January 2012 - ffmpeg-devel - lists.ffmpeg.org

Trans.: a64multienc.c and drawutils.c optimisations
by yann.lepetitcorps＠free.fr 02 Jan '12

02 Jan '12

Hi, I have begin to analyse FFMPEG 0.9 sources and think to has found somes areas that can to be a little optimised such as in a64multienc.c or drawutils.c For example in the a64multienc.c source file, the computation of the current src line adress can to be only make at each line, not for each pixel (cf. this divide by width the number of lines address into the luma computation) static void to_meta_with_crop(AVCodecContext *avctx, AVFrame *p, int *dest) { int blockx, blocky, x, y; int luma = 0; int height = FFMIN(avctx->height, C64YRES); int width = FFMIN(avctx->width , C64XRES); uint8_t *src = p->data[0]; uint8_t *src2; for (blocky = 0; blocky < C64YRES; blocky += 8) { for (blockx = 0; blockx < C64XRES; blockx += 8) { for (y = blocky; y < blocky + 8 && y < C64YRES; y++) { if ( y >= height) return; src2 = src + y * p->linesize[0]; for (x = blockx; x < blockx + 8 && x < C64XRES; x += 2) { /* if(x < width && y < height) { */ if ( x < width ) { /* build average over 2 pixels */ /* luma = (src[(x + 0 + y * p->linesize[0])] + src[(x + 1 + y * p->linesize[0])]) / 2; */ luma = ((int)(src2[x]) + (int)(src2[x+1])) / 2; /* write blocks as linear data now so they are suitable for elbg */ dest[0] = luma; } dest++; } } } } } I have too begin tomodify the ff_fill_line_with_color() func in drawutils.c, for to use differents paths for char, short, int or others filling types where the format is packed void memset8(char *dst, char *val, int num) { memset(dst, *val, num); } void memset16(short int *dst, short int *val, int num) { int i; short int set16 = *val; for(i=0;i<num;i++) *dst++ = set16; } void memset32(int *dst, int *val, int num) { int i; int set32 = *val; for(i=0;i<num;i++) *dst++ = set32; } void memset24(char *dst, char *src, int num) { int i; for (i = 0; i < num; i++) memcpy(dst + i * 3, src, 3); } void memset_sized(char *dst, char *src, int num, int stepsize) { int i; for (i = 0; i < num; i++) memcpy(dst + i * stepsize, src, stepsize); } int ff_fill_line_with_color(uint8_t *line[4], int pixel_step[4], int w, uint8_t dst_color[4], enum PixelFormat pix_fmt, uint8_t rgba_color[4], int *is_packed_rgba, uint8_t rgba_map_ptr[4]) { uint8_t rgba_map[4] = {0}; int i; const AVPixFmtDescriptor *pix_desc = &av_pix_fmt_descriptors[pix_fmt]; int hsub = pix_desc->log2_chroma_w; *is_packed_rgba = ff_fill_rgba_map(rgba_map, pix_fmt) >= 0; if (*is_packed_rgba) { pixel_step[0] = (av_get_bits_per_pixel(pix_desc))>>3; for (i = 0; i < 4; i++) dst_color[rgba_map[i]] = rgba_color[i]; line[0] = av_malloc(w * pixel_step[0]); /* for (i = 0; i < w; i++) memcpy(line[0] + i * pixel_step[0], dst_color, pixel_step[0]); */ swich(av_get_bits_per_pixel(pix_desc)) { case 8 : memset8((char *)line[0], (char *)dst_color, w); break; case 16 : memset16((short int *)line[0], (short int *)dst_color, w); break; case 24 : memset24((char *)line[0], (char *)dst_color, w); break; case 32 : memset32((int *)line[0], (int *)dst_color, w); break; default : memset_sized((char *)line[0], (char *)dst_color, w, pixel_step[0]); break; } if (rgba_map_ptr) memcpy(rgba_map_ptr, rgba_map, sizeof(rgba_map[0]) * 4); } ... return 0; } I think to spend some time to analyse the 0.9 FFMPEG code source for to find similars parts that can be upgraded => what are the FFMPEG parts that are the more CPU limited (for that I test to find some optimisations rooms into them) ? @+ Yannoo ----- Fin du message transféré -----

