December 2009 - ffmpeg-devel - lists.ffmpeg.org

[PATCH] x4->params.analyse.b_weighted_bipred gets wrong bool value in libx264.c
by erik＠slagter.name 06 Dec '09

06 Dec '09

I am not sure about the _exact_ semantics of this field, but according to the source of x264 and compare runs it seems that value of this field must be either 0 or 1, and not 0 or 32 what ffmpeg uses. A shallow glance at the code inside libx264 suggests it does make a difference to use 32 or 1 for true, so I'd rather have it correct. Index: libavcodec/libx264.c =================================================================== --- libavcodec/libx264.c (revision 20746) +++ libavcodec/libx264.c (working copy) @@ -227,7 +227,7 @@ x4->params.analyse.i_direct_mv_pred = avctx->directpred; - x4->params.analyse.b_weighted_bipred = avctx->flags2 & CODEC_FLAG2_WPRED; + x4->params.analyse.b_weighted_bipred = !!(avctx->flags2 & CODEC_FLAG2_WPRED); x4->params.analyse.i_weighted_pred = avctx->weighted_p_pred; if (avctx->me_method == ME_EPZS)

2 2

[PATCH 2/6] RTMP output: small fixes
by piratfm＠gmail.com 06 Dec '09

06 Dec '09

2 5

Check return of url_fseek in av_seek_frame_binary
by sean.soria＠gmail.com 06 Dec '09

06 Dec '09

Calls to url_fseek should have their return value checked in av_seek_frame_binary, just as they do in av_seek_frame_generic. Otherwise, function may return success even though url_fseek reported failure. --- libavformat/utils.c (revision 20697) +++ libavformat/utils.c (working copy) @@ -1311,6 +1311,7 @@ int av_seek_frame_binary(AVFormatContext *s, int s int64_t av_uninit(pos_min), av_uninit(pos_max), pos, pos_limit; int64_t ts_min, ts_max, ts; int index; + int64_t ret; AVStream *st; if (stream_index < 0) @@ -1363,7 +1364,8 @@ int av_seek_frame_binary(AVFormatContext *s, int s return -1; /* do the seek */ - url_fseek(s->pb, pos, SEEK_SET); + if ((ret = url_fseek(s->pb, pos, SEEK_SET)) < 0) + return ret; av_update_cur_dts(s, st, ts);

3 3

[PATCH 1/6] RTMP output: move application part of mrl to protocol handler context
by piratfm＠gmail.com 05 Dec '09

05 Dec '09

2009/12/2 Kostya <kostya.shishkov at gmail.com>: > Part 5 - moving application path to context is fine, but replacing > parsing code chunk is not It's for wowza. If I trying to send to "rtmp://localhost/live/_definst_/doPublish=hackme/1tv0stream" (simple authorized). Current parser split path on this parts: Proto = rtmp, path = /live/_definst_/doPublish=hackme/test0, app = live, fname = _definst_/doPublish=hackme/test0 But sending works with this values: Proto = rtmp, path = /live/_definst_/doPublish=hackme/test0, app = live/_definst_/doPublish=hackme/, fname = test0

3 2

[PATCH] fail before printing output format
by baptiste.coudurier＠gmail.com 05 Dec '09

05 Dec '09

Guys, Patch makes ffmpeg print error and fail before printing output format so the error is the last line printed on terminal, this should greatly help users who cannot find the error message in between. Example: Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '[Taka]_Naruto_Shippuuden_134_[480p][BCB62A7B].mp4': Duration: 00:23:24.77, start: 0.000000, bitrate: 1213 kb/s Stream #0.0(und): Video: h264, yuv420p, 848x480, 1081 kb/s, 119.88 tbr, 360k tbn, 48 tbc Stream #0.1(und): Audio: libfaad, 48000 Hz, 2 channels, s16, 128 kb/s Metadata major_brand : isom minor_version : 1 compatible_brands: isom [mxf @ 0x13e34a0]unsupported video frame rate Output #0, mxf, to 'test.mxf': Stream #0.0(und): Video: mpeg2video, yuv420p, 848x480, q=2-31, 200 kb/s, 90k tbn, 60 tbc Stream #0.1(und): Audio: pcm_s16le, 48000 Hz, 2 channels, s16, 1536 kb/s Stream mapping: Stream #0.0 -> #0.0 Stream #0.1 -> #0.1 Could not write header for output file #0 (incorrect codec parameters ?) With patch: Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '[Taka]_Naruto_Shippuuden_134_[480p][BCB62A7B].mp4': Duration: 00:23:24.77, start: 0.000000, bitrate: 1213 kb/s Stream #0.0(und): Video: h264, yuv420p, 848x480, 1081 kb/s, 119.88 tbr, 360k tbn, 48 tbc Stream #0.1(und): Audio: libfaad, 48000 Hz, 2 channels, s16, 128 kb/s Metadata major_brand : isom minor_version : 1 compatible_brands: isom [mxf @ 0x13e34a0]unsupported video frame rate Could not write header for output file #0 (incorrect codec parameters ?) -- Baptiste COUDURIER Key fingerprint 8D77134D20CC9220201FC5DB0AC9325C5C1ABAAA FFmpeg maintainer http://www.ffmpeg.org

3 10

[PATCH] Add a slice_dir parameter to avfilter_start_frame()
by stefano.sabatini-lala＠poste.it 04 Dec '09

04 Dec '09

Hi, this is useful for the scale, slicify and pad filters, which have otherwise to heuristically determine the slice direction, so duplicating code. Also I'm not sure about which values to use for the slice_dir, now I'm using 1/-1 but >0/<0 may be better, suggestions? This technically requires a major bump, but since the API is declared non-stable we can simply bump minor. Regards. -- FFmpeg = Foolish and Fostering Minimalistic Portentous Extensive Ghost

2 6

[PATCH] Remove PixFmtInfo.name
by stefano.sabatini-lala＠poste.it 04 Dec '09

04 Dec '09

Hi, as in subject. -- FFmpeg = Funny Foolish Mortal Pitiful Erudite Glue

2 6

[PATCH] more APE optimization
by lorenm＠u.washington.edu 04 Dec '09

04 Dec '09

After applying this, there is only one reference to scalarproduct_int16(), in acelp_pitch_delay.c. AFAICT, acelp_pitch_delay.c hasn't ever been compiled. Is it someone's work-in-progress, or can I delete scalarproduct_int16? --Loren Merritt

4 3

[PATCH 6/6] RTMP output: cosmetics
by piratfm＠gmail.com 04 Dec '09

04 Dec '09

2 2

[PATCH 5/6] RTMP output: implement rtmp-output
by piratfm＠gmail.com 04 Dec '09

04 Dec '09

Also for handshake with red5, 1st part of server response, received by client must be sent back.

2 6