pixformat);
153 return AVERROR_PATCHWELCOME;
154 }
155
156 avctx->pix_fmt = PIX_FMT_NONE;
157 switch (vclass) {
158 case XWD_STATIC_GRAY:
159 case XWD_GRAY_SCALE:
160 if (bpp != 1)
161 return AVERROR_INVALIDDATA;
162 if (pixdepth == 1)
163 avctx->pix_fmt = PIX_FMT_MONOWHITE;
164 break;
165 case XWD_STATIC_COLOR:
166 case XWD_PSEUDO_COLOR:
167 if (bpp == 8)
168 avctx->pix_fmt = PIX_FMT_PAL8;
169 break;
170 case XWD_TRUE_COLOR:
171 case XWD_DIRECT_COLOR:
172 if (bpp != 16 && bpp != 24 && bpp != 32)
173 return AVERROR_INVALIDDATA;
174 if (bpp == 16 && pixdepth == 15) {
175 if (rgb[0] == 0x7C00 && rgb[1] == 0x3E0 && rgb[2] == 0x1F)
176 avctx->pix_fmt = be ? PIX_FMT_RGB555BE : PIX_FMT_RGB555LE;
177 else if (rgb[0] == 0x1F && rgb[1] == 0x3E0 && rgb[2] == 0x7C00)
178 avctx->pix_fmt = be ? PIX_FMT_BGR555BE : PIX_FMT_BGR555LE;
179 } else if (bpp == 16 && pixdepth == 16) {
180 if (rgb[0] == 0xF800 && rgb[1] == 0x7E0 && rgb[2] == 0x1F)
181 avctx->pix_fmt = be ? PIX_FMT_RGB565BE : PIX_FMT_RGB565LE;
182 else if (rgb[0] == 0x1F && rgb[1] == 0x7E0 && rgb[2] == 0xF800)
183 avctx->pix_fmt = be ? PIX_FMT_BGR565BE : PIX_FMT_BGR565LE;
184 } else if (bpp == 24) {
185 if (rgb[0] == 0xFF0000 && rgb[1] == 0xFF00 && rgb[2] == 0xFF)
186 avctx->pix_fmt = be ? PIX_FMT_RGB24 : PIX_FMT_BGR24;
187 else if (rgb[0] == 0xFF && rgb[1] == 0xFF00 && rgb[2] == 0xFF0000)
188 avctx->pix_fmt = be ? PIX_FMT_BGR24 : PIX_FMT_RGB24;
189 } else if (bpp == 32) {
190 if (rgb[0] == 0xFF0000 && rgb[1] == 0xFF00 && rgb[2] == 0xFF)
191 avctx->pix_fmt = be ? PIX_FMT_ARGB : PIX_FMT_BGRA;
192 else if (rgb[0] == 0xFF && rgb[1] == 0xFF00 && rgb[2] == 0xFF0000)
193 avctx->pix_fmt = be ? PIX_FMT_ABGR : PIX_FMT_RGBA;
194 }
195 buf += ncolors * XWD_CMAP_SIZE;
196 break;
197 default:
198 av_log(avctx, AV_LOG_ERROR, "invalid visual class\n");
199 return AVERROR_INVALIDDATA;
200 }
201
202 if (avctx->pix_fmt == PIX_FMT_NONE) {
203 av_log_ask_for_sample(avctx, "unknown file: bpp %d, pixdepth %d, vclass %d\n", bpp, pixdepth, vclass);
204 return AVERROR_PATCHWELCOME;
205 }
206
207 if (p->data[0])
208 avctx->release_buffer(avctx, p);
209
210 p->reference = 0;
211 if ((ret = avctx->get_buffer(avctx, p)) < 0) {
212 av_log(avctx, AV_LOG_ERROR, "get_buffer() failed\n");
213 return ret;
214 }
215
216 p->key_frame = 1;
217 p->pict_type = AV_PICTURE_TYPE_I;
218
219 if (avctx->pix_fmt == PIX_FMT_PAL8) {
220 uint32_t *dst = (uint32_t *)p->data[1];
221 uint8_t red, green, blue;
222
223 for (i = 0; i < ncolors; i++) {
224
225 buf += 4; // skip colormap entry number
226 red = *buf; buf += 2;
227 green = *buf; buf += 2;
228 blue = *buf; buf += 2;
229 buf += 2; // skip bitmask flag and padding
230
231 dst[i] = red << 16 | green << 8 | blue;
232 }
233 }
234
235 ptr = p->data[0];
236 for (i = 0; i < avctx->height; i++) {
237 bytestream_get_buffer(&buf, ptr, rsize);
238 buf += lsize - rsize;
239 ptr += p->linesize[0];
240 }
241
242 *data_size = sizeof(AVFrame);
243 *(AVFrame *)data = *p;
244
245 return buf_size;
246}
247
248static av_cold int xwd_decode_close(AVCodecContext *avctx)
249{
250 if (avctx->coded_frame->data[0])
251 avctx->release_buffer(avctx, avctx->coded_frame);
252
253 av_freep(&avctx->coded_frame);
254
255 return 0;
256}
257
258AVCodec ff_xwd_decoder = {
259 .name = "xwd",
260 .type = AVMEDIA_TYPE_VIDEO,
261 .id = CODEC_ID_XWD,
262 .init = xwd_decode_init,
263 .close = xwd_decode_close,
264 .decode = xwd_decode_frame,
265 .capabilities = CODEC_CAP_DR1,
266 .long_name = NULL_IF_CONFIG_SMALL("XWD (X Window Dump) image"),
267};
libavcodec/xwdenc.c+246−0
@@ -0,0 +1,246 @@
1/*
2 * XWD image format
3 *
4 * Copyright (c) 2012 Paul B Mahol
5 *
6 * This file is part of Libav.
