diff options
Diffstat (limited to 'libavcodec/ffv1.c')
-rw-r--r-- | libavcodec/ffv1.c | 955 |
1 files changed, 681 insertions, 274 deletions
diff --git a/libavcodec/ffv1.c b/libavcodec/ffv1.c index f290b940e8..0dedc0d589 100644 --- a/libavcodec/ffv1.c +++ b/libavcodec/ffv1.c @@ -1,22 +1,22 @@ /* * FFV1 codec for libavcodec * - * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at> + * Copyright (c) 2003-2012 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * 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. * - * Libav is distributed in the hope that it will be useful, + * 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 Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -26,7 +26,13 @@ */ #include "libavutil/avassert.h" +#include "libavutil/crc.h" +#include "libavutil/opt.h" +#include "libavutil/imgutils.h" +#include "libavutil/pixdesc.h" +#include "libavutil/timer.h" #include "avcodec.h" +#include "internal.h" #include "get_bits.h" #include "put_bits.h" #include "dsputil.h" @@ -34,6 +40,11 @@ #include "golomb.h" #include "mathops.h" +#ifdef __INTEL_COMPILER +#undef av_flatten +#define av_flatten +#endif + #define MAX_PLANES 4 #define CONTEXT_SIZE 32 @@ -119,12 +130,12 @@ static const int8_t quant11[256] = { }; static const uint8_t ver2_state[256] = { - 0, 10, 10, 10, 10, 16, 16, 16, 28, 16, 16, 29, 42, 49, 20, 49, - 59, 25, 26, 26, 27, 31, 33, 33, 33, 34, 34, 37, 67, 38, 39, 39, - 40, 40, 41, 79, 43, 44, 45, 45, 48, 48, 64, 50, 51, 52, 88, 52, - 53, 74, 55, 57, 58, 58, 74, 60, 101, 61, 62, 84, 66, 66, 68, 69, - 87, 82, 71, 97, 73, 73, 82, 75, 111, 77, 94, 78, 87, 81, 83, 97, - 85, 83, 94, 86, 99, 89, 90, 99, 111, 92, 93, 134, 95, 98, 105, 98, + 0, 10, 10, 10, 10, 16, 16, 16, 28, 16, 16, 29, 42, 49, 20, 49, + 59, 25, 26, 26, 27, 31, 33, 33, 33, 34, 34, 37, 67, 38, 39, 39, + 40, 40, 41, 79, 43, 44, 45, 45, 48, 48, 64, 50, 51, 52, 88, 52, + 53, 74, 55, 57, 58, 58, 74, 60, 101, 61, 62, 84, 66, 66, 68, 69, + 87, 82, 71, 97, 73, 73, 82, 75, 111, 77, 94, 78, 87, 81, 83, 97, + 85, 83, 94, 86, 99, 89, 90, 99, 111, 92, 93, 134, 95, 98, 105, 98, 105, 110, 102, 108, 102, 118, 103, 106, 106, 113, 109, 112, 114, 112, 116, 125, 115, 116, 117, 117, 126, 119, 125, 121, 121, 123, 145, 124, 126, 131, 127, 129, 165, 130, 132, 138, 133, 135, 145, 136, 137, 139, 146, 141, 143, 142, 144, 148, @@ -156,6 +167,7 @@ typedef struct PlaneContext { #define MAX_SLICES 256 typedef struct FFV1Context { + AVClass *class; AVCodecContext *avctx; RangeCoder c; GetBitContext gb; @@ -163,13 +175,18 @@ typedef struct FFV1Context { uint64_t rc_stat[256][2]; uint64_t (*rc_stat2[MAX_QUANT_TABLES])[32][2]; int version; + int minor_version; int width, height; int chroma_h_shift, chroma_v_shift; + int chroma_planes; + int transparency; int flags; int picture_number; AVFrame picture; + AVFrame last_picture; int plane_count; - int ac; // 1 = range coder <-> 0 = golomb rice + int ac; ///< 1=range coder <-> 0=golomb rice + int ac_byte_count; ///< number of bytes used for AC coding PlaneContext plane[MAX_PLANES]; int16_t quant_table[MAX_CONTEXT_INPUTS][256]; int16_t quant_tables[MAX_QUANT_TABLES][MAX_CONTEXT_INPUTS][256]; @@ -180,6 +197,10 @@ typedef struct FFV1Context { int colorspace; int16_t *sample_buffer; int gob_count; + int packed_at_lsb; + int ec; + int slice_damaged; + int key_frame_ok; int quant_table_count; @@ -193,6 +214,7 @@ typedef struct FFV1Context { int slice_height; int slice_x; int slice_y; + int bits_per_raw_sample; } FFV1Context; static av_always_inline int fold(int diff, int bits) @@ -261,10 +283,10 @@ static void find_best_state(uint8_t best_state[256][256], occ[j] = 1.0; for (k = 0; k < 256; k++) { double newocc[256] = { 0 }; - for (m = 0; m < 256; m++) + for (m = 1; m < 256; m++) if (occ[m]) { - len -= occ[m] * (p * l2tab[m] + - (1 - p) * l2tab[256 - m]); + len -=occ[m]*( p *l2tab[ m] + + (1-p)*l2tab[256-m]); } if (len < best_len[k]) { best_len[k] = len; @@ -272,7 +294,7 @@ static void find_best_state(uint8_t best_state[256][256], } for (m = 0; m < 256; m++) if (occ[m]) { - newocc[one_state[m]] += occ[m] * p; + newocc[ one_state[ m]] += occ[m] * p; newocc[256 - one_state[256 - m]] += occ[m] * (1 - p); } memcpy(occ, newocc, sizeof(occ)); @@ -407,7 +429,7 @@ static inline void put_vlc_symbol(PutBitContext *pb, VlcState *const state, i += i; } - assert(k <= 8); + av_assert2(k<=13); #if 0 // JPEG LS if (k == 0 && 2 * state->drift <= -state->count) @@ -562,59 +584,78 @@ static void encode_plane(FFV1Context *s, uint8_t *src, int w, int h, for (i = 0; i < ring_size; i++) sample[i] = s->sample_buffer + (w + 6) * ((h + i - y) % ring_size) + 3; - sample[0][-1] = sample[1][0]; - sample[1][w] = sample[1][w - 1]; + sample[0][-1]= sample[1][0 ]; + sample[1][ w]= sample[1][w-1]; // { START_TIMER - if (s->avctx->bits_per_raw_sample <= 8) { + if (s->bits_per_raw_sample <= 8) { for (x = 0; x < w; x++) sample[0][x] = src[x + stride * y]; encode_line(s, w, sample, plane_index, 8); } else { - for (x = 0; x < w; x++) - sample[0][x] = ((uint16_t *)(src + stride * y))[x] >> - (16 - s->avctx->bits_per_raw_sample); - encode_line(s, w, sample, plane_index, s->avctx->bits_per_raw_sample); + if (s->packed_at_lsb) { + for (x = 0; x < w; x++) { + sample[0][x] = ((uint16_t*)(src + stride*y))[x]; + } + } else { + for (x = 0; x < w; x++) { + sample[0][x] = ((uint16_t*)(src + stride*y))[x] >> (16 - s->bits_per_raw_sample); + } + } + encode_line(s, w, sample, plane_index, s->bits_per_raw_sample); } // STOP_TIMER("encode line") } } } -static void encode_rgb_frame(FFV1Context *s, uint32_t *src, int w, int h, - int stride) +static void encode_rgb_frame(FFV1Context *s, uint8_t *src[3], int w, int h, int stride[3]) { int x, y, p, i; const int ring_size = s->avctx->context_model ? 