Width and height expressions in scale and scale2ref filters can now reference frame index, timestamp and packet position.
Unified · Split
| @@ -16139,6 +16139,19 @@ pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1. · 16138 unmodified lines | |||
| 16139 | @item ovsub | 16139 | @item ovsub |
| 16140 | horizontal and vertical output chroma subsample values. For example for the | 16140 | horizontal and vertical output chroma subsample values. For example for the |
| 16141 | pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1. | 16141 | pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1. |
| 16142 | |||
| 16143 | @item n | ||
| 16144 | The (sequential) number of the input frame, starting from 0. | ||
| 16145 | Only available with @code{eval=frame}. | ||
| 16146 | |||
| 16147 | @item t | ||
| 16148 | The presentation timestamp of the input frame, expressed as a number of | ||
| 16149 | seconds. Only available with @code{eval=frame}. | ||
| 16150 | |||
| 16151 | @item pos | ||
| 16152 | The position (byte offset) of the frame in the input stream, or NaN if | ||
| 16153 | this information is unavailable and/or meaningless (for example in case of synthetic video). | ||
| 16154 | Only available with @code{eval=frame}. | ||
| 16142 | @end table | 16155 | @end table |
| 16143 | 16156 | ||
| 16144 | @subsection Examples | 16157 | @subsection Examples |
| @@ -16362,6 +16375,19 @@ The main input video's display aspect ratio. Calculated from · 217 unmodified lines | |||
| 16362 | The main input video's horizontal and vertical chroma subsample values. | 16375 | The main input video's horizontal and vertical chroma subsample values. |
| 16363 | For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} | 16376 | For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} |
| 16364 | is 1. | 16377 | is 1. |
| 16378 | |||
| 16379 | @item main_n | ||
| 16380 | The (sequential) number of the main input frame, starting from 0. | ||
| 16381 | Only available with @code{eval=frame}. | ||
| 16382 | |||
| 16383 | @item main_t | ||
| 16384 | The presentation timestamp of the main input frame, expressed as a number of | ||
| 16385 | seconds. Only available with @code{eval=frame}. | ||
| 16386 | |||
| 16387 | @item main_pos | ||
| 16388 | The position (byte offset) of the frame in the main input stream, or NaN if | ||
| 16389 | this information is unavailable and/or meaningless (for example in case of synthetic video). | ||
| 16390 | Only available with @code{eval=frame}. | ||
| 16365 | @end table | 16391 | @end table |
| 16366 | 16392 | ||
| 16367 | @subsection Examples | 16393 | @subsection Examples |
| @@ -54,6 +54,9 @@ static const char *const var_names[] = { · 53 unmodified lines | |||
| 54 | "vsub", | 54 | "vsub", |
| 55 | "ohsub", | 55 | "ohsub", |
| 56 | "ovsub", | 56 | "ovsub", |
| 57 | "n", | ||
| 58 | "t", | ||
| 59 | "pos", | ||
| 57 | "main_w", | 60 | "main_w", |
| 58 | "main_h", | 61 | "main_h", |
| 59 | "main_a", | 62 | "main_a", |
| @@ -61,6 +64,9 @@ static const char *const var_names[] = { · 1 unmodified line | |||
| 61 | "main_dar", "mdar", | 64 | "main_dar", "mdar", |
| 62 | "main_hsub", | 65 | "main_hsub", |
| 63 | "main_vsub", | 66 | "main_vsub", |
| 67 | "main_n", | ||
| 68 | "main_t", | ||
| 69 | "main_pos", | ||
| 64 | NULL | 70 | NULL |
| 65 | }; | 71 | }; |
| 66 | 72 | ||
| @@ -76,6 +82,9 @@ enum var_name { · 9 unmodified lines | |||
| 76 | VAR_VSUB, | 82 | VAR_VSUB, |
| 77 | VAR_OHSUB, | 83 | VAR_OHSUB, |
| 78 | VAR_OVSUB, | 84 | VAR_OVSUB, |
| 85 | VAR_N, | ||