6 19

[RFC] lavfi: automatic buffer ordering
by Nicolas George 02 Jan '12

02 Jan '12

Hi. Mark's message yesterday on the users mailing list ( http://ffmpeg.org/pipermail/ffmpeg-user/2011-December/004045.html ) has shown that there may be actual problems when filters output frames unasked (like split does) while the other end is not capable to deal with it (like overlay; it is not a coincidence if it happens mostly with filters with more than one input or output). In this example, the filters enter an endless loop of one asking a frame on one link and the other sending one on the other, eating all stack and heap space. The problem can be easily fixed by inserting a fifo filter at the adequate place; actually, it is the whole purpose of the fifo filter. I believe it would be better if it was done automatically. For that, I have the following proposal, that I intend to implement, but I would like to have some feedback before: struct AVFilterPad { ... unsigned flags; }; /** * The pad is safe with regard to asynchronous pushes. * For output pads, that means that it will only push a buffer (start_frame * or filter_samples) as a result of a request. * For input pads, that means that it can accept a buffer at any time, * regardless of whether it has requested it. * If an unrequested buffer is pushed on a pad with this flag cleared, it * will be queued until the appropriate time. */ #define AVFILTER_PAD_FLAG_ASYNC_PUSH_SAFE 0x1 Note 1: I chose this polarity for the flag so that pads where the flag was forgotten or not yet added will be considered unsafe by default. Note 2: with this semantic, a simple one-to-one filter like scale will have a safe input and an unsafe output. Therefore, the combination "safe input + scale" will have its output unsafe even though it is actually safe. I believe it does not matter much. As for the implementation, rather than having the filtergraph building process automatically insert fifos, I believe it is simpler to add the proper checks and queuing near the end of avfilter_start_frame and avfilter_filter_samples, and the start of avfilter_request_frame and avfilter_poll_frame. Note 3: if it gets implemented this way, then the fifo filter becomes trivial. The queuing logic in astreamsync and amerge gets a lot simpler too. Regards, -- Nicolas George

2 1

[PATCH 0/5] silence various compiler warnings
by Jean First 02 Jan '12

02 Jan '12

silence various compiler warnings and other related changes. Jean First (5): af_astreamsync: remove unused variable mxfdec: change av_log formatting %lx to PRIx64 and %li to PRIi64 j2kdec: av_log formatting use %lx instead of %x srtdec: use the default format for snprintf mov: av_dlog use PRId64 instead of %ld libavcodec/j2kdec.c | 2 +- libavcodec/srtdec.c | 4 ++-- libavfilter/af_astreamsync.c | 1 - libavformat/mov.c | 2 +- libavformat/mxfdec.c | 10 +++++----- 5 files changed, 9 insertions(+), 10 deletions(-) -- 1.7.7.3

4 13

[PATCH] cosmetics: replace -1 with PIX_FMT_NONE
by Paul B Mahol 01 Jan '12

01 Jan '12

--- libavcodec/j2kdec.c | 2 +- libavcodec/mjpegenc.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/libavcodec/j2kdec.c b/libavcodec/j2kdec.c index d605b13..c4cf017 100644 --- a/libavcodec/j2kdec.c +++ b/libavcodec/j2kdec.c @@ -1078,5 +1078,5 @@ AVCodec ff_jpeg2000_decoder = { .capabilities = CODEC_CAP_EXPERIMENTAL, .long_name = NULL_IF_CONFIG_SMALL("JPEG 2000"), .pix_fmts = - (enum PixelFormat[]) {PIX_FMT_GRAY8, PIX_FMT_RGB24, -1} + (enum PixelFormat[]) {PIX_FMT_GRAY8, PIX_FMT_RGB24, PIX_FMT_NONE} }; diff --git a/libavcodec/mjpegenc.c b/libavcodec/mjpegenc.c index 0fd5e5a..454a804 100644 --- a/libavcodec/mjpegenc.c +++ b/libavcodec/mjpegenc.c @@ -490,6 +490,6 @@ AVCodec ff_amv_encoder = { .init = MPV_encode_init, .encode = amv_encode_picture, .close = MPV_encode_end, - .pix_fmts= (enum PixelFormat[]){PIX_FMT_YUVJ420P, PIX_FMT_YUVJ422P, -1}, + .pix_fmts= (enum PixelFormat[]){PIX_FMT_YUVJ420P, PIX_FMT_YUVJ422P, PIX_FMT_NONE}, .long_name = NULL_IF_CONFIG_SMALL("AMV Video"), }; -- 1.7.7