3 : 2; - int16_t *sample[3][3]; + int16_t *sample[4][3]; + int lbd = s->avctx->bits_per_raw_sample <= 8; + int bits = s->avctx->bits_per_raw_sample > 0 ? s->avctx->bits_per_raw_sample : 8; + int offset = 1 << bits; s->run_index = 0; - memset(s->sample_buffer, 0, ring_size * 3 * (w + 6) * sizeof(*s->sample_buffer)); + memset(s->sample_buffer, 0, ring_size * 4 * (w + 6) * sizeof(*s->sample_buffer)); for (y = 0; y < h; y++) { for (i = 0; i < ring_size; i++) - for (p = 0; p < 3; p++) - sample[p][i] = s->sample_buffer + p * ring_size * (w + 6) + - ((h + i - y) % ring_size) * (w + 6) + 3; + for (p = 0; p < 4; p++) + sample[p][i]= s->sample_buffer + p*ring_size*(w+6) + ((h+i-y)%ring_size)*(w+6) + 3; for (x = 0; x < w; x++) { - int v = src[x + stride * y]; - int b = v & 0xFF; - int g = (v >> 8) & 0xFF; - int r = (v >> 16) & 0xFF; + int b, g, r, av_uninit(a); + if (lbd) { + unsigned v = *((uint32_t*)(src[0] + x*4 + stride[0]*y)); + b = v & 0xFF; + g = (v >> 8) & 0xFF; + r = (v >> 16) & 0xFF; + a = v >> 24; + } else { + b = *((uint16_t*)(src[0] + x*2 + stride[0]*y)); + g = *((uint16_t*)(src[1] + x*2 + stride[1]*y)); + r = *((uint16_t*)(src[2] + x*2 + stride[2]*y)); + } b -= g; r -= g; g += (b + r) >> 2; - b += 0x100; - r += 0x100; + b += offset; + r += offset; sample[0][0][x] = g; sample[1][0][x] = b; sample[2][0][x] = r; + sample[3][0][x] = a; } - for (p = 0; p < 3; p++) { - sample[p][0][-1] = sample[p][1][0]; - sample[p][1][w] = sample[p][1][w - 1]; - encode_line(s, w, sample[p], FFMIN(p, 1), 9); + for (p = 0; p < 3 + s->transparency; p++) { + sample[p][0][-1] = sample[p][1][0 ]; + sample[p][1][ w] = sample[p][1][w-1]; + if (lbd) + encode_line(s, w, sample[p], (p+1)/2, 9); + else + encode_line(s, w, sample[p], (p+1)/2, bits+1); } } } @@ -658,16 +699,16 @@ static void write_header(FFV1Context *f) put_symbol(c, state, f->state_transition[i] - c->one_state[i], 1); } - put_symbol(c, state, f->colorspace, 0); // YUV cs type + put_symbol(c, state, f->colorspace, 0); //YUV cs type if (f->version > 0) - put_symbol(c, state, f->avctx->bits_per_raw_sample, 0); - put_rac(c, state, 1); // chroma planes + put_symbol(c, state, f->bits_per_raw_sample, 0); + put_rac(c, state, f->chroma_planes); put_symbol(c, state, f->chroma_h_shift, 0); put_symbol(c, state, f->chroma_v_shift, 0); - put_rac(c, state, 0); // no transparency plane + put_rac(c, state, f->transparency); write_quant_tables(c, f->quant_table); - } else { + } else if (f->version < 3) { put_symbol(c, state, f->slice_count, 0); for (i = 0; i < f->slice_count; i++) { FFV1Context *fs = f->slice_context[i]; @@ -695,15 +736,19 @@ static av_cold int common_init(AVCodecContext *avctx) { FFV1Context *s = avctx->priv_data; + if (!avctx->width || !avctx->height) + return AVERROR_INVALIDDATA; + s->avctx = avctx; s->flags = avctx->flags; + avcodec_get_frame_defaults(&s->picture); + ff_dsputil_init(&s->dsp, avctx); s->width = avctx->width; s->height = avctx->height; - assert(s->width && s->height); // defaults s->num_h_slices = 1; s->num_v_slices = 1; @@ -711,12 +756,12 @@ static av_cold int common_init(AVCodecContext *avctx) return 0; } -static int init_slice_state(FFV1Context *f) +static int init_slice_state(FFV1Context *f, FFV1Context *fs) { - int i, j; + int j; - for (i = 0; i < f->slice_count; i++) { - FFV1Context *fs = f->slice_context[i]; + fs->plane_count = f->plane_count; + fs->transparency = f->transparency; for (j = 0; j < f->plane_count; j++) { PlaneContext *const p = &fs->plane[j]; @@ -737,15 +782,24 @@ static int init_slice_state(FFV1Context *f) if (fs->ac > 1) { // FIXME: only redo if state_transition changed for (j = 1; j < 256; j++) { - fs->c.one_state[j] = fs->state_transition[j]; + fs->c. one_state[ j] = f->state_transition[j]; fs->c.zero_state[256 - j] = 256 - fs->c.one_state[j]; } } - } return 0; } +static int init_slices_state(FFV1Context *f) { + int i; + for (i = 0; i < f->slice_count; i++) { + FFV1Context *fs = f->slice_context[i]; + if (init_slice_state(f, fs) < 0) + return -1; + } + return 0; +} + static av_cold int init_slice_contexts(FFV1Context *f) { int i; @@ -769,8 +823,7 @@ static av_cold int init_slice_contexts(FFV1Context *f) fs->slice_x = sxs; fs->slice_y = sys; - fs->sample_buffer = av_malloc(9 * (fs->width + 6) * - sizeof(*fs->sample_buffer)); + fs->sample_buffer = av_malloc(3*4 * (fs->width+6) * sizeof(*fs->sample_buffer)); if (!fs->sample_buffer) return AVERROR(ENOMEM); } @@ -799,6 +852,7 @@ static int write_extra_header(FFV1Context *f) uint8_t state[CONTEXT_SIZE]; int i, j, k; uint8_t state2[32][CONTEXT_SIZE]; + unsigned v; memset(state2, 128, sizeof(state2)); memset(state, 128, sizeof(state)); @@ -809,16 +863,21 @@ static int write_extra_header(FFV1Context *f) ff_build_rac_states(c, 0.05 * (1LL << 32), 256 - 8); put_symbol(c, state, f->version, 0); + if (f->version > 2) { + if (f->version == 3) + f->minor_version = 2; + put_symbol(c, state, f->minor_version, 0); + } put_symbol(c, state, f->ac, 0); if (f->ac > 1) for (i = 1; i < 256; i++) put_symbol(c, state, f->state_transition[i] - c->one_state[i], 1); put_symbol(c, state, f->colorspace, 0); // YUV cs type - put_symbol(c, state, f->avctx->bits_per_raw_sample, 0); - put_rac(c, state, 1); // chroma planes + put_symbol(c, state, f->bits_per_raw_sample, 0); + put_rac(c, state, f->chroma_planes); put_symbol(c, state, f->chroma_h_shift, 0); put_symbol(c, state, f->chroma_v_shift, 0); - put_rac(c, state, 0); // no transparency plane + put_rac(c, state, f->transparency); put_symbol(c, state, f->num_h_slices - 1, 0); put_symbol(c, state, f->num_v_slices - 1, 0); @@ -843,7 +902,14 @@ static int write_extra_header(FFV1Context *f) } } + if (f->version > 2) { + put_symbol(c, state, f->ec, 0); + } + f->avctx->extradata_size = ff_rac_terminate(c); + v = av_crc(av_crc_get_table(AV_CRC_32_IEEE), 0, f->avctx->extradata, f->avctx->extradata_size); + AV_WL32(f->avctx->extradata + f->avctx->extradata_size, v); + f->avctx->extradata_size += 4; return 0; } @@ -898,89 +964,167 @@ static int sort_stt(FFV1Context *s, uint8_t stt[256]) static av_cold int encode_init(AVCodecContext *avctx) { + const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(avctx->pix_fmt); FFV1Context *s = avctx->priv_data; int i, j, k, m; common_init(avctx); s->version = 0; - s->ac = avctx->coder_type ? 