7 *
8 * Libav is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2.1 of the License, or (at your option) any later version.
12 *
13 * Libav is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
17 *
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with Libav; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21 */
22
23#include "libavutil/intreadwrite.h"
24#include "libavutil/pixdesc.h"
25#include "avcodec.h"
26#include "bytestream.h"
27#include "xwd.h"
28
29#define WINDOW_NAME "lavcxwdenc"
30#define WINDOW_NAME_SIZE 11
31
32static av_cold int xwd_encode_init(AVCodecContext *avctx)
33{
34 avctx->coded_frame = avcodec_alloc_frame();
35 if (!avctx->coded_frame)
36 return AVERROR(ENOMEM);
37
38 return 0;
39}
40
41static int xwd_encode_frame(AVCodecContext *avctx, uint8_t *buf,
42 int buf_size, void *data)
43{
44 AVFrame *p = data;
45 enum PixelFormat pix_fmt = avctx->pix_fmt;
46 uint32_t pixdepth, bpp, bpad, ncolors = 0, lsize, vclass, be = 0;
47 uint32_t rgb[3] = { 0 };
48 uint32_t header_size;
49 int i, out_size;
50 uint8_t *ptr;
51
52 pixdepth = av_get_bits_per_pixel(&av_pix_fmt_descriptors[pix_fmt]);
53 if (av_pix_fmt_descriptors[pix_fmt].flags & PIX_FMT_BE)
54 be = 1;
55 switch (pix_fmt) {
56 case PIX_FMT_ARGB:
57 case PIX_FMT_BGRA:
58 case PIX_FMT_RGBA:
59 case PIX_FMT_ABGR:
60 if (pix_fmt == PIX_FMT_ARGB ||
61 pix_fmt == PIX_FMT_ABGR)
62 be = 1;
63 if (pix_fmt == PIX_FMT_ABGR ||
64 pix_fmt == PIX_FMT_RGBA) {
65 rgb[0] = 0xFF;
66 rgb[1] = 0xFF00;
67 rgb[2] = 0xFF0000;
68 } else {
69 rgb[0] = 0xFF0000;
70 rgb[1] = 0xFF00;
71 rgb[2] = 0xFF;
72 }
73 bpp = 32;
74 pixdepth = 24;
75 vclass = XWD_TRUE_COLOR;
76 bpad = 32;
77 break;
78 case PIX_FMT_BGR24:
79 case PIX_FMT_RGB24:
80 if (pix_fmt == PIX_FMT_RGB24)
81 be = 1;
82 bpp = 24;
83 vclass = XWD_TRUE_COLOR;
84 bpad = 32;
85 rgb[0] = 0xFF0000;
86 rgb[1] = 0xFF00;
87 rgb[2] = 0xFF;
88 break;
89 case PIX_FMT_RGB565LE:
90 case PIX_FMT_RGB565BE:
91 case PIX_FMT_BGR565LE:
92 case PIX_FMT_BGR565BE:
93 if (pix_fmt == PIX_FMT_BGR565LE ||
94 pix_fmt == PIX_FMT_BGR565BE) {
95 rgb[0] = 0x1F;
96 rgb[1] = 0x7E0;
97 rgb[2] = 0xF800;
98 } else {
99 rgb[0] = 0xF800;
100 rgb[1] = 0x7E0;
101 rgb[2] = 0x1F;
102 }
103 bpp = 16;
104 vclass = XWD_TRUE_COLOR;
105 bpad = 16;
106 break;
107 case PIX_FMT_RGB555LE:
108 case PIX_FMT_RGB555BE:
109 case PIX_FMT_BGR555LE:
110 case PIX_FMT_BGR555BE:
111 if (pix_fmt == PIX_FMT_BGR555LE ||
112 pix_fmt == PIX_FMT_BGR555BE) {
113 rgb[0] = 0x1F;
114 rgb[1] = 0x3E0;
115 rgb[2] = 0x7C00;
116 } else {
117 rgb[0] = 0x7C00;
118 rgb[1] = 0x3E0;
119 rgb[2] = 0x1F;
120 }
121 bpp = 16;
122 vclass = XWD_TRUE_COLOR;
123 bpad = 16;
124 break;
125 case PIX_FMT_RGB8:
126 case PIX_FMT_BGR8:
127 case PIX_FMT_RGB4_BYTE:
128 case PIX_FMT_BGR4_BYTE:
129 case PIX_FMT_PAL8:
130 bpp = 8;
131 vclass = XWD_PSEUDO_COLOR;
132 bpad = 8;
133 ncolors = 256;
134 break;
135 case PIX_FMT_MONOWHITE:
136 bpp = 1;
