RetroSearch Browse

Tue Jan 30 02:53:47 CET 2007 · https://ffmpeg.org/pipermail/ffmpeg-devel/2007-January/028231.html

Hi,

* Michael Niedermayer <michaelni at gmx.at> [2007-01-30 01:34:19 +0100]:

> Hi
> 
> On Tue, Jan 23, 2007 at 11:06:53AM +0800, Limin Wang wrote:
> > Hi,
> > 
> > > this looks better but it still has problems
> > > first it overrides the demuxer timebase blindly, which breaks some
> > > h264 in avi/mov, also it breaks user specified timebases 
> > > (ffmpeg -r 12 myfile.h264 out.abc)
> > > 
> > > so the check should not be for CODEC_ID_H264 but rather for the lack of
> > > a set timebase (this needs some change to libavformat/raw.c more specifically
> > > the hardcoded 1/25 timebase there should be changed to 0/0) then utils.c
> > > somewhere should check for this and set it to the codec timebase
> > > st->codec->time_base with av_set_pts_info or if that also is 0/0 then 
> > > use some random default like 1/25
> > 
> > I have fixed by your suggestion, please review the attached patch. One problem
> > is it'll cause av_rescale_q() in utils.c floating point exception, I don't know
> > what's the background. After comment it out, I can pass the test without
> 
> av_rescale_q cant work with 0/0 timebase
> a if(timebase.den != 0) around that should fix it

OK, fix by this way instead of comment out. Please review regenerated patch
in attached file.

> also do the regresion tests pass with your patch (make test) ?
> and video grabing should also be tested (by someone) if all works fine then
> this patch (with the av_rescale_q if fix) should be ok

How to do the regresion tests? Any automatic test script? I have pass the test
with (h264->yuv, mov->yuv, > mp4->yuv). Who can do the video grabing testing?
I don't know how to test it.

Thanks,
Limin 
> [...]
> -- 
> Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB
> 
> Democracy is the form of government in which you can choose your dictator



> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at mplayerhq.hu
> http://lists.mplayerhq.hu/mailman/listinfo/ffmpeg-devel
-------------- next part --------------
Index: ffmpeg.c
===================================================================

--- ffmpeg.c	(revision 7763)
+++ ffmpeg.c	(working copy)
@@ -107,8 +107,8 @@
 static int frame_leftBand  = 0;
 static int frame_rightBand = 0;
 static int max_frames[4] = {INT_MAX, INT_MAX, INT_MAX, INT_MAX};
-static int frame_rate = 25;
-static int frame_rate_base = 1;
+static int frame_rate = 0;
+static int frame_rate_base = 0;
 static float video_qscale = 0;
 static int video_qdiff = 3;
 static uint16_t *intra_matrix = NULL;
Index: libavformat/utils.c
===================================================================
--- libavformat/utils.c	(revision 7763)
+++ libavformat/utils.c	(working copy)
@@ -1939,8 +1939,8 @@
              st->codec->codec_id == CODEC_ID_SHORTEN ||
              (st->codec->codec_id == CODEC_ID_MPEG4 && !st->need_parsing))*/)
             try_decode_frame(st, pkt->data, pkt->size);
-
-        if (av_rescale_q(st->codec_info_duration, st->time_base, AV_TIME_BASE_Q) >= ic->max_analyze_duration) {
+        if (st->time_base.den != 0 &&
+            av_rescale_q(st->codec_info_duration, st->time_base, AV_TIME_BASE_Q) >= ic->max_analyze_duration) {
             break;
         }
         count++;
@@ -1986,6 +1986,12 @@
                     st->r_frame_rate.den = st->time_base.num;
                 }
             }
+			
+            /* update with real frame rate if user didn't force it */
+            if( !st->time_base.num )
+                av_set_pts_info(st, 64, st->r_frame_rate.num, st->r_frame_rate.den);
+            else	
+                av_set_pts_info(st, 64, 1, 25);
         }
     }
 
Index: libavformat/raw.c
===================================================================
--- libavformat/raw.c	(revision 7763)
+++ libavformat/raw.c	(working copy)
@@ -304,11 +304,13 @@
     /* for mjpeg, specify frame rate */
     /* for mpeg4 specify it too (most mpeg4 streams dont have the fixed_vop_rate set ...)*/
     if (ap->time_base.num) {
+        /* user force fps */
         av_set_pts_info(st, 64, ap->time_base.num, ap->time_base.den);
     } else if ( st->codec->codec_id == CODEC_ID_MJPEG ||
                 st->codec->codec_id == CODEC_ID_MPEG4 ||
                 st->codec->codec_id == CODEC_ID_H264) {
-        av_set_pts_info(st, 64, 1, 25);
+        /* default is 0/0 */
+        av_set_pts_info(st, 64, 0, 0);
     }
 
     return 0;
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 481 bytes
Desc: not available
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20070130/9ff5064a/attachment.pgp>



Home - News ( United States | United Kingdom | Italy | Germany ) - Football scores

Showing content from https://ffmpeg.org/pipermail/ffmpeg-devel/2007-January/028231.html below:

[Ffmpeg-devel] [PATCH] a fix in ffmpeg.c