2 1

[PATCH]Add missing PIX_FMT_NONE to libopenjpegenc
by Carl-Eugen Hoyos 01 Jan '12

01 Jan '12

Hi! Attached patch fixes ticket #813. Please comment, Carl Eugen

4 4

[PATCH] Port tinterlace filter from MPlayer.
by stefano.sabatini-lala＠poste.it 01 Jan '12

01 Jan '12

--- configure | 1 + doc/filters.texi | 26 +++++ libavfilter/Makefile | 1 + libavfilter/allfilters.c | 1 + libavfilter/avfilter.c | 2 +- libavfilter/vf_tinterlace.c | 260 +++++++++++++++++++++++++++++++++++++++++++ 6 files changed, 290 insertions(+), 1 deletions(-) create mode 100644 libavfilter/vf_tinterlace.c diff --git a/configure b/configure index 0538201..d3e2c14 100755 --- a/configure +++ b/configure @@ -1402,6 +1402,7 @@ udp_protocol_deps="network" # filters ocv_smooth_filter_deps="libopencv" +tinterlace_filter_deps="gpl" yadif_filter_deps="gpl" # libraries diff --git a/doc/filters.texi b/doc/filters.texi index d000276..3397989 100644 --- a/doc/filters.texi +++ b/doc/filters.texi @@ -390,6 +390,32 @@ not specified it will use the default value of 16. Adding this in the beginning of filter chains should make filtering faster due to better use of the memory cache. + at section tinterlace + +Perform temporal field interlacing - merge pairs of frames into an +interlaced frame, halving the framerate. Even frames are moved into +the upper field, odd frames to the lower field. + +The frames are counted starting from 1, so the first input frame is +considered odd. + +Accepts a parameters specifying the mode. Available modes are: + +0 Move odd frames into the upper field of the new frame, even into the + lower field, generating a double height frame at half framerate. + +1 Only output odd frames, even frames are dropped; height + unchanged, half framerate + +2 Only output even frames, odd frames are dropped; height + unchanged, half framerate + +3 Expand each frame to full height, but pad alternate lines with + black; framerate unchanged and double height. + +4 Interleave even lines from even frames with odd lines from odd + frames. Height unchanged at half framerate. + @section unsharp Sharpen or blur the input video. diff --git a/libavfilter/Makefile b/libavfilter/Makefile index 174b83e..431fe33 100644 --- a/libavfilter/Makefile +++ b/libavfilter/Makefile @@ -34,6 +34,7 @@ OBJS-$(CONFIG_PIXDESCTEST_FILTER) += vf_pixdesctest.o OBJS-$(CONFIG_PIXELASPECT_FILTER) += vf_aspect.o OBJS-$(CONFIG_SCALE_FILTER) += vf_scale.o OBJS-$(CONFIG_SLICIFY_FILTER) += vf_slicify.o +OBJS-$(CONFIG_TINTERLACE_FILTER) += vf_tinterlace.o OBJS-$(CONFIG_UNSHARP_FILTER) += vf_unsharp.o OBJS-$(CONFIG_VFLIP_FILTER) += vf_vflip.o OBJS-$(CONFIG_YADIF_FILTER) += vf_yadif.o diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c index ec7d933..5224fde 100644 --- a/libavfilter/allfilters.c +++ b/libavfilter/allfilters.c @@ -54,6 +54,7 @@ void avfilter_register_all(void) REGISTER_FILTER (PIXELASPECT, pixelaspect, vf); REGISTER_FILTER (SCALE, scale, vf); REGISTER_FILTER (SLICIFY, slicify, vf); + REGISTER_FILTER (TINTERLACE, tinterlace, vf); REGISTER_FILTER (UNSHARP, unsharp, vf); REGISTER_FILTER (VFLIP, vflip, vf); REGISTER_FILTER (YADIF, yadif, vf); diff --git a/libavfilter/avfilter.c b/libavfilter/avfilter.c index 221dfaa..72566e5 100644 --- a/libavfilter/avfilter.c +++ b/libavfilter/avfilter.c @@ -19,7 +19,7 @@ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ -/* #define DEBUG */ +#define DEBUG #include "libavcodec/audioconvert.c" #include "libavutil/pixdesc.h" diff --git a/libavfilter/vf_tinterlace.c b/libavfilter/vf_tinterlace.c new file mode 100644 index 0000000..771c6be --- /dev/null +++ b/libavfilter/vf_tinterlace.c @@ -0,0 +1,260 @@ +/* + * Copyright (C) 2010 Stefano Sabatini + * Copyright (C) 2003 Michael Zucchi <notzed at ximian.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * 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 General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with FFmpeg if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include "libavcore/imgutils.h" +#include "avfilter.h" + +typedef struct { + /** + * 0: Move odd frames into the upper field of the new image, even into the lower field, generating a double-height video at half framerate. + * 1: Only output odd frames, even frames are dropped; height unchanged, half framerate + * 2: Only output even frames, odd frames are dropped; height unchanged, half framerate + * 3: Expand each frame to double height, but pad alternate lines with black; framerate unchanged. + * 4: Interleave even lines from even frames with odd lines from odd frames. Height unchanged at half framerate. + */ + int mode; + + int vsub; ///< chroma vertical subsampling + int frame; ///< frame number + int frame_pending; ///< tell if there is a pending frame ready to be sent + AVFilterBufferRef *picref; ///< previously stored picref +} TInterlaceContext; + +static int query_formats(AVFilterContext *ctx) +{ + static const enum PixelFormat pix_fmts[] = { + PIX_FMT_YUV420P, PIX_FMT_YUV422P, PIX_FMT_YUV410P, + PIX_FMT_YUV444P, PIX_FMT_GRAY8, PIX_FMT_NONE + }; + + avfilter_set_common_formats(ctx, avfilter_make_format_list(pix_fmts)); + + return 0; +} + +static av_cold int init(AVFilterContext *ctx, const char *args, void *opaque) +{ + TInterlaceContext *tinterlace = ctx->priv; + tinterlace->mode = 0; + tinterlace->picref = NULL; + tinterlace->frame = 1; + tinterlace->frame_pending = 0; + + if (args) sscanf(args, "%d", &tinterlace->mode); + + if (tinterlace->mode < 0 || tinterlace->mode > 4) { + av_log(ctx, AV_LOG_ERROR, + "Unknown mode %d, use a value between 0 and 4\n", tinterlace->mode); + return AVERROR(EINVAL); + } + + return 0; +} + +static av_cold void uninit(AVFilterContext *ctx) +{ + TInterlaceContext *tinterlace = ctx->priv; + + if (tinterlace->picref) avfilter_unref_buffer(tinterlace->picref); +} + +static int config_out_props(AVFilterLink *outlink) +{ + AVFilterContext *ctx = outlink->src; + AVFilterLink *inlink = outlink->src->inputs[0]; + TInterlaceContext *tinterlace = ctx->priv; + + tinterlace->vsub = av_pix_fmt_descriptors[outlink->format].log2_chroma_h; + outlink->w = inlink->w; + outlink->h = (tinterlace->mode == 0 || tinterlace->mode == 3) ? inlink->h*2 : inlink->h; + + av_log(ctx, AV_LOG_INFO, "mode:%d in_h:%d out_h:%d\n", + tinterlace->mode, inlink->h, outlink->h); + + return 0; +} + +static void copy_picture_fields(uint8_t *dst[4], int dst_linesize[4], + const uint8_t *src[4], int src_linesize[4], + enum PixelFormat format, int w, int + src_h, int lower_fields, + int only_half_src_fields, int src_lower_fields) +{ + int i, k = 1, vsub = av_pix_fmt_descriptors[format].log2_chroma_h; + if (only_half_src_fields) { + src_h /= 2; + k = 2; + } + + for (i = 0; i < 3; i++) { + int h = i == 1 || i == 2 ? src_h >> vsub : src_h; + int linesize = av_image_get_linesize(format, w, i); + + if (lower_fields) + dst[i] += dst_linesize[i]; + if (only_half_src_fields && src_lower_fields) + src[i] += src_linesize[i]; + av_image_copy_plane(dst[i], dst_linesize[i]*2, + src[i], src_linesize[i]*k, linesize, h); + } +} + +static void end_frame(AVFilterLink *inlink) +{ + AVFilterContext *ctx = inlink->dst; + AVFilterLink *outlink = ctx->outputs[0]; + TInterlaceContext *tinterlace = ctx->priv; + AVFilterBufferRef *outpicref; + AVFilterBufferRef *picref = inlink->cur_buf; + + if (tinterlace->mode == 0) { + if (!tinterlace->picref) { + outpicref = avfilter_get_video_buffer(outlink, AV_PERM_WRITE, outlink->w, outlink->h); + avfilter_copy_buffer_ref_props(outpicref, picref); + outpicref->video->w = outlink->w; + outpicref->video->h = outlink->h; + outpicref->video->interlaced = 1; + outpicref->video->top_field_first = 1; + + /* write odd frame fields into the upper fields of the new frame */ + copy_picture_fields(outpicref->data, outpicref->linesize, + picref ->data, picref ->linesize, + outlink->format, inlink->w, inlink->h, 0, 0, 0); + tinterlace->picref = outpicref; + } else { + outpicref = tinterlace->picref; + + /* write even frame fields into the lower fields of the new frame */ + copy_picture_fields(outpicref->data, outpicref->linesize, + picref ->data, picref ->linesize, + outlink->format, inlink->w, inlink->h, 1, 0, 0); + tinterlace->frame_pending = 1; + } + } else if ((tinterlace->mode == 1 && tinterlace->frame & 1) || /* send only odd frames */ + (tinterlace->mode == 2 && !(tinterlace->frame & 1))) { /* send only even frames */ + tinterlace->picref = avfilter_ref_buffer(picref, AV_PERM_READ); + tinterlace->frame_pending = 1; + } else if (tinterlace->mode == 3) { + outpicref = avfilter_get_video_buffer(outlink, AV_PERM_WRITE, outlink->w, outlink->h); + avfilter_copy_buffer_ref_props(outpicref, picref); + outpicref->video->w = outlink->w; + outpicref->video->h = outlink->h; + outpicref->video->interlaced = 1; + outpicref->video->top_field_first = tinterlace->frame & 1; + + /* cleanup the destination picture */ + memset(outpicref->data[0], 0, outpicref->linesize[0] * outlink->h); + if (outpicref->data[1]) + memset(outpicref->data[1], 