137 bpad = 8;
138 vclass = XWD_STATIC_GRAY;
139 break;
140 default:
141 av_log(avctx, AV_LOG_INFO, "unsupported pixel format\n");
142 return AVERROR(EINVAL);
143 }
144
145 lsize = FFALIGN(bpp * avctx->width, bpad) / 8;
146 header_size = XWD_HEADER_SIZE + WINDOW_NAME_SIZE;
147 out_size = header_size + ncolors * XWD_CMAP_SIZE + avctx->height * lsize;
148
149 if (buf_size < out_size) {
150 av_log(avctx, AV_LOG_ERROR, "output buffer too small\n");
151 return AVERROR(ENOMEM);
152 }
153
154 avctx->coded_frame->key_frame = 1;
155 avctx->coded_frame->pict_type = AV_PICTURE_TYPE_I;
156
157 bytestream_put_be32(&buf, header_size);
158 bytestream_put_be32(&buf, XWD_VERSION); // file version
159 bytestream_put_be32(&buf, XWD_Z_PIXMAP); // pixmap format
160 bytestream_put_be32(&buf, pixdepth); // pixmap depth in pixels
161 bytestream_put_be32(&buf, avctx->width); // pixmap width in pixels
162 bytestream_put_be32(&buf, avctx->height); // pixmap height in pixels
163 bytestream_put_be32(&buf, 0); // bitmap x offset
164 bytestream_put_be32(&buf, be); // byte order
165 bytestream_put_be32(&buf, 32); // bitmap unit
166 bytestream_put_be32(&buf, be); // bit-order of image data
167 bytestream_put_be32(&buf, bpad); // bitmap scan-line pad in bits
168 bytestream_put_be32(&buf, bpp); // bits per pixel
169 bytestream_put_be32(&buf, lsize); // bytes per scan-line
170 bytestream_put_be32(&buf, vclass); // visual class
171 bytestream_put_be32(&buf, rgb[0]); // red mask
172 bytestream_put_be32(&buf, rgb[1]); // green mask
173 bytestream_put_be32(&buf, rgb[2]); // blue mask
174 bytestream_put_be32(&buf, 8); // size of each bitmask in bits
175 bytestream_put_be32(&buf, ncolors); // number of colors
176 bytestream_put_be32(&buf, ncolors); // number of entries in color map
177 bytestream_put_be32(&buf, avctx->width); // window width
178 bytestream_put_be32(&buf, avctx->height); // window height
179 bytestream_put_be32(&buf, 0); // window upper left X coordinate
180 bytestream_put_be32(&buf, 0); // window upper left Y coordinate
181 bytestream_put_be32(&buf, 0); // window border width
182 bytestream_put_buffer(&buf, WINDOW_NAME, WINDOW_NAME_SIZE);
183
184 for (i = 0; i < ncolors; i++) {
185 uint32_t val;
186 uint8_t red, green, blue;
187
188 val = AV_RN32A(p->data[1] + i * 4);
189 red = (val >> 16) & 0xFF;
190 green = (val >> 8) & 0xFF;
191 blue = val & 0xFF;
192
193 bytestream_put_be32(&buf, i); // colormap entry number
194 bytestream_put_be16(&buf, red << 8);
195 bytestream_put_be16(&buf, green << 8);
196 bytestream_put_be16(&buf, blue << 8);
197 bytestream_put_byte(&buf, 0x7); // bitmask flag
198 bytestream_put_byte(&buf, 0); // padding
199 }
200
201 ptr = p->data[0];
202 for (i = 0; i < avctx->height; i++) {
203 bytestream_put_buffer(&buf, ptr, lsize);
204 ptr += p->linesize[0];
205 }
206
207 return out_size;
208}
209
210static av_cold int xwd_encode_close(AVCodecContext *avctx)
211{
