https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/04b89e8ae33b https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/64e2fb3f9d89 https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/0463f5d6d56d https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/c5f314309067 https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/c33b4048859a https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/a2b090da7932 https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/1dddb930aaf0 https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/4e47ebf38b97 https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/50bc87263576 https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/d794b36a7788 https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/51c0b9e829be https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/e3c4442b249a https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/6fd1533057ff https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/ded0334d214f https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/70887d44ffa3 https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/fe100bc556d7 https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/fb70e0611bd7 https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/031f1561cd28 https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/1c6fd7d756af https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/96748ac54f99 https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/fe196fd29a67 --- configure.orig 2021-10-24 20:47:11 UTC +++ configure @@ -6430,7 +6430,7 @@ enabled libsrt && require_pkg_config libsrt enabled libssh && require_pkg_config libssh libssh libssh/sftp.h sftp_init enabled libspeex && require_pkg_config libspeex speex speex/speex.h speex_decoder_init enabled libsrt && require_pkg_config libsrt "srt >= 1.3.0" srt/srt.h srt_socket -enabled libsvtav1 && require_pkg_config libsvtav1 "SvtAv1Enc >= 0.8.4" EbSvtAv1Enc.h svt_av1_enc_init_handle +enabled libsvtav1 && require_pkg_config libsvtav1 "SvtAv1Enc >= 0.9.0" EbSvtAv1Enc.h svt_av1_enc_init_handle enabled libtensorflow && require libtensorflow tensorflow/c/c_api.h TF_Version -ltensorflow enabled libtesseract && require_pkg_config libtesseract tesseract tesseract/capi.h TessBaseAPICreate enabled libtheora && require libtheora theora/theoraenc.h th_info_init -ltheoraenc -ltheoradec -logg --- doc/encoders.texi.orig 2021-10-24 20:47:07 UTC +++ doc/encoders.texi @@ -1754,28 +1754,15 @@ Set the operating point tier. @item tier Set the operating point tier. -@item rc -Set the rate control mode to use. - -Possible modes: -@table @option -@item cqp -Constant quantizer: use fixed values of qindex (dependent on the frame type) -throughout the stream. This mode is the default. - -@item vbr -Variable bitrate: use a target bitrate for the whole stream. - -@item cvbr -Constrained variable bitrate: use a target bitrate for each GOP. -@end table - @item qmax Set the maximum quantizer to use when using a bitrate mode. @item qmin Set the minimum quantizer to use when using a bitrate mode. +@item crf +Constant rate factor value used in crf rate control mode (0-63). + @item qp Set the quantizer used in cqp rate control mode (0-63). @@ -1786,14 +1773,18 @@ Set number of frames to look ahead (0-120). Set number of frames to look ahead (0-120). @item preset -Set the quality-speed tradeoff, in the range 0 to 8. Higher values are -faster but lower quality. Defaults to 8 (highest speed). +Set the quality-speed tradeoff, in the range 0 to 13. Higher values are +faster but lower quality. @item tile_rows Set log2 of the number of rows of tiles to use (0-6). @item tile_columns Set log2 of the number of columns of tiles to use (0-4). + +@item svtav1-params +Set SVT-AV1 options using a list of @var{key}=@var{value} pairs separated +by ":". See the SVT-AV1 encoder user guide for a list of accepted parameters. @end table --- libavcodec/libsvtav1.c.orig 2021-10-24 20:47:07 UTC +++ libavcodec/libsvtav1.c @@ -60,17 +60,20 @@ typedef struct SvtContext { EOS_STATUS eos_flag; // User options. + AVDictionary *svtav1_opts; +#if FF_API_SVTAV1_OPTS int hierarchical_level; int la_depth; - int enc_mode; - int rc_mode; int scd; - int qp; int tier; int tile_columns; int tile_rows; +#endif + int enc_mode; + int crf; + int qp; } SvtContext; static const struct { @@ -120,16 +123,12 @@ static int alloc_buffer(EbSvtAv1EncConfiguration *conf static int alloc_buffer(EbSvtAv1EncConfiguration *config, SvtContext *svt_enc) { - const int pack_mode_10bit = - (config->encoder_bit_depth > 8) && (config->compressed_ten_bit_format == 0) ? 1 : 0; - const size_t luma_size_8bit = - config->source_width * config->source_height * (1 << pack_mode_10bit); - const size_t luma_size_10bit = - (config->encoder_bit_depth > 8 && pack_mode_10bit == 0) ? luma_size_8bit : 0; + const size_t luma_size = config->source_width * config->source_height * + (config->encoder_bit_depth > 8 ? 2 : 1); EbSvtIOFormat *in_data; - svt_enc->raw_size = (luma_size_8bit + luma_size_10bit) * 3 / 2; + svt_enc->raw_size = luma_size * 3 / 2; // allocate buffer for in and out svt_enc->in_buf = av_mallocz(sizeof(*svt_enc->in_buf)); @@ -151,11 +150,132 @@ static int config_enc_params(EbSvtAv1EncConfiguration { SvtContext *svt_enc = avctx->priv_data; const AVPixFmtDescriptor *desc; + AVDictionaryEntry *en = NULL; + // Update param from options +#if FF_API_SVTAV1_OPTS + if (svt_enc->hierarchical_level >= 0) + param->hierarchical_levels = svt_enc->hierarchical_level; + if (svt_enc->tier >= 0) + param->tier = svt_enc->tier; + if (svt_enc->scd >= 0) + param->scene_change_detection = svt_enc->scd; + if (svt_enc->tile_columns >= 0) + param->tile_columns = svt_enc->tile_columns; + if (svt_enc->tile_rows >= 0) + param->tile_rows = svt_enc->tile_rows; + + if (svt_enc->la_depth >= 0) + param->look_ahead_distance = svt_enc->la_depth; +#endif + + if (svt_enc->enc_mode >= 0) + param->enc_mode = svt_enc->enc_mode; + + if (avctx->bit_rate) { + param->target_bit_rate = avctx->bit_rate; + if (avctx->rc_max_rate != avctx->bit_rate) + param->rate_control_mode = 1; + else + param->rate_control_mode = 2; + + param->max_qp_allowed = avctx->qmax; + param->min_qp_allowed = avctx->qmin; + } + param->max_bit_rate = avctx->rc_max_rate; + if (avctx->bit_rate && avctx->rc_buffer_size) + param->maximum_buffer_size_ms = avctx->rc_buffer_size * 1000LL / avctx->bit_rate; + + if (svt_enc->crf > 0) { + param->qp = svt_enc->crf; + param->rate_control_mode = 0; + } else if (svt_enc->qp > 0) { + param->qp = svt_enc->qp; + param->rate_control_mode = 0; + param->enable_adaptive_quantization = 0; + } + + desc = av_pix_fmt_desc_get(avctx->pix_fmt); + param->color_primaries = avctx->color_primaries; + param->matrix_coefficients = (desc->flags & AV_PIX_FMT_FLAG_RGB) ? + AVCOL_SPC_RGB : avctx->colorspace; + param->transfer_characteristics = avctx->color_trc; + + if (avctx->color_range != AVCOL_RANGE_UNSPECIFIED) + param->color_range = avctx->color_range == AVCOL_RANGE_JPEG; + else + param->color_range = !!