128, outpicref->linesize[1] * (outlink->h >> tinterlace->vsub)); + if (outpicref->data[2]) + memset(outpicref->data[2], 128, outpicref->linesize[2] * (outlink->h >> tinterlace->vsub)); + + /* copy fields */ + copy_picture_fields(outpicref->data, outpicref->linesize, + picref ->data, picref ->linesize, + outlink->format, inlink->w, inlink->h, tinterlace->frame & 1, 0, 0); + tinterlace->picref = outpicref; + tinterlace->frame_pending = 1; + } else if (tinterlace->mode == 4) { + /* Interleave even lines from even frames with odd lines from odd frames. Height unchanged at half framerate. */ + if (!tinterlace->picref) { + outpicref = avfilter_get_video_buffer(outlink, AV_PERM_WRITE, outlink->w, outlink->h); + avfilter_copy_buffer_ref_props(outpicref, picref); + outpicref->video->w = outlink->w; + outpicref->video->h = outlink->h; + outpicref->video->interlaced = 1; + outpicref->video->top_field_first = 1; + + /* write odd frame odd fields into the upper fields of the new frame */ + copy_picture_fields(outpicref->data, outpicref->linesize, + picref ->data, picref ->linesize, + outlink->format, inlink->w, inlink->h, 0, 1, 0); + tinterlace->picref = outpicref; + } else { + outpicref = tinterlace->picref; + + /* write even frame even fields into the lower fields of the new frame */ + copy_picture_fields(outpicref->data, outpicref->linesize, + picref ->data, picref ->linesize, + outlink->format, inlink->w, inlink->h, 1, 1, 1); + tinterlace->frame_pending = 1; + } + } + + tinterlace->frame++; + avfilter_unref_buffer(inlink->cur_buf); +} + +static void null_start_frame(AVFilterLink *inlink, AVFilterBufferRef *picref) { } + +static int request_frame(AVFilterLink *outlink) +{ + AVFilterContext *ctx = outlink->src; + TInterlaceContext *tinterlace = ctx->priv; + + while (!tinterlace->frame_pending) { + int ret; + + if ((ret = avfilter_request_frame(outlink->src->inputs[0]))) + return ret; + } + + avfilter_start_frame(outlink, tinterlace->picref); + avfilter_draw_slice(outlink, 0, outlink->h, 1); + avfilter_end_frame(outlink); + tinterlace->picref = NULL; + tinterlace->frame_pending = 0; + return 0; +} + +static int poll_frame(AVFilterLink *outlink) +{ + TInterlaceContext *tinterlace = outlink->src->priv; + + return tinterlace->frame_pending; +} + +static void null_draw_slice(AVFilterLink *link, int y, int h, int slice_dir) { } + +AVFilter avfilter_vf_tinterlace = { + .name = "tinterlace", + .description = NULL_IF_CONFIG_SMALL("Perform temporal field interlacing."), + + .priv_size = sizeof(TInterlaceContext), + .init = init, + .uninit = uninit, + .query_formats = query_formats, + + .inputs = (AVFilterPad[]) {{ .name = "default", + .type = AVMEDIA_TYPE_VIDEO, + .start_frame = null_start_frame, + .draw_slice = null_draw_slice, + .end_frame = end_frame, }, + { .name = NULL}}, + + .outputs = (AVFilterPad[]) {{ .name = "default", + .type = AVMEDIA_TYPE_VIDEO, + .config_props = config_out_props, + .poll_frame = poll_frame, + .request_frame = request_frame, }, + + { .name = NULL}}, +}; -- 1.7.1