212 av_freep(&avctx->coded_frame);
213
214 return 0;
215}
216
217AVCodec ff_xwd_encoder = {
218 .name = "xwd",
219 .type = AVMEDIA_TYPE_VIDEO,
220 .id = CODEC_ID_XWD,
221 .init = xwd_encode_init,
222 .encode = xwd_encode_frame,
223 .close = xwd_encode_close,
224 .pix_fmts = (const enum PixelFormat[]) { PIX_FMT_BGRA,
225 PIX_FMT_RGBA,
226 PIX_FMT_ARGB,
227 PIX_FMT_ABGR,
228 PIX_FMT_RGB24,
229 PIX_FMT_BGR24,
230 PIX_FMT_RGB565BE,
231 PIX_FMT_RGB565LE,
232 PIX_FMT_BGR565BE,
233 PIX_FMT_BGR565LE,
234 PIX_FMT_RGB555BE,
235 PIX_FMT_RGB555LE,
236 PIX_FMT_BGR555BE,
237 PIX_FMT_BGR555LE,
238 PIX_FMT_RGB8,
239 PIX_FMT_BGR8,
240 PIX_FMT_RGB4_BYTE,
241 PIX_FMT_BGR4_BYTE,
242 PIX_FMT_PAL8,
243 PIX_FMT_MONOWHITE,
244 PIX_FMT_NONE },
245 .long_name = NULL_IF_CONFIG_SMALL("XWD (X Window Dump) image"),
246};
libavformat/Makefile+2−2
@@ -326,8 +326,8 @@ OBJS-$(CONFIG_WC3_DEMUXER) += wc3movie.o · 325 unmodified lines
326OBJS-$(CONFIG_WEBM_MUXER) += matroskaenc.o matroska.o \326OBJS-$(CONFIG_WEBM_MUXER) += matroskaenc.o matroska.o \
327 riff.o isom.o avc.o \327 riff.o isom.o avc.o \
328 flacenc_header.o avlanguage.o328 flacenc_header.o avlanguage.o
329OBJS-$(CONFIG_WSAUD_DEMUXER) += westwood.o329OBJS-$(CONFIG_WSAUD_DEMUXER) += westwood_aud.o
330OBJS-$(CONFIG_WSVQA_DEMUXER) += westwood.o330OBJS-$(CONFIG_WSVQA_DEMUXER) += westwood_vqa.o
331OBJS-$(CONFIG_WTV_DEMUXER) += wtvdec.o wtv.o asfdec.o asf.o asfcrypt.o \331OBJS-$(CONFIG_WTV_DEMUXER) += wtvdec.o wtv.o asfdec.o asf.o asfcrypt.o \
332 avlanguage.o mpegts.o isom.o riff.o332 avlanguage.o mpegts.o isom.o riff.o
333OBJS-$(CONFIG_WTV_MUXER) += wtvenc.o wtv.o asf.o asfenc.o riff.o333OBJS-$(CONFIG_WTV_MUXER) += wtvenc.o wtv.o asf.o asfenc.o riff.o
libavformat/img2.c+2−1
@@ -91,6 +91,7 @@ static const IdStrMap img_tags[] = { · 90 unmodified lines
91 { CODEC_ID_JPEG2000 , "jpc"},91 { CODEC_ID_JPEG2000 , "jpc"},
92 { CODEC_ID_DPX , "dpx"},92 { CODEC_ID_DPX , "dpx"},
93 { CODEC_ID_PICTOR , "pic"},93 { CODEC_ID_PICTOR , "pic"},
94 { CODEC_ID_XWD , "xwd"},
94 { CODEC_ID_NONE , NULL}95 { CODEC_ID_NONE , NULL}
95};96};
96 97
@@ -528,7 +529,7 @@ AVOutputFormat ff_image2_muxer = { · 431 unmodified lines
528 .name = "image2",529 .name = "image2",
529 .long_name = NULL_IF_CONFIG_SMALL("image2 sequence"),530 .long_name = NULL_IF_CONFIG_SMALL("image2 sequence"),
530 .extensions = "bmp,dpx,jls,jpeg,jpg,ljpg,pam,pbm,pcx,pgm,pgmyuv,png,"531 .extensions = "bmp,dpx,jls,jpeg,jpg,ljpg,pam,pbm,pcx,pgm,pgmyuv,png,"
531 "ppm,sgi,tga,tif,tiff,jp2,j2c",532 "ppm,sgi,tga,tif,tiff,jp2,j2c,xwd",
532 .priv_data_size = sizeof(VideoData),533 .priv_data_size = sizeof(VideoData),
533 .video_codec = CODEC_ID_MJPEG,534 .video_codec = CODEC_ID_MJPEG,
534 .write_header = write_header,535 .write_header = write_header,
libavformat/westwood_aud.c+173−0
@@ -0,0 +1,173 @@
1/*
2 * Westwood Studios AUD Format Demuxer
3 * Copyright (c) 2003 The ffmpeg Project
4 *
5 * This file is part of Libav.