| 86 | VAR_T, | ||
| 87 | VAR_POS, | ||
| 79 | VAR_S2R_MAIN_W, | 88 | VAR_S2R_MAIN_W, |
| 80 | VAR_S2R_MAIN_H, | 89 | VAR_S2R_MAIN_H, |
| 81 | VAR_S2R_MAIN_A, | 90 | VAR_S2R_MAIN_A, |
| @@ -83,6 +92,9 @@ enum var_name { · 1 unmodified line | |||
| 83 | VAR_S2R_MAIN_DAR, VAR_S2R_MDAR, | 92 | VAR_S2R_MAIN_DAR, VAR_S2R_MDAR, |
| 84 | VAR_S2R_MAIN_HSUB, | 93 | VAR_S2R_MAIN_HSUB, |
| 85 | VAR_S2R_MAIN_VSUB, | 94 | VAR_S2R_MAIN_VSUB, |
| 95 | VAR_S2R_MAIN_N, | ||
| 96 | VAR_S2R_MAIN_T, | ||
| 97 | VAR_S2R_MAIN_POS, | ||
| 86 | VARS_NB | 98 | VARS_NB |
| 87 | }; | 99 | }; |
| 88 | 100 | ||
| @@ -184,11 +196,25 @@ static int check_exprs(AVFilterContext *ctx) · 95 unmodified lines | |||
| 184 | vars_w[VAR_S2R_MAIN_DAR] || vars_h[VAR_S2R_MAIN_DAR] || | 196 | vars_w[VAR_S2R_MAIN_DAR] || vars_h[VAR_S2R_MAIN_DAR] || |
| 185 | vars_w[VAR_S2R_MDAR] || vars_h[VAR_S2R_MDAR] || | 197 | vars_w[VAR_S2R_MDAR] || vars_h[VAR_S2R_MDAR] || |
| 186 | vars_w[VAR_S2R_MAIN_HSUB] || vars_h[VAR_S2R_MAIN_HSUB] || | 198 | vars_w[VAR_S2R_MAIN_HSUB] || vars_h[VAR_S2R_MAIN_HSUB] || |
| 187 | vars_w[VAR_S2R_MAIN_VSUB] || vars_h[VAR_S2R_MAIN_VSUB]) ) { | 199 | vars_w[VAR_S2R_MAIN_VSUB] || vars_h[VAR_S2R_MAIN_VSUB] || |
| 200 | vars_w[VAR_S2R_MAIN_N] || vars_h[VAR_S2R_MAIN_N] || | ||
| 201 | vars_w[VAR_S2R_MAIN_T] || vars_h[VAR_S2R_MAIN_T] || | ||
| 202 | vars_w[VAR_S2R_MAIN_POS] || vars_h[VAR_S2R_MAIN_POS]) ) { | ||
| 188 | av_log(ctx, AV_LOG_ERROR, "Expressions with scale2ref variables are not valid in scale filter.\n"); | 203 | av_log(ctx, AV_LOG_ERROR, "Expressions with scale2ref variables are not valid in scale filter.\n"); |
| 189 | return AVERROR(EINVAL); | 204 | return AVERROR(EINVAL); |
| 190 | } | 205 | } |
| 191 | 206 | ||
| 207 | if (scale->eval_mode == EVAL_MODE_INIT && | ||
| 208 | (vars_w[VAR_N] || vars_h[VAR_N] || | ||
| 209 | vars_w[VAR_T] || vars_h[VAR_T] || | ||
| 210 | vars_w[VAR_POS] || vars_h[VAR_POS] || | ||
| 211 | vars_w[VAR_S2R_MAIN_N] || vars_h[VAR_S2R_MAIN_N] || | ||
| 212 | vars_w[VAR_S2R_MAIN_T] || vars_h[VAR_S2R_MAIN_T] || | ||
| 213 | vars_w[VAR_S2R_MAIN_POS] || vars_h[VAR_S2R_MAIN_POS]) ) { | ||
| 214 | av_log(ctx, AV_LOG_ERROR, "Expressions with frame variables 'n', 't', 'pos' are not valid in init eval_mode.\n"); | ||
| 215 | return AVERROR(EINVAL); | ||
| 216 | } | ||
| 217 | |||
| 192 | return 0; | 218 | return 0; |
| 193 | } | 219 | } |
| 194 | 220 | ||
| @@ -622,6 +648,8 @@ static int scale_slice(AVFilterLink *link, AVFrame *out_buf, AVFrame *cur_pic, s · 427 unmodified lines | |||
| 622 | out,out_stride); | 648 | out,out_stride); |
| 623 | } | 649 | } |
| 624 | 650 | ||
| 651 | #define TS2T(ts, tb) ((ts) == AV_NOPTS_VALUE ? NAN : (double)(ts) * av_q2d(tb)) | ||
| 652 | |||
| 625 | static int scale_frame(AVFilterLink *link, AVFrame *in, AVFrame **frame_out) | 653 | static int scale_frame(AVFilterLink *link, AVFrame *in, AVFrame **frame_out) |
| 626 | { | 654 | { |
| 627 | AVFilterContext *ctx = link->dst; | 655 | AVFilterContext *ctx = link->dst; |
| @@ -643,10 +671,20 @@ static int scale_frame(AVFilterLink *link, AVFrame *in, AVFrame **frame_out) · 15 unmodified lines | |||
| 643 | in->sample_aspect_ratio.den != link->sample_aspect_ratio.den || | 671 | in->sample_aspect_ratio.den != link->sample_aspect_ratio.den || |
| 644 | in->sample_aspect_ratio.num != link->sample_aspect_ratio.num; | 672 | in->sample_aspect_ratio.num != link->sample_aspect_ratio.num; |
| 645 | 673 | ||
| 646 | if (frame_changed || | 674 | if (scale->eval_mode == EVAL_MODE_FRAME || frame_changed) { |