9 25

[PATCH 0/3] h264 random access fixes
by Michael Niedermayer 01 Jan '12

01 Jan '12

This patch series fixes some issues with trash being displayed after seeking Please test, especially with samples that had issues with seeking in the past I will commit this tomorrow if noone finds a bug in it or has objections.

1 4

[PATCH] vf_tinterlace: implement interlace mode 5
by Stefano Sabatini 01 Jan '12

01 Jan '12

Allow creating interlaced bottom field first video. --- doc/filters.texi | 4 ++++ libavfilter/vf_tinterlace.c | 23 +++++++++++++---------- 2 files changed, 17 insertions(+), 10 deletions(-) diff --git a/doc/filters.texi b/doc/filters.texi index 387a05e..17983e4 100644 --- a/doc/filters.texi +++ b/doc/filters.texi @@ -2463,6 +2463,10 @@ generating a frame with double height at the same input framerate. @item 4 Interleave the upper field from odd frames with the lower field from even frames, generating a frame with unchanged height at half framerate. + +@item 5 +Interleave the lower field from odd frames with the upper field from +even frames, generating a frame with unchanged height at half framerate. @end table Default mode is 0. diff --git a/libavfilter/vf_tinterlace.c b/libavfilter/vf_tinterlace.c index 7ba164e..8407d1f 100644 --- a/libavfilter/vf_tinterlace.c +++ b/libavfilter/vf_tinterlace.c @@ -68,9 +68,9 @@ static av_cold int init(AVFilterContext *ctx, const char *args, void *opaque) if (args) { n = sscanf(args, "%d", &tinterlace->mode); - if (n != 1 || tinterlace->mode < 0 || tinterlace->mode > 4) { + if (n != 1 || tinterlace->mode < 0 || tinterlace->mode > 5) { av_log(ctx, AV_LOG_ERROR, - "Invalid mode '%s', use an integer between 0 and 4\n", args); + "Invalid mode '%s', use an integer between 0 and 5\n", args); return AVERROR(EINVAL); } } @@ -179,7 +179,7 @@ static void end_frame(AVFilterLink *inlink) AVFilterBufferRef *cur = tinterlace->cur; AVFilterBufferRef *next = tinterlace->next; AVFilterBufferRef *out = NULL; - int field; + int field, tff; /* we need at least two frames */ if (!tinterlace->cur) @@ -234,23 +234,26 @@ static void end_frame(AVFilterLink *inlink) FIELD_UPPER_AND_LOWER, 1, !field); break; - case 4: /* interleave upper lines from odd frames with lower lines from even frames, - * halving the frame rate and preserving image height */ + /* interleave upper/lower lines from odd frames with lower/upper lines from even frames, + * halving the frame rate and preserving image height */ + case 4: /* top field first */ + case 5: /* bottom field first */ + tff = tinterlace->mode == 4; out = avfilter_get_video_buffer(outlink, AV_PERM_WRITE, outlink->w, outlink->h); avfilter_copy_buffer_ref_props(out, cur); out->video->interlaced = 1; - out->video->top_field_first = 1; + out->video->top_field_first = tff; - /* copy upper field from cur */ + /* copy upper/lower field from cur */ copy_picture_field(out->data, out->linesize, cur->data, cur->linesize, inlink->format, inlink->w, inlink->h, - FIELD_UPPER, 1, FIELD_UPPER); - /* copy lower fields from next */ + tff ? FIELD_UPPER : FIELD_LOWER, 1, tff ? FIELD_UPPER : FIELD_LOWER); + /* copy lower/upper field from next */ copy_picture_field(out->data, out->linesize, next->data, next->linesize, inlink->format, inlink->w, inlink->h, - FIELD_LOWER, 1, FIELD_LOWER); + tff ? FIELD_LOWER : FIELD_UPPER, 1, tff ? FIELD_LOWER : FIELD_UPPER); avfilter_unref_buffer(tinterlace->next); tinterlace->next = NULL; break; -- 1.7.5.4

2 2

new year bugs
by Michael Niedermayer 01 Jan '12

01 Jan '12

Hi Todays merge enabled the automatic cpu counting and thus multithreading. Anyone following gitlog probably already spoted a bunch of bugfixes from me But there are likely more things that need fixing. Thus if you spot a bug try disabling threads first These bugs arent really new, its just the threads=1 default that hid them. -- Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB Dictatorship naturally arises out of democracy, and the most aggravated form of tyranny and slavery out of the most extreme liberty. -- Plato

1 0

[PATCH] waveform monitor filter
by Mark Himsley 01 Jan '12

01 Jan '12

This is quite an early/raw version of a luminance waveform monitor. I'd like to add to it with vector-scope, RGB parade, etc. At this stage I'm just looking for some comments from you experienced folks whether I'm doing something crazy. I've been testing this filter with this, to overlay the waveform monitor on top of the original video (thanks to Nicholas for the fifo workaround). ffmpeg -i input.mov -vf "split[one][two];[two]wfm_luma,scale=256:256[wfm];[one]fifo,[wfm]overlay=0:0" -aspect 16:9 -vcodec dvvideo -acodec pcm_s16le -ac 2 -y output.mov Thanks. -- Mark

2 2