(desc->flags & AV_PIX_FMT_FLAG_RGB); + +#if SVT_AV1_CHECK_VERSION(1, 0, 0) + if (avctx->chroma_sample_location != AVCHROMA_LOC_UNSPECIFIED) { + const char *name = + av_chroma_location_name(avctx->chroma_sample_location); + + switch (avctx->chroma_sample_location) { + case AVCHROMA_LOC_LEFT: + param->chroma_sample_position = EB_CSP_VERTICAL; + break; + case AVCHROMA_LOC_TOPLEFT: + param->chroma_sample_position = EB_CSP_COLOCATED; + break; + default: + if (!name) + break; + + av_log(avctx, AV_LOG_WARNING, + "Specified chroma sample location %s is unsupported " + "on the AV1 bit stream level. Usage of a container that " + "allows passing this information - such as Matroska - " + "is recommended.\n", + name); + break; + } + } +#endif + + if (avctx->profile != FF_PROFILE_UNKNOWN) + param->profile = avctx->profile; + + if (avctx->level != FF_LEVEL_UNKNOWN) + param->level = avctx->level; + + if (avctx->gop_size > 0) + param->intra_period_length = avctx->gop_size - 1; + + if (avctx->framerate.num > 0 && avctx->framerate.den > 0) { + param->frame_rate_numerator = avctx->framerate.num; + param->frame_rate_denominator = avctx->framerate.den; + } else { + param->frame_rate_numerator = avctx->time_base.den; + param->frame_rate_denominator = avctx->time_base.num * avctx->ticks_per_frame; + } + + /* 2 = IDR, closed GOP, 1 = CRA, open GOP */ + param->intra_refresh_type = avctx->flags & AV_CODEC_FLAG_CLOSED_GOP ? 2 : 1; + +#if SVT_AV1_CHECK_VERSION(0, 9, 1) + while ((en = av_dict_get(svt_enc->svtav1_opts, "", en, AV_DICT_IGNORE_SUFFIX))) { + EbErrorType ret = svt_av1_enc_parse_parameter(param, en->key, en->value); + if (ret != EB_ErrorNone) { + int level = (avctx->err_recognition & AV_EF_EXPLODE) ? AV_LOG_ERROR : AV_LOG_WARNING; + av_log(avctx, level, "Error parsing option %s: %s.\n", en->key, en->value); + if (avctx->err_recognition & AV_EF_EXPLODE) + return AVERROR(EINVAL); + } + } +#else + if ((en = av_dict_get(svt_enc->svtav1_opts, "", NULL, AV_DICT_IGNORE_SUFFIX))) { + int level = (avctx->err_recognition & AV_EF_EXPLODE) ? AV_LOG_ERROR : AV_LOG_WARNING; + av_log(avctx, level, "svt-params needs libavcodec to be compiled with SVT-AV1 " + "headers >= 0.9.1.\n"); + if (avctx->err_recognition & AV_EF_EXPLODE) + return AVERROR(ENOSYS); + } +#endif + param->source_width = avctx->width; param->source_height = avctx->height; - desc = av_pix_fmt_desc_get(avctx->pix_fmt); param->encoder_bit_depth = desc->comp[0].depth; if (desc->log2_chroma_w == 1 && desc->log2_chroma_h == 1) @@ -169,12 +289,6 @@ static int config_enc_params(EbSvtAv1EncConfiguration return AVERROR(EINVAL); } - if (avctx->profile != FF_PROFILE_UNKNOWN) - param->profile = avctx->profile; - - if (avctx->level != FF_LEVEL_UNKNOWN) - param->level = avctx->level; - if ((param->encoder_color_format == EB_YUV422 || param->encoder_bit_depth > 10) && param->profile != FF_PROFILE_AV1_PROFESSIONAL ) { av_log(avctx, AV_LOG_WARNING, "Forcing Professional profile\n"); @@ -184,40 +298,21 @@ static int config_enc_params(EbSvtAv1EncConfiguration param->profile = FF_PROFILE_AV1_HIGH; } - // Update param from options - param->hierarchical_levels = svt_enc->hierarchical_level; - param->enc_mode = svt_enc->enc_mode; - param->tier = svt_enc->tier; - param->rate_control_mode = svt_enc->rc_mode; - param->scene_change_detection = svt_enc->scd; - param->qp = svt_enc->qp; + avctx->bit_rate = param->rate_control_mode > 0 ? + param->target_bit_rate : 0; + avctx->rc_max_rate = param->max_bit_rate; + avctx->rc_buffer_size = param->maximum_buffer_size_ms * avctx->bit_rate / 1000LL; - param->target_bit_rate = avctx->bit_rate; + if (avctx->bit_rate || avctx->rc_max_rate || avctx->rc_buffer_size) { + AVCPBProperties *cpb_props = ff_add_cpb_side_data(avctx); + if (!cpb_props) + return AVERROR(ENOMEM); - if (avctx->gop_size > 0) - param->intra_period_length = avctx->gop_size - 1; - - if (avctx->framerate.num > 0 && avctx->framerate.den > 0) { - param->frame_rate_numerator = avctx->framerate.num; - param->frame_rate_denominator = avctx->framerate.den; - } else { - param->frame_rate_numerator = avctx->time_base.den; - param->frame_rate_denominator = avctx->time_base.num * avctx->ticks_per_frame; + cpb_props->buffer_size = avctx->rc_buffer_size; + cpb_props->max_bitrate = avctx->rc_max_rate; + cpb_props->avg_bitrate = avctx->bit_rate; } - if (param->rate_control_mode) { - param->max_qp_allowed = avctx->qmax; - param->min_qp_allowed = avctx->qmin; - } - - param->intra_refresh_type = 2; /* Real keyframes only */ - - if (svt_enc->la_depth >= 0) - param->look_ahead_distance = svt_enc->la_depth; - - param->tile_columns = svt_enc->tile_columns; - param->tile_rows = svt_enc->tile_rows; - return 0; } @@ -330,11 +425,8 @@ static int eb_send_frame(AVCodecContext *avctx, const if (svt_enc->eos_flag == EOS_SENT) return 0; - headerPtrLast.n_alloc_len = 0; - headerPtrLast.n_filled_len = 0; - headerPtrLast.n_tick_count = 0; - headerPtrLast.p_app_private = NULL; - headerPtrLast.p_buffer = NULL; + memset(&headerPtrLast, 0, sizeof(headerPtrLast)); + headerPtrLast.pic_type = EB_AV1_INVALID_PICTURE; headerPtrLast.flags = EB_BUFFERFLAG_EOS; svt_av1_enc_send_picture(svt_enc->svt_handle, &headerPtrLast); @@ -350,6 +442,16 @@ static int eb_send_frame(AVCodecContext *avctx, const headerPtr->p_app_private = NULL; headerPtr->pts = frame->pts; + switch (frame->pict_type) { + case AV_PICTURE_TYPE_I: + headerPtr->pic_type = EB_AV1_KEY_PICTURE; + break; + default: + // Actually means auto, or default. + headerPtr->pic_type = EB_AV1_INVALID_PICTURE; + break; + } + svt_av1_enc_send_picture(svt_enc->svt_handle, headerPtr); return 0; @@ -472,21 +574,22 @@ static const AVOption options[] = { #define OFFSET(x) offsetof(SvtContext, x) #define VE AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_ENCODING_PARAM static const AVOption options[] = { - { "hielevel", "Hierarchical prediction levels setting", OFFSET(hierarchical_level), - AV_OPT_TYPE_INT, { .i64 = 4 }, 3, 4, VE , "hielevel"}, +#if FF_API_SVTAV1_OPTS + { "hielevel", "Hierarchical prediction levels setting (Deprecated, use svtav1-params)", OFFSET(hierarchical_level), + AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 4, VE | AV_OPT_FLAG_DEPRECATED , "hielevel"}, { "3level", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = 3 }, INT_MIN, INT_MAX, VE, "hielevel" }, { "4level", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = 4 }, INT_MIN, INT_MAX, VE, "hielevel" }, - { "la_depth", "Look ahead distance [0, 120]", OFFSET(la_depth), - AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 120, VE }, + { "la_depth", "Look ahead distance [0, 120] (Deprecated, use svtav1-params)", OFFSET(la_depth), + AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 120, VE | AV_OPT_FLAG_DEPRECATED }, - { "preset", "Encoding preset [0, 8]", - OFFSET(enc_mode), AV_OPT_TYPE_INT, { .i64 = MAX_ENC_PRESET }, 0, MAX_ENC_PRESET, VE }, - - { "tier", "Set operating point tier", OFFSET(tier), - AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 1, VE, "tier" }, + { "tier", "Set operating point tier (Deprecated, use svtav1-params)", OFFSET(tier), + AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 1, VE | AV_OPT_FLAG_DEPRECATED, "tier" }, { "main", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = 0 }, 0, 0, VE, "tier" }, { "high", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = 1 }, 0, 0, VE, "tier" }, +#endif + { "preset", "Encoding preset", + OFFSET(enc_mode), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, MAX_ENC_PRESET, VE }, FF_AV1_PROFILE_OPTS @@ -518,21 +621,20 @@ static const AVOption options[] = { { LEVEL("7.3", 73) }, #undef LEVEL - { "rc", "Bit rate control mode", OFFSET(rc_mode), - AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 3, VE , "rc"}, - { "cqp", "Constant quantizer", 0, AV_OPT_TYPE_CONST, { .i64 = 0 }, INT_MIN, INT_MAX, VE, "rc" }, - { "vbr", "Variable Bit Rate, use a target bitrate for the entire stream", 0, AV_OPT_TYPE_CONST, { .i64 = 1 }, INT_MIN, INT_MAX, VE, "rc" }, - { "cvbr", "Constrained Variable Bit Rate, use a target bitrate for each GOP", 0, AV_OPT_TYPE_CONST,{ .i64 = 2 }, INT_MIN, INT_MAX, VE, "rc" }, + { "crf", "Constant Rate Factor value", OFFSET(crf), + AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 63, VE }, + { "qp", "Initial Quantizer level value", OFFSET(qp), + AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 63, VE }, +#if FF_API_SVTAV1_OPTS + { "sc_detection", "Scene change detection (Deprecated, use svtav1-params)", OFFSET(scd), + AV_OPT_TYPE_BOOL, { .i64 = -1 }, -1, 1, VE | AV_OPT_FLAG_DEPRECATED }, - { "qp", "Quantizer to use with cqp rate control mode", OFFSET(qp), - AV_OPT_TYPE_INT, { .i64 = 50 }, 0, 63, VE }, + { "tile_columns", "Log2 of number of tile columns to use (Deprecated, use svtav1-params)", OFFSET(tile_columns), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 4, VE | AV_OPT_FLAG_DEPRECATED }, + { "tile_rows", "Log2 of number of tile rows to use (Deprecated, use svtav1-params)", OFFSET(tile_rows), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 6, VE | AV_OPT_FLAG_DEPRECATED }, +#endif - { "sc_detection", "Scene change detection", OFFSET(scd), - AV_OPT_TYPE_BOOL, { .i64 = 0 }, 0, 1, VE }, + { "svtav1-params", "Set the SVT-AV1 configuration using a :-separated list of key=value parameters", OFFSET(svtav1_opts), AV_OPT_TYPE_DICT, { 0 }, 0, 0, VE }, - { "tile_columns", "Log2 of number of tile columns to use", OFFSET(tile_columns), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 4, VE}, - { "tile_rows", "Log2 of number of tile rows to use", OFFSET(tile_rows), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 6, VE}, - {NULL}, }; @@ -544,9 +646,10 @@ static const AVCodecDefault eb_enc_defaults[] = { }; static const AVCodecDefault eb_enc_defaults[] = { - { "b", "7M" }, + { "b", "0" }, + { "flags", "+cgop" }, { "g", "-1" }, - { "qmin", "0" }, + { "qmin", "1" }, { "qmax", "63" }, { NULL }, }; @@ -561,12 +664,11 @@ AVCodec ff_libsvtav1_encoder = { .receive_packet = eb_receive_packet, .close = eb_enc_close, .capabilities = AV_CODEC_CAP_DELAY | AV_CODEC_CAP_OTHER_THREADS, - .caps_internal = FF_CODEC_CAP_AUTO_THREADS, + .caps_internal = FF_CODEC_CAP_AUTO_THREADS | FF_CODEC_CAP_INIT_CLEANUP, .pix_fmts = (const enum AVPixelFormat[]){ AV_PIX_FMT_YUV420P, AV_PIX_FMT_YUV420P10, AV_PIX_FMT_NONE }, .priv_class = &class, .defaults = eb_enc_defaults, - .caps_internal = FF_CODEC_CAP_INIT_CLEANUP, .wrapper_name = "libsvtav1", }; --- libavcodec/version.h.orig 2021-10-24 20:47:07 UTC +++ libavcodec/version.h @@ -168,5 +168,8 @@ #ifndef FF_API_INIT_PACKET #define FF_API_INIT_PACKET (LIBAVCODEC_VERSION_MAJOR < 60) #endif +#ifndef FF_API_SVTAV1_OPTS +#define FF_API_SVTAV1_OPTS (LIBAVCODEC_VERSION_MAJOR < 60) +#endif #endif /* AVCODEC_VERSION_H */