| 647 | (scale->eval_mode == EVAL_MODE_FRAME && | ||
| 648 | ctx->filter == &ff_vf_scale2ref) ) { | ||
| 649 | int ret; | 675 | int ret; |
| 676 | unsigned vars_w[VARS_NB] = { 0 }, vars_h[VARS_NB] = { 0 }; | ||
| 677 | |||
| 678 | av_expr_count_vars(scale->w_pexpr, vars_w, VARS_NB); | ||
| 679 | av_expr_count_vars(scale->h_pexpr, vars_h, VARS_NB); | ||
| 680 | |||
| 681 | if (scale->eval_mode == EVAL_MODE_FRAME && | ||
| 682 | !frame_changed && | ||
| 683 | ctx->filter != &ff_vf_scale2ref && | ||
| 684 | !(vars_w[VAR_N] || vars_w[VAR_T] || vars_w[VAR_POS]) && | ||
| 685 | !(vars_h[VAR_N] || vars_h[VAR_T] || vars_h[VAR_POS]) && | ||
| 686 | scale->w && scale->h) | ||
| 687 | goto scale; | ||
| 650 | 688 | ||
| 651 | if (scale->eval_mode == EVAL_MODE_INIT) { | 689 | if (scale->eval_mode == EVAL_MODE_INIT) { |
| 652 | snprintf(buf, sizeof(buf)-1, "%d", outlink->w); | 690 | snprintf(buf, sizeof(buf)-1, "%d", outlink->w); |
| @@ -663,6 +701,16 @@ static int scale_frame(AVFilterLink *link, AVFrame *in, AVFrame **frame_out) · 10 unmodified lines | |||
| 663 | return ret; | 701 | return ret; |
| 664 | } | 702 | } |
| 665 | 703 | ||
| 704 | if (ctx->filter == &ff_vf_scale2ref) { | ||
| 705 | scale->var_values[VAR_S2R_MAIN_N] = link->frame_count_out; | ||
| 706 | scale->var_values[VAR_S2R_MAIN_T] = TS2T(in->pts, link->time_base); | ||
| 707 | scale->var_values[VAR_S2R_MAIN_POS] = in->pkt_pos == -1 ? NAN : in->pkt_pos; | ||
| 708 | } else { | ||
| 709 | scale->var_values[VAR_N] = link->frame_count_out; | ||
| 710 | scale->var_values[VAR_T] = TS2T(in->pts, link->time_base); | ||
| 711 | scale->var_values[VAR_POS] = in->pkt_pos == -1 ? NAN : in->pkt_pos; | ||
| 712 | } | ||
| 713 | |||
| 666 | link->dst->inputs[0]->format = in->format; | 714 | link->dst->inputs[0]->format = in->format; |
| 667 | link->dst->inputs[0]->w = in->width; | 715 | link->dst->inputs[0]->w = in->width; |
| 668 | link->dst->inputs[0]->h = in->height; | 716 | link->dst->inputs[0]->h = in->height; |
| @@ -674,6 +722,7 @@ static int scale_frame(AVFilterLink *link, AVFrame *in, AVFrame **frame_out) · 5 unmodified lines | |||
| 674 | return ret; | 722 | return ret; |
| 675 | } | 723 | } |
| 676 | 724 | ||
| 725 | scale: | ||
| 677 | if (!scale->sws) { | 726 | if (!scale->sws) { |
| 678 | *frame_out = in; | 727 | *frame_out = in; |
| 679 | return 0; | 728 | return 0; |
| @@ -780,6 +829,7 @@ static int filter_frame(AVFilterLink *link, AVFrame *in) · 100 unmodified lines | |||
| 780 | 829 | ||
| 781 | static int filter_frame_ref(AVFilterLink *link, AVFrame *in) | 830 | static int filter_frame_ref(AVFilterLink *link, AVFrame *in) |
| 782 | { | 831 | { |
| 832 | ScaleContext *scale = link->dst->priv; | ||
| 783 | AVFilterLink *outlink = link->dst->outputs[1]; | 833 | AVFilterLink *outlink = link->dst->outputs[1]; |
| 784 | int frame_changed; | 834 | int frame_changed; |
| 785 | 835 | ||
| @@ -799,6 +849,12 @@ static int filter_frame_ref(AVFilterLink *link, AVFrame *in) · 13 unmodified lines | |||
| 799 | config_props_ref(outlink); | 849 | config_props_ref(outlink); |
| 800 | } | 850 | } |
| 801 | 851 | ||
| 852 | if (scale->eval_mode == EVAL_MODE_FRAME) { | ||
| 853 | scale->var_values[VAR_N] = link->frame_count_out; | ||
| 854 | scale->var_values[VAR_T] = TS2T(in->pts, link->time_base); | ||
| 855 | scale->var_values[VAR_POS] = in->pkt_pos == -1 ? NAN : in->pkt_pos; | ||
| 856 | } | ||
| 857 | |||
| 802 | return ff_filter_frame(outlink, in); | 858 | return ff_filter_frame(outlink, in); |
| 803 | } | 859 | } |
| 804 | 860 | ||