6 *
7 * Libav is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Lesser General Public
9 * License as published by the Free Software Foundation; either
10 * version 2.1 of the License, or (at your option) any later version.
11 *
12 * Libav is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * Lesser General Public License for more details.
16 *
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with Libav; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20 */
21
22/**
23 * @file
24 * Westwood Studios AUD file demuxer
25 * by Mike Melanson (melanson@pcisys.net)
26 * for more information on the Westwood file formats, visit:
27 * http://www.pcisys.net/~melanson/codecs/
28 * http://www.geocities.com/SiliconValley/8682/aud3.txt
29 *
30 * Implementation note: There is no definite file signature for AUD files.
31 * The demuxer uses a probabilistic strategy for content detection. This
32 * entails performing sanity checks on certain header values in order to
33 * qualify a file. Refer to wsaud_probe() for the precise parameters.
34 */
35
36#include "libavutil/intreadwrite.h"
37#include "avformat.h"
38#include "internal.h"
39
40#define AUD_HEADER_SIZE 12
41#define AUD_CHUNK_PREAMBLE_SIZE 8
42#define AUD_CHUNK_SIGNATURE 0x0000DEAF
43
44typedef struct WsAudDemuxContext {
45 int audio_samplerate;
46 int audio_channels;
47 int audio_bits;
48 enum CodecID audio_type;
49 int audio_stream_index;
50 int64_t audio_frame_counter;
51} WsAudDemuxContext;
52
53static int wsaud_probe(AVProbeData *p)
54{
55 int field;
56
57 /* Probabilistic content detection strategy: There is no file signature
58 * so perform sanity checks on various header parameters:
59 * 8000 <= sample rate (16 bits) <= 48000 ==> 40001 acceptable numbers
60 * flags <= 0x03 (2 LSBs are used) ==> 4 acceptable numbers
61 * compression type (8 bits) = 1 or 99 ==> 2 acceptable numbers
62 * first audio chunk signature (32 bits) ==> 1 acceptable number
63 * The number space contains 2^64 numbers. There are 40001 * 4 * 2 * 1 =
64 * 320008 acceptable number combinations.
65 */
66
67 if (p->buf_size < AUD_HEADER_SIZE + AUD_CHUNK_PREAMBLE_SIZE)
68 return 0;
69
70 /* check sample rate */
71 field = AV_RL16(&p->buf[0]);
72 if ((field < 8000) || (field > 48000))
73 return 0;
74
75 /* enforce the rule that the top 6 bits of this flags field are reserved (0);
76 * this might not be true, but enforce it until deemed unnecessary */
77 if (p->buf[10] & 0xFC)
78 return 0;
79
80 /* note: only check for WS IMA (type 99) right now since there is no
81 * support for type 1 */
82 if (p->buf[11] != 99)
83 return 0;
84
85 /* read ahead to the first audio chunk and validate the first header signature */
86 if (AV_RL32(&p->buf[16]) != AUD_CHUNK_SIGNATURE)
87 return 0;
88
89 /* return 1/2 certainty since this file check is a little sketchy */
90 return AVPROBE_SCORE_MAX / 2;
91}
92
93static int wsaud_read_header(AVFormatContext *s,
94 AVFormatParameters *ap)
95{
96 WsAudDemuxContext *wsaud = s->priv_data;
97 AVIOContext *pb = s->pb;
98 AVStream *st;
99 unsigned char header[AUD_HEADER_SIZE];
100
101 if (avio_read(pb, header, AUD_HEADER_SIZE) != AUD_HEADER_SIZE)
102 return AVERROR(EIO);
103 wsaud->audio_samplerate = AV_RL16(&header[0]);
104 if (header[11] == 99)
105 wsaud->audio_type = CODEC_ID_ADPCM_IMA_WS;