2 : 0; - if (s->ac > 1) - for (i = 1; i < 256; i++) - s->state_transition[i] = ver2_state[i]; + if ((avctx->flags & (CODEC_FLAG_PASS1|CODEC_FLAG_PASS2)) || avctx->slices>1) + s->version = FFMAX(s->version, 2); - s->plane_count = 2; - for (i = 0; i < 256; i++) { - s->quant_table_count = 2; - if (avctx->bits_per_raw_sample <= 8) { - s->quant_tables[0][0][i] = quant11[i]; - s->quant_tables[0][1][i] = quant11[i] * 11; - s->quant_tables[0][2][i] = quant11[i] * 11 * 11; - s->quant_tables[1][0][i] = quant11[i]; - s->quant_tables[1][1][i] = quant11[i] * 11; - s->quant_tables[1][2][i] = quant5[i] * 11 * 11; - s->quant_tables[1][3][i] = quant5[i] * 5 * 11 * 11; - s->quant_tables[1][4][i] = quant5[i] * 5 * 5 * 11 * 11; - } else { - s->quant_tables[0][0][i] = quant9_10bit[i]; - s->quant_tables[0][1][i] = quant9_10bit[i] * 11; - s->quant_tables[0][2][i] = quant9_10bit[i] * 11 * 11; - s->quant_tables[1][0][i] = quant9_10bit[i]; - s->quant_tables[1][1][i] = quant9_10bit[i] * 11; - s->quant_tables[1][2][i] = quant5_10bit[i] * 11 * 11; - s->quant_tables[1][3][i] = quant5_10bit[i] * 5 * 11 * 11; - s->quant_tables[1][4][i] = quant5_10bit[i] * 5 * 5 * 11 * 11; - } + if (avctx->level == 3) { + s->version = 3; } - s->context_count[0] = (11 * 11 * 11 + 1) / 2; - s->context_count[1] = (11 * 11 * 5 * 5 * 5 + 1) / 2; - memcpy(s->quant_table, s->quant_tables[avctx->context_model], - sizeof(s->quant_table)); - for (i = 0; i < s->plane_count; i++) { - PlaneContext *const p = &s->plane[i]; - - memcpy(p->quant_table, s->quant_table, sizeof(p->quant_table)); - p->quant_table_index = avctx->context_model; - p->context_count = s->context_count[p->quant_table_index]; + if (s->ec < 0) { + s->ec = (s->version >= 3); } - if (allocate_initial_states(s) < 0) - return AVERROR(ENOMEM); + if (s->version >= 2 && avctx->strict_std_compliance > FF_COMPLIANCE_EXPERIMENTAL) { + av_log(avctx, AV_LOG_ERROR, "Version 2 needed for requested features but version 2 is experimental and not enabled\n"); + return AVERROR_INVALIDDATA; + } - avctx->coded_frame = &s->picture; - switch (avctx->pix_fmt) { + s->ac = avctx->coder_type > 0 ? 2 : 0; + + s->plane_count = 3; + switch(avctx->pix_fmt) { + case AV_PIX_FMT_YUV444P9: + case AV_PIX_FMT_YUV422P9: + case AV_PIX_FMT_YUV420P9: + if (!avctx->bits_per_raw_sample) + s->bits_per_raw_sample = 9; + case AV_PIX_FMT_YUV444P10: + case AV_PIX_FMT_YUV420P10: + case AV_PIX_FMT_YUV422P10: + s->packed_at_lsb = 1; + if (!avctx->bits_per_raw_sample && !s->bits_per_raw_sample) + s->bits_per_raw_sample = 10; + case AV_PIX_FMT_GRAY16: case AV_PIX_FMT_YUV444P16: case AV_PIX_FMT_YUV422P16: case AV_PIX_FMT_YUV420P16: - if (avctx->bits_per_raw_sample <= 8) { + if (!avctx->bits_per_raw_sample && !s->bits_per_raw_sample) { + s->bits_per_raw_sample = 16; + } else if (!s->bits_per_raw_sample) { + s->bits_per_raw_sample = avctx->bits_per_raw_sample; + } + if (s->bits_per_raw_sample <= 8) { av_log(avctx, AV_LOG_ERROR, "bits_per_raw_sample invalid\n"); - return -1; + return AVERROR_INVALIDDATA; + } + if (!s->ac && avctx->coder_type == -1) { + av_log(avctx, AV_LOG_INFO, "bits_per_raw_sample > 8, forcing coder 1\n"); + s->ac = 2; } if (!s->ac) { - av_log(avctx, AV_LOG_ERROR, - "bits_per_raw_sample of more than 8 needs -coder 1 currently\n"); - return -1; + av_log(avctx, AV_LOG_ERROR, "bits_per_raw_sample of more than 8 needs -coder 1 currently\n"); + return AVERROR_INVALIDDATA; } s->version = FFMAX(s->version, 1); + case AV_PIX_FMT_GRAY8: case AV_PIX_FMT_YUV444P: + case AV_PIX_FMT_YUV440P: case AV_PIX_FMT_YUV422P: case AV_PIX_FMT_YUV420P: case AV_PIX_FMT_YUV411P: case AV_PIX_FMT_YUV410P: + s->chroma_planes = desc->nb_components < 3 ? 0 : 1; s->colorspace = 0; break; + case AV_PIX_FMT_YUVA444P: + case AV_PIX_FMT_YUVA422P: + case AV_PIX_FMT_YUVA420P: + s->chroma_planes = 1; + s->colorspace = 0; + s->transparency = 1; + break; case AV_PIX_FMT_RGB32: s->colorspace = 1; + s->transparency = 1; + break; + case AV_PIX_FMT_0RGB32: + s->colorspace = 1; + break; + case AV_PIX_FMT_GBRP9: + if (!avctx->bits_per_raw_sample) + s->bits_per_raw_sample = 9; + case AV_PIX_FMT_GBRP10: + if (!avctx->bits_per_raw_sample && !s->bits_per_raw_sample) + s->bits_per_raw_sample = 10; + case AV_PIX_FMT_GBRP12: + if (!avctx->bits_per_raw_sample && !s->bits_per_raw_sample) + s->bits_per_raw_sample = 12; + case AV_PIX_FMT_GBRP14: + if (!avctx->bits_per_raw_sample && !s->bits_per_raw_sample) + s->bits_per_raw_sample = 14; + else if (!s->bits_per_raw_sample) + s->bits_per_raw_sample = avctx->bits_per_raw_sample; + s->colorspace = 1; + s->chroma_planes = 1; + s->version = FFMAX(s->version, 1); break; default: av_log(avctx, AV_LOG_ERROR, "format not supported\n"); - return -1; + return AVERROR_INVALIDDATA; } - avcodec_get_chroma_sub_sample(avctx->pix_fmt, &s->chroma_h_shift, - &s->chroma_v_shift); + if (s->transparency) { + av_log(avctx, AV_LOG_WARNING, "Storing alpha plane, this will require a recent FFV1 decoder to playback!\n"); + } + if (avctx->context_model > 1U) { + av_log(avctx, AV_LOG_ERROR, "Invalid context model %d, valid values are 0 and 1\n", avctx->context_model); + return AVERROR(EINVAL); + } + + if (s->ac > 1) + for (i = 1; i < 256; i++) + s->state_transition[i] = ver2_state[i]; + + for (i = 0; i < 256; i++) { + s->quant_table_count = 2; + if (s->bits_per_raw_sample <= 8) { + s->quant_tables[0][0][i]= quant11[i]; + s->quant_tables[0][1][i]= 11*quant11[i]; + s->quant_tables[0][2][i]= 11*11*quant11[i]; + s->quant_tables[1][0][i]= quant11[i]; + s->quant_tables[1][1][i]= 11*quant11[i]; + s->quant_tables[1][2][i]= 11*11*quant5 [i]; + s->quant_tables[1][3][i]= 5*11*11*quant5 [i]; + s->quant_tables[1][4][i]= 5*5*11*11*quant5 [i]; + } else { + s->quant_tables[0][0][i]= quant9_10bit[i]; + s->quant_tables[0][1][i]= 11*quant9_10bit[i]; + s->quant_tables[0][2][i]= 11*11*quant9_10bit[i]; + s->quant_tables[1][0][i]= quant9_10bit[i]; + s->quant_tables[1][1][i]= 11*quant9_10bit[i]; + s->quant_tables[1][2][i]= 11*11*quant5_10bit[i]; + s->quant_tables[1][3][i]= 5*11*11*quant5_10bit[i]; + s->quant_tables[1][4][i]= 5*5*11*11*quant5_10bit[i]; + } + } + s->context_count[0] = (11 * 11 * 11 + 1) / 2; + s->context_count[1] = (11 * 11 * 5 * 5 * 5 + 1) / 2; + memcpy(s->quant_table, s->quant_tables[avctx->context_model], + sizeof(s->quant_table)); + + for (i = 0; i < s->plane_count; i++) { + PlaneContext *const p = &s->plane[i]; + + memcpy(p->quant_table, s->quant_table, sizeof(p->quant_table)); + p->quant_table_index = avctx->context_model; + p->context_count = s->context_count[p->quant_table_index]; + } + + if (allocate_initial_states(s) < 0) + return AVERROR(ENOMEM); + avctx->coded_frame = &s->picture; + if (!s->transparency) + s->plane_count = 2; + avcodec_get_chroma_sub_sample(avctx->pix_fmt, &s->chroma_h_shift, &s->chroma_v_shift); s->picture_number = 0; if (avctx->flags & (CODEC_FLAG_PASS1 | CODEC_FLAG_PASS2)) { @@ -999,7 +1143,7 @@ static av_cold int encode_init(AVCodecContext *avctx) av_assert0(s->version >= 2); - for (;; ) { + for (;;) { for (j = 0; j < 256; j++) for (i = 0; i < 2; i++) { s->rc_stat[j][i] = strtol(p, &next, 0); @@ -1019,15 +1163,15 @@ static av_cold int encode_init(AVCodecContext *avctx) av_log(avctx, AV_LOG_ERROR, "2Pass file invalid at %d %d %d %d [%s]\n", i, j, k, m, p); - return -1; + return AVERROR_INVALIDDATA; } p = next; } } gob_count = strtol(p, &next, 0); - if (next == p || gob_count < 0) { + if (next == p || gob_count <= 0) { av_log(avctx, AV_LOG_ERROR, "2Pass file invalid\n"); - return -1; + return AVERROR_INVALIDDATA; } p = next; while (*p == '\n' || *p == ' ') @@ -1056,14 +1200,21 @@ static av_cold int encode_init(AVCodecContext *avctx) } if (s->version > 1) { - s->num_h_slices = 2; - s->num_v_slices = 2; + for (s->num_v_slices = 2; s->num_v_slices < 9; s->num_v_slices++) { + for (s->num_h_slices = s->num_v_slices; s->num_h_slices < 2*s->num_v_slices; s->num_h_slices++) { + if (avctx->slices == s->num_h_slices * s->num_v_slices && avctx->slices <= 64 || !avctx->slices) + goto slices_ok; + } + } + av_log(avctx, AV_LOG_ERROR, "Unsupported number %d of slices requested, please specify a supported number with -slices (ex:4,6,9,12,16, ...)\n", avctx->slices); + return -1; + slices_ok: write_extra_header(s); } if (init_slice_contexts(s) < 0) return -1; - if (init_slice_state(s) < 0) + if (init_slices_state(s) < 0) return -1; #define STATS_OUT_SIZE 1024 * 1024 * 6 @@ -1085,12 +1236,11 @@ static av_cold int encode_init(AVCodecContext *avctx) #endif /* CONFIG_FFV1_ENCODER */ -static void clear_state(FFV1Context *f) + +static void clear_slice_state(FFV1Context *f, FFV1Context *fs) { - int i, si, j; + int i, j; - for (si = 0; si < f->slice_count; si++) { - FFV1Context *fs = f->slice_context[si]; for (i = 0; i < f->plane_count; i++) { PlaneContext *p = &fs->plane[i]; @@ -1112,10 +1262,31 @@ static void clear_state(FFV1Context *f) } } } - } } #if CONFIG_FFV1_ENCODER + +static void encode_slice_header(FFV1Context *f, FFV1Context *fs) +{ + RangeCoder *c = &fs->c; + uint8_t state[CONTEXT_SIZE]; + int j; + memset(state, 128, sizeof(state)); + + put_symbol(c, state, (fs->slice_x +1)*f->num_h_slices / f->width , 0); + put_symbol(c, state, (fs->slice_y +1)*f->num_v_slices / f->height , 0); + put_symbol(c, state, (fs->slice_width +1)*f->num_h_slices / f->width -1, 0); + put_symbol(c, state, (fs->slice_height+1)*f->num_v_slices / f->height-1, 0); + for (j=0; j<f->plane_count; j++) { + put_symbol(c, state, f->plane[j].quant_table_index, 0); + av_assert0(f->plane[j].quant_table_index == f->avctx->context_model); + } + if (!f->picture.interlaced_frame) put_symbol(c, state, 3, 0); + else put_symbol(c, state, 1 + !f->picture.top_field_first, 0); + put_symbol(c, state, f->picture.sample_aspect_ratio.num, 0); + put_symbol(c, state, f->picture.sample_aspect_ratio.den, 0); +} + static int encode_slice(AVCodecContext *c, void *arg) { FFV1Context *fs = *(void **)arg; @@ -1125,6 +1296,19 @@ static int encode_slice(AVCodecContext *c, void *arg) int x = fs->slice_x; int y = fs->slice_y; AVFrame *const p = &f->picture; + const int ps = av_pix_fmt_desc_get(c->pix_fmt)->comp[0].step_minus1 + 1; + + if (p->key_frame) + clear_slice_state(f, fs); + if (f->version > 2) { + encode_slice_header(f, fs); + } + if (!fs->ac) { + if (f->version > 2) + put_rac(&fs->c, (uint8_t[]) {129}, 0); + fs->ac_byte_count = f->version > 2 || (!x&&!y) ? ff_rac_terminate(&fs->c) : 0; + init_put_bits(&fs->pb, fs->c.bytestream_start + fs->ac_byte_count, fs->c.bytestream_end - fs->c.bytestream_start - fs->ac_byte_count); + } if (f->colorspace == 0) { const int chroma_width = -((-width) >> f->chroma_h_shift); @@ -1132,17 +1316,19 @@ static int encode_slice(AVCodecContext *c, void *arg) const int cx = x >> f->chroma_h_shift; const int cy = y >> f->chroma_v_shift; - encode_plane(fs, p->data[0] + x + y * p->linesize[0], - width, height, p->linesize[0], 0); + encode_plane(fs, p->data[0] + ps*x + y*p->linesize[0], width, height, p->linesize[0], 0); - encode_plane(fs, p->data[1] + cx + cy * p->linesize[1], - chroma_width, chroma_height, p->linesize[1], 1); - encode_plane(fs, p->data[2] + cx + cy * p->linesize[2], - chroma_width, chroma_height, p->linesize[2], 1); + if (f->chroma_planes) { + encode_plane(fs, p->data[1] + ps*cx+cy*p->linesize[1], chroma_width, chroma_height, p->linesize[1], 1); + encode_plane(fs, p->data[2] + ps*cx+cy*p->linesize[2], chroma_width, chroma_height, p->linesize[2], 1); + } + if (fs->transparency) + encode_plane(fs, p->data[3] + ps*x + y*p->linesize[3], width, height, p->linesize[3], 2); } else { - encode_rgb_frame(fs, (uint32_t *)(p->data[0]) + - x + y * (p->linesize[0] / 4), - width, height, p->linesize[0] / 4); + uint8_t *planes[3] = {p->data[0] + ps*x + y*p->linesize[0], + p->data[1] + ps*x + y*p->linesize[1], + p->data[2] + ps*x + y*p->linesize[2]}; + encode_rgb_frame(fs, planes, width, height, p->linesize); } emms_c(); @@ -1160,12 +1346,9 @@ static int encode_frame(AVCodecContext *avctx, AVPacket *pkt, uint8_t *buf_p; int i, ret; - if (!pkt->data && - (ret = av_new_packet(pkt, avctx->width * avctx->height * - ((8 * 2 + 1 + 1) * 4) / 8 + FF_MIN_BUFFER_SIZE)) < 0) { - av_log(avctx, AV_LOG_ERROR, "Error getting output packet.\n"); + if ((ret = ff_alloc_packet2(avctx, pkt, avctx->width*avctx->height*((8*2+1+1)*4)/8 + + FF_MIN_BUFFER_SIZE)) < 0) return ret; - } ff_init_range_encoder(c, pkt->data, pkt->size); ff_build_rac_states(c, 0.05 * (1LL << 32), 256 - 8); @@ -1178,17 +1361,12 @@ static int encode_frame(AVCodecContext *avctx, AVPacket *pkt, p->key_frame = 1; f->gob_count++; write_header(f); - clear_state(f); } else { put_rac(c, &keystate, 0); p->key_frame = 0; } - if (!f->ac) { - used_count += ff_rac_terminate(c); - init_put_bits(&f->slice_context[0]->pb, pkt->data + used_count, - pkt->size - used_count); - } else if (f->ac > 1) { + if (f->ac > 1) { int i; for (i = 1; i < 256; i++) { c->one_state[i] = f->state_transition[i]; @@ -1198,13 +1376,9 @@ static int encode_frame(AVCodecContext *avctx, AVPacket *pkt, for (i = 1; i < f->slice_count; i++) { FFV1Context *fs = f->slice_context[i]; - uint8_t *start = pkt->data + (pkt->size - used_count) * i / f->slice_count; + uint8_t *start = pkt->data + (pkt->size - used_count) * (int64_t)i / f->slice_count; int len = pkt->size / f->slice_count; - - if (fs->ac) - ff_init_range_encoder(&fs->c, start, len); - else - init_put_bits(&fs->pb, start, len); + ff_init_range_encoder(&fs->c, start, len); } avctx->execute(avctx, encode_slice, &f->slice_context[0], NULL, f->slice_count, sizeof(void *)); @@ -1215,21 +1389,26 @@ static int encode_frame(AVCodecContext *avctx, AVPacket *pkt, int bytes; if (fs->ac) { - uint8_t state = 128; + uint8_t state=129; put_rac(&fs->c, &state, 0); bytes = ff_rac_terminate(&fs->c); } else { flush_put_bits(&fs->pb); // FIXME: nicer padding - bytes = used_count + (put_bits_count(&fs->pb) + 7) / 8; - used_count = 0; + bytes = fs->ac_byte_count + (put_bits_count(&fs->pb) + 7)/8; } - if (i > 0) { + if (i > 0 || f->version > 2) { av_assert0(bytes < pkt->size / f->slice_count); - memmove(buf_p, fs->ac ? fs->c.bytestream_start : fs->pb.buf, bytes); + memmove(buf_p, fs->c.bytestream_start, bytes); av_assert0(bytes < (1 << 24)); AV_WB24(buf_p + bytes, bytes); bytes += 3; } + if (f->ec) { + unsigned v; + buf_p[bytes++] = 0; + v = av_crc(av_crc_get_table(AV_CRC_32_IEEE), 0, buf_p, bytes); + AV_WL32(buf_p + bytes, v); bytes += 4; + } buf_p += bytes; } @@ -1293,6 +1472,8 @@ static av_cold int common_end(AVCodecContext *avctx) if (avctx->codec->decode && s->picture.data[0]) avctx->release_buffer(avctx, &s->picture); + if (avctx->codec->decode && s->last_picture.data[0]) + avctx->release_buffer(avctx, &s->last_picture); for (j = 0; j < s->slice_count; j++) { FFV1Context *fs = s->slice_context[j]; @@ -1419,69 +1600,165 @@ static void decode_plane(FFV1Context *s, uint8_t *src, for (x = 0; x < w; x++) src[x + stride * y] = sample[1][x]; } else { - decode_line(s, w, sample, plane_index, - s->avctx->bits_per_raw_sample); - for (x = 0; x < w; x++) - ((uint16_t *)(src + stride * y))[x] = - sample[1][x] << (16 - s->avctx->bits_per_raw_sample); + decode_line(s, w, sample, plane_index, s->avctx->bits_per_raw_sample); + if (s->packed_at_lsb) { + for (x = 0; x < w; x++) { + ((uint16_t*)(src + stride*y))[x] = sample[1][x]; + } + } else { + for (x = 0; x < w; x++) { + ((uint16_t*)(src + stride*y))[x] = sample[1][x] << (16 - s->avctx->bits_per_raw_sample); + } + } } // STOP_TIMER("decode-line") } } } -static void decode_rgb_frame(FFV1Context *s, uint32_t *src, - int w, int h, int stride) +static void decode_rgb_frame(FFV1Context *s, uint8_t *src[3], int w, int h, int stride[3]) { int x, y, p; - int16_t *sample[3][2]; - for (x = 0; x < 3; x++) { + int16_t *sample[4][2]; + int lbd = s->avctx->bits_per_raw_sample <= 8; + int bits = s->avctx->bits_per_raw_sample > 0 ? s->avctx->bits_per_raw_sample : 8; + int offset = 1 << bits; + for (x = 0; x < 4; x++) { sample[x][0] = s->sample_buffer + x * 2 * (w + 6) + 3; sample[x][1] = s->sample_buffer + (x * 2 + 1) * (w + 6) + 3; } s->run_index = 0; - memset(s->sample_buffer, 0, 6 * (w + 6) * sizeof(*s->sample_buffer)); + memset(s->sample_buffer, 0, 8*(w+6)*sizeof(*s->sample_buffer)); for (y = 0; y < h; y++) { - for (p = 0; p < 3; p++) { + for (p = 0; p < 3 + s->transparency; p++) { int16_t *temp = sample[p][0]; // FIXME: try a normal buffer sample[p][0] = sample[p][1]; sample[p][1] = temp; - sample[p][1][-1] = sample[p][0][0]; - sample[p][0][w] = sample[p][0][w - 1]; - decode_line(s, w, sample[p], FFMIN(p, 1), 9); + sample[p][1][-1]= sample[p][0][0 ]; + sample[p][0][ w]= sample[p][0][w-1]; + if (lbd) + decode_line(s, w, sample[p], (p+1)/2, 9); + else + decode_line(s, w, sample[p], (p+1)/2, bits+1); } for (x = 0; x < w; x++) { int g = sample[0][1][x]; int b = sample[1][1][x]; int r = sample[2][1][x]; + int a = sample[3][1][x]; -// assert(g >= 0 && b >= 0 && r >= 0); -// assert(g < 256 && b < 512 && r < 512); - - b -= 0x100; - r -= 0x100; + b -= offset; + r -= offset; g -= (b + r) >> 2; b += g; r += g; - src[x + stride * y] = b + (g << 8) + (r << 16) + (0xFF << 24); + if (lbd) + *((uint32_t*)(src[0] + x*4 + stride[0]*y)) = b + (g<<8) + (r<<16) + (a<<24); + else { + *((uint16_t*)(src[0] + x*2 + stride[0]*y)) = b; + *((uint16_t*)(src[1] + x*2 + stride[1]*y)) = g; + *((uint16_t*)(src[2] + x*2 + stride[2]*y)) = r; + } } } } +static int decode_slice_header(FFV1Context *f, FFV1Context *fs) +{ + RangeCoder *c = &fs->c; + uint8_t state[CONTEXT_SIZE]; + unsigned ps, i, context_count; + memset(state, 128, sizeof(state)); + + av_assert0(f->version > 2); + + fs->slice_x = get_symbol(c, state, 0) * f->width ; + fs->slice_y = get_symbol(c, state, 0) * f->height; + fs->slice_width = (get_symbol(c, state, 0) + 1) * f->width + fs->slice_x; + fs->slice_height = (get_symbol(c, state, 0) + 1) * f->height + fs->slice_y; + + fs->slice_x /= f->num_h_slices; + fs->slice_y /= f->num_v_slices; + fs->slice_width = fs->slice_width /f->num_h_slices - fs->slice_x; + fs->slice_height = fs->slice_height/f->num_v_slices - fs->slice_y; + if ((unsigned)fs->slice_width > f->width || (unsigned)fs->slice_height > f->height) + return -1; + if ( (unsigned)fs->slice_x + (uint64_t)fs->slice_width > f->width + || (unsigned)fs->slice_y + (uint64_t)fs->slice_height > f->height) + return -1; + + for (i = 0; i < f->plane_count; i++) { + PlaneContext * const p = &fs->plane[i]; + int idx = get_symbol(c, state, 0); + if (idx > (unsigned)f->quant_table_count) { + av_log(f->avctx, AV_LOG_ERROR, "quant_table_index out of range\n"); + return -1; + } + p->quant_table_index = idx; + memcpy(p->quant_table, f->quant_tables[idx], sizeof(p->quant_table)); + context_count = f->context_count[idx]; + + if (p->context_count < context_count) { + av_freep(&p->state); + av_freep(&p->vlc_state); + } + p->context_count = context_count; + } + + ps = get_symbol(c, state, 0); + if (ps == 1) { + f->picture.interlaced_frame = 1; + f->picture.top_field_first = 1; + } else if (ps == 2) { + f->picture.interlaced_frame = 1; + f->picture.top_field_first = 0; + } else if (ps == 3) { + f->picture.interlaced_frame = 0; + } + f->picture.sample_aspect_ratio.num = get_symbol(c, state, 0); + f->picture.sample_aspect_ratio.den = get_symbol(c, state, 0); + + return 0; +} + static int decode_slice(AVCodecContext *c, void *arg) { FFV1Context *fs = *(void **)arg; FFV1Context *f = fs->avctx->priv_data; - int width = fs->slice_width; - int height = fs->slice_height; - int x = fs->slice_x; - int y = fs->slice_y; - AVFrame *const p = &f->picture; + int width, height, x, y; + const int ps = av_pix_fmt_desc_get(c->pix_fmt)->comp[0].step_minus1 + 1; + AVFrame * const p = &f->picture; + + if (f->version > 2) { + if (init_slice_state(f, fs) < 0) + return AVERROR(ENOMEM); + if (decode_slice_header(f, fs) < 0) { + fs->slice_damaged = 1; + return AVERROR_INVALIDDATA; + } + } + if (init_slice_state(f, fs) < 0) + return AVERROR(ENOMEM); + if (f->picture.key_frame) + clear_slice_state(f, fs); + width = fs->slice_width; + height= fs->slice_height; + x= fs->slice_x; + y= fs->slice_y; + + if (!fs->ac) { + if (f->version == 3 && f->minor_version > 1 || f->version > 3) + get_rac(&fs->c, (uint8_t[]) {129}); + fs->ac_byte_count = f->version > 2 || (!x&&!y) ? fs->c.bytestream - fs->c.bytestream_start - 1 : 0; + init_get_bits(&fs->gb, + fs->c.bytestream_start + fs->ac_byte_count, + (fs->c.bytestream_end - fs->c.bytestream_start - fs->ac_byte_count) * 8); + } av_assert1(width && height); if (f->colorspace == 0) { @@ -1489,18 +1766,28 @@ static int decode_slice(AVCodecContext *c, void *arg) const int chroma_height = -((-height) >> f->chroma_v_shift); const int cx = x >> f->chroma_h_shift; const int cy = y >> f->chroma_v_shift; + decode_plane(fs, p->data[0] + ps*x + y*p->linesize[0], width, height, p->linesize[0], 0); - decode_plane(fs, p->data[0] + x + y * p->linesize[0], - width, height, p->linesize[0], 0); - - decode_plane(fs, p->data[1] + cx + cy * p->linesize[1], - chroma_width, chroma_height, p->linesize[1], 1); - decode_plane(fs, p->data[2] + cx + cy * p->linesize[1], - chroma_width, chroma_height, p->linesize[2], 1); + if (f->chroma_planes) { + decode_plane(fs, p->data[1] + ps*cx+cy*p->linesize[1], chroma_width, chroma_height, p->linesize[1], 1); + decode_plane(fs, p->data[2] + ps*cx+cy*p->linesize[2], chroma_width, chroma_height, p->linesize[2], 1); + } + if (fs->transparency) + decode_plane(fs, p->data[3] + ps*x + y*p->linesize[3], width, height, p->linesize[3], 2); } else { - decode_rgb_frame(fs, - (uint32_t *)p->data[0] + x + y * (p->linesize[0] / 4), - width, height, p->linesize[0] / 4); + uint8_t *planes[3] = {p->data[0] + ps*x + y*p->linesize[0], + p->data[1] + ps*x + y*p->linesize[1], + p->data[2] + ps*x + y*p->linesize[2]}; + decode_rgb_frame(fs, planes, width, height, p->linesize); + } + if (fs->ac && f->version > 2) { + int v; + get_rac(&fs->c, (uint8_t[]) {129}); + v = fs->c.bytestream_end - fs->c.bytestream - 2 - 5*f->ec; + if (v) { + av_log(f->avctx, AV_LOG_ERROR, "bytestream end mismatching by %d\n", v); + fs->slice_damaged = 1; + } } emms_c(); @@ -1517,9 +1804,9 @@ static int read_quant_table(RangeCoder *c, int16_t *quant_table, int scale) memset(state, 128, sizeof(state)); for (v = 0; i < 128; v++) { - int len = get_symbol(c, state, 0) + 1; + unsigned len = get_symbol(c, state, 0) + 1; - if (len + i > 128) + if (len > 128 - i) return -1; while (len--) { @@ -1564,17 +1851,22 @@ static int read_extra_header(FFV1Context *f) ff_build_rac_states(c, 0.05 * (1LL << 32), 256 - 8); f->version = get_symbol(c, state, 0); - f->ac = f->avctx->coder_type = get_symbol(c, state, 0); + if (f->version > 2) { + c->bytestream_end -= 4; + f->minor_version = get_symbol(c, state, 0); + } + f->ac = f->avctx->coder_type = get_symbol(c, state, 0); if (f->ac > 1) for (i = 1; i < 256; i++) f->state_transition[i] = get_symbol(c, state, 1) + c->one_state[i]; + f->colorspace = get_symbol(c, state, 0); // YUV cs type f->avctx->bits_per_raw_sample = get_symbol(c, state, 0); - get_rac(c, state); // no chroma = false + f->chroma_planes = get_rac(c, state); f->chroma_h_shift = get_symbol(c, state, 0); f->chroma_v_shift = get_symbol(c, state, 0); - get_rac(c, state); // transparency plane - f->plane_count = 2; + f->transparency = get_rac(c, state); + f->plane_count = 2 + f->transparency; f->num_h_slices = 1 + get_symbol(c, state, 0); f->num_v_slices = 1 + get_symbol(c, state, 0); @@ -1608,19 +1900,38 @@ static int read_extra_header(FFV1Context *f) f->initial_states[i][j][k] = (pred + get_symbol(c, state2[k], 1)) & 0xFF; } + + if (f->version > 2) { + f->ec = get_symbol(c, state, 0); + } + + if (f->version > 2) { + unsigned v; + v = av_crc(av_crc_get_table(AV_CRC_32_IEEE), 0, f->avctx->extradata, f->avctx->extradata_size); + if (v) { + av_log(f->avctx, AV_LOG_ERROR, "CRC mismatch %X!