106 else
107 return AVERROR_INVALIDDATA;
108
109 /* flag 0 indicates stereo */
110 wsaud->audio_channels = (header[10] & 0x1) + 1;
111 /* flag 1 indicates 16 bit audio */
112 wsaud->audio_bits = (((header[10] & 0x2) >> 1) + 1) * 8;
113
114 /* initialize the audio decoder stream */
115 st = avformat_new_stream(s, NULL);
116 if (!st)
117 return AVERROR(ENOMEM);
118 avpriv_set_pts_info(st, 33, 1, wsaud->audio_samplerate);
119 st->codec->codec_type = AVMEDIA_TYPE_AUDIO;
120 st->codec->codec_id = wsaud->audio_type;
121 st->codec->codec_tag = 0; /* no tag */
122 st->codec->channels = wsaud->audio_channels;
123 st->codec->sample_rate = wsaud->audio_samplerate;
124 st->codec->bits_per_coded_sample = wsaud->audio_bits;
125 st->codec->bit_rate = st->codec->channels * st->codec->sample_rate *
126 st->codec->bits_per_coded_sample / 4;
127 st->codec->block_align = st->codec->channels * st->codec->bits_per_coded_sample;
128
129 wsaud->audio_stream_index = st->index;
130 wsaud->audio_frame_counter = 0;
131
132 return 0;
133}
134
135static int wsaud_read_packet(AVFormatContext *s,
136 AVPacket *pkt)
137{
138 WsAudDemuxContext *wsaud = s->priv_data;
139 AVIOContext *pb = s->pb;
140 unsigned char preamble[AUD_CHUNK_PREAMBLE_SIZE];
141 unsigned int chunk_size;
142 int ret = 0;
143
144 if (avio_read(pb, preamble, AUD_CHUNK_PREAMBLE_SIZE) !=
145 AUD_CHUNK_PREAMBLE_SIZE)
146 return AVERROR(EIO);
147
148 /* validate the chunk */
149 if (AV_RL32(&preamble[4]) != AUD_CHUNK_SIGNATURE)
150 return AVERROR_INVALIDDATA;
151
152 chunk_size = AV_RL16(&preamble[0]);
153 ret= av_get_packet(pb, pkt, chunk_size);
154 if (ret != chunk_size)
155 return AVERROR(EIO);
156 pkt->stream_index = wsaud->audio_stream_index;
157 pkt->pts = wsaud->audio_frame_counter;
158 pkt->pts /= wsaud->audio_samplerate;
159
160 /* 2 samples/byte, 1 or 2 samples per frame depending on stereo */
161 wsaud->audio_frame_counter += (chunk_size * 2) / wsaud->audio_channels;
162
163 return ret;
164}
165
166AVInputFormat ff_wsaud_demuxer = {
167 .name = "wsaud",
168 .long_name = NULL_IF_CONFIG_SMALL("Westwood Studios audio format"),
169 .priv_data_size = sizeof(WsAudDemuxContext),
170 .read_probe = wsaud_probe,
171 .read_header = wsaud_read_header,
172 .read_packet = wsaud_read_packet,
173};
libavformat/westwood_vqa.crenamed from libavformat/westwood.c+2−145
@@ -1,5 +1,5 @@
1/*1/*
2 * Westwood Studios Multimedia Formats Demuxer (VQA, AUD)2 * Westwood Studios VQA Format Demuxer
3 * Copyright (c) 2003 The ffmpeg Project3 * Copyright (c) 2003 The ffmpeg Project
4 *4 *
5 * This file is part of FFmpeg.5 * This file is part of FFmpeg.
@@ -21,26 +21,17 @@ · 15 unmodified lines
21 21
22/**22/**
23 * @file23 * @file
24 * Westwood Studios VQA & AUD file demuxers24 * Westwood Studios VQA file demuxer
25 * by Mike Melanson (melanson@pcisys.net)25 * by Mike Melanson (melanson@pcisys.net)
26 * for more information on the Westwood file formats, visit:26 * for more information on the Westwood file formats, visit:
27 * http://www.pcisys.net/~melanson/codecs/27 * http://www.pcisys.net/~melanson/codecs/
28 * http://www.geocities.com/SiliconValley/8682/aud3.txt28 * http://www.geocities.com/SiliconValley/8682/aud3.txt
29 *
30 * Implementation note: There is no definite file signature for AUD files.