\n", v); + return AVERROR_INVALIDDATA; + } + } + return 0; } static int read_header(FFV1Context *f) { uint8_t state[CONTEXT_SIZE]; - int i, j, context_count; + int i, j, context_count = -1; //-1 to avoid warning RangeCoder *const c = &f->slice_context[0]->c; memset(state, 128, sizeof(state)); if (f->version < 2) { - f->version = get_symbol(c, state, 0); + unsigned v= get_symbol(c, state, 0); + if (v >= 2) { + av_log(f->avctx, AV_LOG_ERROR, "invalid version %d in ver01 header\n", v); + return AVERROR_INVALIDDATA; + } + f->version = v; f->ac = f->avctx->coder_type = get_symbol(c, state, 0); if (f->ac > 1) for (i = 1; i < 256; i++) @@ -1628,46 +1939,65 @@ static int read_header(FFV1Context *f) f->colorspace = get_symbol(c, state, 0); // YUV cs type if (f->version > 0) f->avctx->bits_per_raw_sample = get_symbol(c, state, 0); - get_rac(c, state); // no chroma = false + f->chroma_planes = get_rac(c, state); f->chroma_h_shift = get_symbol(c, state, 0); f->chroma_v_shift = get_symbol(c, state, 0); - get_rac(c, state); // transparency plane - f->plane_count = 2; + f->transparency = get_rac(c, state); + f->plane_count = 2 + f->transparency; } if (f->colorspace == 0) { - if (f->avctx->bits_per_raw_sample <= 8) { - switch (16 * f->chroma_h_shift + f->chroma_v_shift) { - case 0x00: - f->avctx->pix_fmt = AV_PIX_FMT_YUV444P; - break; - case 0x10: - f->avctx->pix_fmt = AV_PIX_FMT_YUV422P; - break; - case 0x11: - f->avctx->pix_fmt = AV_PIX_FMT_YUV420P; - break; - case 0x20: - f->avctx->pix_fmt = AV_PIX_FMT_YUV411P; - break; - case 0x22: - f->avctx->pix_fmt = AV_PIX_FMT_YUV410P; - break; + if (!f->transparency && !f->chroma_planes) { + if (f->avctx->bits_per_raw_sample <= 8) + f->avctx->pix_fmt = AV_PIX_FMT_GRAY8; + else + f->avctx->pix_fmt = AV_PIX_FMT_GRAY16; + } else if (f->avctx->bits_per_raw_sample<=8 && !f->transparency) { + switch(16 * f->chroma_h_shift + f->chroma_v_shift) { + case 0x00: f->avctx->pix_fmt = AV_PIX_FMT_YUV444P; break; + case 0x01: f->avctx->pix_fmt = AV_PIX_FMT_YUV440P; break; + case 0x10: f->avctx->pix_fmt = AV_PIX_FMT_YUV422P; break; + case 0x11: f->avctx->pix_fmt = AV_PIX_FMT_YUV420P; break; + case 0x20: f->avctx->pix_fmt = AV_PIX_FMT_YUV411P; break; + case 0x22: f->avctx->pix_fmt = AV_PIX_FMT_YUV410P; break; + default: + av_log(f->avctx, AV_LOG_ERROR, "format not supported\n"); + return -1; + } + } else if (f->avctx->bits_per_raw_sample <= 8 && f->transparency) { + switch(16*f->chroma_h_shift + f->chroma_v_shift) { + case 0x00: f->avctx->pix_fmt = AV_PIX_FMT_YUVA444P; break; + case 0x10: f->avctx->pix_fmt = AV_PIX_FMT_YUVA422P; break; + case 0x11: f->avctx->pix_fmt = AV_PIX_FMT_YUVA420P; break; + default: + av_log(f->avctx, AV_LOG_ERROR, "format not supported\n"); + return -1; + } + } else if (f->avctx->bits_per_raw_sample == 9) { + f->packed_at_lsb = 1; + switch(16 * f->chroma_h_shift + f->chroma_v_shift) { + case 0x00: f->avctx->pix_fmt = AV_PIX_FMT_YUV444P9; break; + case 0x10: f->avctx->pix_fmt = AV_PIX_FMT_YUV422P9; break; + case 0x11: f->avctx->pix_fmt = AV_PIX_FMT_YUV420P9; break; + default: + av_log(f->avctx, AV_LOG_ERROR, "format not supported\n"); + return -1; + } + } else if (f->avctx->bits_per_raw_sample == 10) { + f->packed_at_lsb = 1; + switch(16 * f->chroma_h_shift + f->chroma_v_shift) { + case 0x00: f->avctx->pix_fmt = AV_PIX_FMT_YUV444P10; break; + case 0x10: f->avctx->pix_fmt = AV_PIX_FMT_YUV422P10; break; + case 0x11: f->avctx->pix_fmt = AV_PIX_FMT_YUV420P10; break; default: av_log(f->avctx, AV_LOG_ERROR, "format not supported\n"); return -1; } } else { - switch (16 * f->chroma_h_shift + f->chroma_v_shift) { - case 0x00: - f->avctx->pix_fmt = AV_PIX_FMT_YUV444P16; - break; - case 0x10: - f->avctx->pix_fmt = AV_PIX_FMT_YUV422P16; - break; - case 0x11: - f->avctx->pix_fmt = AV_PIX_FMT_YUV420P16; - break; + switch(16 * f->chroma_h_shift + f->chroma_v_shift) { + case 0x00: f->avctx->pix_fmt = AV_PIX_FMT_YUV444P16; break; + case 0x10: f->avctx->pix_fmt = AV_PIX_FMT_YUV422P16; break; + case 0x11: f->avctx->pix_fmt = AV_PIX_FMT_YUV420P16; break; default: av_log(f->avctx, AV_LOG_ERROR, "format not supported\n"); return -1; @@ -1679,7 +2009,17 @@ static int read_header(FFV1Context *f) "chroma subsampling not supported in this colorspace\n"); return -1; } - f->avctx->pix_fmt = AV_PIX_FMT_RGB32; + if ( f->avctx->bits_per_raw_sample == 9) + f->avctx->pix_fmt = AV_PIX_FMT_GBRP9; + else if (f->avctx->bits_per_raw_sample == 10) + f->avctx->pix_fmt = AV_PIX_FMT_GBRP10; + else if (f->avctx->bits_per_raw_sample == 12) + f->avctx->pix_fmt = AV_PIX_FMT_GBRP12; + else if (f->avctx->bits_per_raw_sample == 14) + f->avctx->pix_fmt = AV_PIX_FMT_GBRP14; + else + if (f->transparency) f->avctx->pix_fmt = AV_PIX_FMT_RGB32; + else f->avctx->pix_fmt = AV_PIX_FMT_0RGB32; } else { av_log(f->avctx, AV_LOG_ERROR, "colorspace not supported\n"); return -1; @@ -1694,19 +2034,33 @@ static int read_header(FFV1Context *f) av_log(f->avctx, AV_LOG_ERROR, "read_quant_table error\n"); return -1; } - } else { + } else if (f->version < 3) { f->slice_count = get_symbol(c, state, 0); - if (f->slice_count > (unsigned)MAX_SLICES) - return -1; + } else { + const uint8_t *p = c->bytestream_end; + for (f->slice_count = 0; f->slice_count < MAX_SLICES && 3 < p - c->bytestream_start; f->slice_count++) { + int trailer = 3 + 5*!!f->ec; + int size = AV_RB24(p-trailer); + if (size + trailer > p - c->bytestream_start) + break; + p -= size + trailer; + } + } + if (f->slice_count > (unsigned)MAX_SLICES || f->slice_count <= 0) { + av_log(f->avctx, AV_LOG_ERROR, "slice count %d is invalid\n", f->slice_count); + return -1; } for (j = 0; j < f->slice_count; j++) { FFV1Context *fs = f->slice_context[j]; fs->ac = f->ac; + fs->packed_at_lsb = f->packed_at_lsb; + + fs->slice_damaged = 0; - if (f->version >= 2) { - fs->slice_x = get_symbol(c, state, 0) * f->width; - fs->slice_y = get_symbol(c, state, 0) * f->height; + if (f->version == 2) { + fs->slice_x = get_symbol(c, state, 0) * f->width ; + fs->slice_y = get_symbol(c, state, 0) * f->height; fs->slice_width = (get_symbol(c, state, 0) + 1) * f->width + fs->slice_x; fs->slice_height = (get_symbol(c, state, 0) + 1) * f->height + fs->slice_y; @@ -1725,7 +2079,7 @@ static int read_header(FFV1Context *f) for (i = 0; i < f->plane_count; i++) { PlaneContext *const p = &fs->plane[i]; - if (f->version >= 2) { + if (f->version == 2) { int idx = get_symbol(c, state, 0); if (idx > (unsigned)f->quant_table_count) { av_log(f->avctx, AV_LOG_ERROR, @@ -1740,14 +2094,16 @@ static int read_header(FFV1Context *f) memcpy(p->quant_table, f->quant_table, sizeof(p->quant_table)); } - if (p->context_count < context_count) { - av_freep(&p->state); - av_freep(&p->vlc_state); + if (f->version <= 2) { + av_assert0(context_count >= 0); + if (p->context_count < context_count) { + av_freep(&p->state); + av_freep(&p->vlc_state); + } + p->context_count = context_count; } - p->context_count = context_count; } } - return 0; } @@ -1766,15 +2122,14 @@ static av_cold int decode_init(AVCodecContext *avctx) return 0; } -static int decode_frame(AVCodecContext *avctx, void *data, - int *data_size, AVPacket *avpkt) +static int decode_frame(AVCodecContext *avctx, void *data, int *data_size, AVPacket *avpkt) { const uint8_t *buf = avpkt->data; int buf_size = avpkt->size; FFV1Context *f = avctx->priv_data; RangeCoder *const c = &f->slice_context[0]->c; AVFrame *const p = &f->picture; - int bytes_read, i; + int i; uint8_t keystate = 128; const uint8_t *buf_p; @@ -1790,64 +2145,91 @@ static int decode_frame(AVCodecContext *avctx, void *data, p->pict_type = AV_PICTURE_TYPE_I; // FIXME: I vs. P if (get_rac(c, &keystate)) { p->key_frame = 1; + f->key_frame_ok = 0; if (read_header(f) < 0) return -1; - if (init_slice_state(f) < 0) - return -1; - - clear_state(f); + f->key_frame_ok = 1; } else { - p->key_frame = 0; - } - if (f->ac > 1) { - int i; - for (i = 1; i < 256; i++) { - c->one_state[i] = f->state_transition[i]; - c->zero_state[256 - i] = 256 - c->one_state[i]; + if (!f->key_frame_ok) { + av_log(avctx, AV_LOG_ERROR, "Cant decode non keyframe without valid keyframe\n"); + return AVERROR_INVALIDDATA; } + p->key_frame= 0; } - p->reference = 0; + p->reference = 3; //for error concealment if (avctx->get_buffer(avctx, p) < 0) { av_log(avctx, AV_LOG_ERROR, "get_buffer() failed\n"); return -1; } - if (avctx->debug & FF_DEBUG_PICT_INFO) - av_log(avctx, AV_LOG_ERROR, "keyframe:%d coder:%d\n", p->key_frame, f->ac); - - if (!f->ac) { - bytes_read = c->bytestream - c->bytestream_start - 1; - if (bytes_read == 0) - av_log(avctx, AV_LOG_ERROR, "error at end of AC stream\n"); // FIXME - init_get_bits(&f->slice_context[0]->gb, buf + bytes_read, - (buf_size - bytes_read) * 8); - } else { - bytes_read = 0; /* avoid warning */ - } + if (avctx->debug&FF_DEBUG_PICT_INFO) + av_log(avctx, AV_LOG_DEBUG, "ver:%d keyframe:%d coder:%d ec:%d slices:%d bps:%d\n", + f->version, p->key_frame, f->ac, f->ec, f->slice_count, f->avctx->bits_per_raw_sample); buf_p = buf + buf_size; - for (i = f->slice_count - 1; i > 0; i--) { + for (i = f->slice_count - 1; i >= 0; i--) { FFV1Context *fs = f->slice_context[i]; - int v = AV_RB24(buf_p - 3) + 3; - if (buf_p - buf <= v) { + int trailer = 3 + 5*!!f->ec; + int v; + + if (i || f->version>2) v = AV_RB24(buf_p-trailer)+trailer; + else v = buf_p - c->bytestream_start; + if (buf_p - c->bytestream_start < v) { av_log(avctx, AV_LOG_ERROR, "Slice pointer chain broken\n"); return -1; } buf_p -= v; - if (fs->ac) + + if (f->ec) { + unsigned crc = av_crc(av_crc_get_table(AV_CRC_32_IEEE), 0, buf_p, v); + if (crc) { + int64_t ts = avpkt->pts != AV_NOPTS_VALUE ? avpkt->pts : avpkt->dts; + av_log(f->avctx, AV_LOG_ERROR, "CRC mismatch %X!", crc); + if (ts != AV_NOPTS_VALUE && avctx->pkt_timebase.num) { + av_log(f->avctx, AV_LOG_ERROR, "at %f seconds\n", ts*av_q2d(avctx->pkt_timebase)); + } else if (ts != AV_NOPTS_VALUE) { + av_log(f->avctx, AV_LOG_ERROR, "at %"PRId64"\n", ts); + } else { + av_log(f->avctx, AV_LOG_ERROR, "\n"); + } + fs->slice_damaged = 1; + } + } + + if (i) { ff_init_range_decoder(&fs->c, buf_p, v); - else - init_get_bits(&fs->gb, buf_p, v * 8); + } else + fs->c.bytestream_end = (uint8_t *)(buf_p + v); + } + + avctx->execute(avctx, decode_slice, &f->slice_context[0], NULL, f->slice_count, sizeof(void*)); + + for (i=f->slice_count-1; i>=0; i--) { + FFV1Context *fs= f->slice_context[i]; + int j; + if (fs->slice_damaged && f->last_picture.data[0]) { + uint8_t *dst[4], *src[4]; + for (j=0; j<4; j++) { + int sh = (j==1 || j==2) ? f->chroma_h_shift : 0; + int sv = (j==1 || j==2) ? f->chroma_v_shift : 0; + dst[j] = f->picture .data[j] + f->picture .linesize[j]* + (fs->slice_y>>sv) + (fs->slice_x>>sh); + src[j] = f->last_picture.data[j] + f->last_picture.linesize[j]* + (fs->slice_y>>sv) + (fs->slice_x>>sh); + } + av_image_copy(dst, f->picture.linesize, (const uint8_t **)src, f->last_picture.linesize, + avctx->pix_fmt, fs->slice_width, fs->slice_height); + } } - avctx->execute(avctx, decode_slice, &f->slice_context[0], - NULL, f->slice_count, sizeof(void *)); f->picture_number++; *picture = *p; *data_size = sizeof(AVFrame); + FFSWAP(AVFrame, f->picture, f->last_picture); + return buf_size; } @@ -1865,6 +2247,26 @@ AVCodec ff_ffv1_decoder = { }; #if CONFIG_FFV1_ENCODER + +#define OFFSET(x) offsetof(FFV1Context, x) +#define VE AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_ENCODING_PARAM +static const AVOption options[] = { + { "slicecrc", "Protect slices with CRCs", OFFSET(ec), AV_OPT_TYPE_INT, {.i64 = -1}, -1, 1, VE}, +{NULL} +}; + +static const AVClass class = { + .class_name = "ffv1 encoder", + .item_name = av_default_item_name, + .option = options, + .version = LIBAVUTIL_VERSION_INT, +}; + +static const AVCodecDefault ffv1_defaults[] = { + { "coder", "-1" }, + { NULL }, +}; + AVCodec ff_ffv1_encoder = { .name = "ffv1", .type = AVMEDIA_TYPE_VIDEO, @@ -1874,13 +2276,18 @@ AVCodec ff_ffv1_encoder = { .encode2 = encode_frame, .close = common_end, .capabilities = CODEC_CAP_SLICE_THREADS, + .defaults = ffv1_defaults, .pix_fmts = (const enum AVPixelFormat[]) { - AV_PIX_FMT_YUV420P, AV_PIX_FMT_YUV422P, AV_PIX_FMT_YUV444P, - AV_PIX_FMT_YUV411P, AV_PIX_FMT_YUV410P, - AV_PIX_FMT_YUV420P16, AV_PIX_FMT_YUV422P16, AV_PIX_FMT_YUV444P16, - AV_PIX_FMT_RGB32, + AV_PIX_FMT_YUV420P, AV_PIX_FMT_YUVA420P, AV_PIX_FMT_YUVA422P, AV_PIX_FMT_YUV444P, + AV_PIX_FMT_YUVA444P, AV_PIX_FMT_YUV440P, AV_PIX_FMT_YUV422P, AV_PIX_FMT_YUV411P, + AV_PIX_FMT_YUV410P, AV_PIX_FMT_0RGB32, AV_PIX_FMT_RGB32, AV_PIX_FMT_YUV420P16, + AV_PIX_FMT_YUV422P16, AV_PIX_FMT_YUV444P16, AV_PIX_FMT_YUV444P9, AV_PIX_FMT_YUV422P9, + AV_PIX_FMT_YUV420P9, AV_PIX_FMT_YUV420P10, AV_PIX_FMT_YUV422P10, AV_PIX_FMT_YUV444P10, + AV_PIX_FMT_GRAY16, AV_PIX_FMT_GRAY8, AV_PIX_FMT_GBRP9, AV_PIX_FMT_GBRP10, + AV_PIX_FMT_GBRP12, AV_PIX_FMT_GBRP14, AV_PIX_FMT_NONE }, .long_name = NULL_IF_CONFIG_SMALL("FFmpeg video codec #1"), + .priv_class = &class, }; #endif |