31 * The demuxer uses a probabilistic strategy for content detection. This
32 * entails performing sanity checks on certain header values in order to
33 * qualify a file. Refer to wsaud_probe() for the precise parameters.
34 */29 */
35 30
36#include "libavutil/intreadwrite.h"31#include "libavutil/intreadwrite.h"
37#include "avformat.h"32#include "avformat.h"
38#include "internal.h"33#include "internal.h"
39 34
40#define AUD_HEADER_SIZE 12
41#define AUD_CHUNK_PREAMBLE_SIZE 8
42#define AUD_CHUNK_SIGNATURE 0x0000DEAF
43
44#define FORM_TAG MKBETAG('F', 'O', 'R', 'M')35#define FORM_TAG MKBETAG('F', 'O', 'R', 'M')
45#define WVQA_TAG MKBETAG('W', 'V', 'Q', 'A')36#define WVQA_TAG MKBETAG('W', 'V', 'Q', 'A')
46#define VQHD_TAG MKBETAG('V', 'Q', 'H', 'D')37#define VQHD_TAG MKBETAG('V', 'Q', 'H', 'D')
@@ -63,15 +54,6 @@ · 16 unmodified lines
63#define VQA_FRAMERATE 1554#define VQA_FRAMERATE 15
64#define VQA_PREAMBLE_SIZE 855#define VQA_PREAMBLE_SIZE 8
65 56
66typedef struct WsAudDemuxContext {
67 int audio_samplerate;
68 int audio_channels;
69 int audio_bits;
70 enum CodecID audio_type;
71 int audio_stream_index;
72 int64_t audio_frame_counter;
73} WsAudDemuxContext;
74
75typedef struct WsVqaDemuxContext {57typedef struct WsVqaDemuxContext {
76 int audio_samplerate;58 int audio_samplerate;
77 int audio_channels;59 int audio_channels;
@@ -83,119 +65,6 @@ typedef struct WsVqaDemuxContext { · 5 unmodified lines
83 int64_t audio_frame_counter;65 int64_t audio_frame_counter;
84} WsVqaDemuxContext;66} WsVqaDemuxContext;
85 67
86static int wsaud_probe(AVProbeData *p)
87{
88 int field;
89
90 /* Probabilistic content detection strategy: There is no file signature
91 * so perform sanity checks on various header parameters:
92 * 8000 <= sample rate (16 bits) <= 48000 ==> 40001 acceptable numbers
93 * flags <= 0x03 (2 LSBs are used) ==> 4 acceptable numbers
94 * compression type (8 bits) = 1 or 99 ==> 2 acceptable numbers
95 * first audio chunk signature (32 bits) ==> 1 acceptable number
96 * The number space contains 2^64 numbers. There are 40001 * 4 * 2 * 1 =
97 * 320008 acceptable number combinations.
98 */
99
100 if (p->buf_size < AUD_HEADER_SIZE + AUD_CHUNK_PREAMBLE_SIZE)
101 return 0;
102
103 /* check sample rate */
104 field = AV_RL16(&p->buf[0]);
105 if ((field < 8000) || (field > 48000))
106 return 0;
107
108 /* enforce the rule that the top 6 bits of this flags field are reserved (0);
109 * this might not be true, but enforce it until deemed unnecessary */
110 if (p->buf[10] & 0xFC)
111 return 0;
112
113 /* note: only check for WS IMA (type 99) right now since there is no
114 * support for type 1 */
115 if (p->buf[11] != 99)
116 return 0;
117
118 /* read ahead to the first audio chunk and validate the first header signature */
119 if (AV_RL32(&p->buf[16]) != AUD_CHUNK_SIGNATURE)
120 return 0;
121
122 /* return 1/2 certainty since this file check is a little sketchy */
123 return AVPROBE_SCORE_MAX / 2;
124}
125
126static int wsaud_read_header(AVFormatContext *s,
127 AVFormatParameters *ap)
128{
129 WsAudDemuxContext *wsaud = s->priv_data;
130 AVIOContext *pb = s->pb;
131 AVStream *st;
132 unsigned char header[AUD_HEADER_SIZE];
133
134 if (avio_read(pb, header, AUD_HEADER_SIZE) != AUD_HEADER_SIZE)
135 return AVERROR(EIO);
136 wsaud->audio_samplerate = AV_RL16(&header[0]);
137 if (header[11] == 99)
138 wsaud->audio_type = CODEC_ID_ADPCM_IMA_WS;
139 else
140 return AVERROR_INVALIDDATA;
141
142 /* flag 0 indicates stereo */
143 wsaud->audio_channels = (header[10] & 0x1) + 1;
144 /* flag 1 indicates 16 bit audio */
145 wsaud->audio_bits = (((header[10] & 0x2) >> 1) + 1) * 8;
146
147 /* initialize the audio decoder stream */
148 st = avformat_new_stream(s, NULL);
149 if (!st)
150 return AVERROR(ENOMEM);
151 avpriv_set_pts_info(st, 33, 1, wsaud->audio_samplerate);
152 st->codec->codec_type = AVMEDIA_TYPE_AUDIO;
153 st->codec->codec_id = wsaud->audio_type;
154 st->codec->codec_tag = 0; /* no tag */
155 st->codec->channels = wsaud->audio_channels;
156 st->codec->sample_rate = wsaud->audio_samplerate;
157 st->codec->bits_per_coded_sample = wsaud->audio_bits;
158 st->codec->bit_rate = st->codec->channels * st->codec->sample_rate *
159 st->codec->bits_per_coded_sample / 4;
160 st->codec->block_align = st->codec->channels * st->codec->bits_per_coded_sample;
161
162 wsaud->audio_stream_index = st->index;
163 wsaud->audio_frame_counter = 0;
164
165 return 0;
166}
167
168static int wsaud_read_packet(AVFormatContext *s,
169 AVPacket *pkt)
170{
171 WsAudDemuxContext *wsaud = s->priv_data;
172 AVIOContext *pb = s->pb;
173 unsigned char preamble[AUD_CHUNK_PREAMBLE_SIZE];
174 unsigned int chunk_size;
175 int ret = 0;
176
177 if (avio_read(pb, preamble, AUD_CHUNK_PREAMBLE_SIZE) !=
178 AUD_CHUNK_PREAMBLE_SIZE)
179 return AVERROR(EIO);
180
181 /* validate the chunk */
182 if (AV_RL32(&preamble[4]) != AUD_CHUNK_SIGNATURE)
183 return AVERROR_INVALIDDATA;
184
185 chunk_size = AV_RL16(&preamble[0]);
186 ret= av_get_packet(pb, pkt, chunk_size);
187 if (ret != chunk_size)
188 return AVERROR(EIO);
189 pkt->stream_index = wsaud->audio_stream_index;
190 pkt->pts = wsaud->audio_frame_counter;
191 pkt->pts /= wsaud->audio_samplerate;
192
193 /* 2 samples/byte, 1 or 2 samples per frame depending on stereo */
194 wsaud->audio_frame_counter += (chunk_size * 2) / wsaud->audio_channels;
195
196 return ret;
197}
198
199static int wsvqa_probe(AVProbeData *p)68static int wsvqa_probe(AVProbeData *p)
200{69{
201 /* need 12 bytes to qualify */70 /* need 12 bytes to qualify */
@@ -367,17 +236,6 @@ static int wsvqa_read_packet(AVFormatContext *s, · 165 unmodified lines
367 return ret;236 return ret;
368}237}
369 238
370#if CONFIG_WSAUD_DEMUXER
371AVInputFormat ff_wsaud_demuxer = {
372 .name = "wsaud",
373 .long_name = NULL_IF_CONFIG_SMALL("Westwood Studios audio format"),
374 .priv_data_size = sizeof(WsAudDemuxContext),
375 .read_probe = wsaud_probe,
376 .read_header = wsaud_read_header,
377 .read_packet = wsaud_read_packet,
378};
379#endif
380#if CONFIG_WSVQA_DEMUXER
381AVInputFormat ff_wsvqa_demuxer = {239AVInputFormat ff_wsvqa_demuxer = {
382 .name = "wsvqa",240 .name = "wsvqa",
383 .long_name = NULL_IF_CONFIG_SMALL("Westwood Studios VQA format"),241 .long_name = NULL_IF_CONFIG_SMALL("Westwood Studios VQA format"),
@@ -386,4 +244,3 @@ AVInputFormat ff_wsvqa_demuxer = { · 2 unmodified lines
386 .read_header = wsvqa_read_header,244 .read_header = wsvqa_read_header,
387 .read_packet = wsvqa_read_packet,245 .read_packet = wsvqa_read_packet,
388};246};
389#endif

Keyboard shortcuts

General

Show keyboard shortcuts
?
Focus the current search
/

Issues and pull requests

Open filters
f

Issues, pull requests and Inbox

Select next item
j↓
Select previous item
k↑
Open selected item
Enter
Dismiss filters or clear selection
Esc

Repository code

Go to file
t