[MPlayer-dev-eng] Proposed patch for ASF lock up due to consecutive video packets with no valid PTS.

[Robert Krakora]

Hello:

I improved the patch (see below)...I don't recall getting any responses on
this yet...

[root at am2mm mplayer]# svn diff
Index: libmpdemux/video.c
===================================================================
--- libmpdemux/video.c  (revision 29752)
+++ libmpdemux/video.c  (working copy)
@@ -426,10 +426,11 @@
     fprintf(stderr, "'\n");
 }

+double last_frame_time = -1.0;
 int video_read_frame(sh_video_t* sh_video,float* frame_time_ptr,unsigned
char** start,int force_fps){
     demux_stream_t *d_video=sh_video->ds;
     demuxer_t *demuxer=d_video->demuxer;
-    float frame_time=1;
+    float frame_time=0.033;
     float pts1=d_video->pts;
     float pts=0;
     int picture_coding_type=0;
@@ -591,7 +592,13 @@
       case DEMUXER_TYPE_ASF: {
         double next_pts = ds_get_next_pts(d_video);
         double d= (next_pts != MP_NOPTS_VALUE) ? next_pts - d_video->pts :
d_video->pts-pts1;
-        if(d>=0){
+       if(d>=0){
+          if (d >= 1.0 && last_frame_time > 0.0)
+                d = last_frame_time;
+          else if (d >= 1.0)
+                d = frame_time;
+          else
+                last_frame_time = d;
           if(d>0){
             if((int)sh_video->fps==1000)
               mp_msg(MSGT_CPLAYER,MSGL_V,"\navg. framerate: %d
fps             \n",(int)(1.0f/d));
@@ -601,7 +608,8 @@
           frame_time = d;
         } else {
           mp_msg(MSGT_CPLAYER,MSGL_WARN,"\nInvalid frame duration value
(%5.3f/%5.3f => %5.3f). Defaulting to %5.3f
sec.\n",d_video->pts,next_pts,d,frame_time);
-          // frame_time = 1/25.0;
+          if (frame_time > 1.0 && last_frame_time > 0.0)
+                frame_time = last_frame_time;
         }
       }
       break;

Best Regards,

-- 
Rob Krakora
Senior Software Engineer
MessageNet Systems
101 East Carmel Dr. Suite 105
Carmel, IN 46032
(317)566-1677 Ext. 206
(317)663-0808 Fax

On Sat, Oct 3, 2009 at 12:47 PM, Robert Krakora <
rob.krakora at messagenetsystems.com> wrote:

> Hello:
>
> I am using a Linksys PTZ camera that employs ASF to encapsulate MPEG4
> video.  I was seeing the main mplayer process lock up decoding video while
> the mplayer forked process was still acquiring HTTP delivered ASF data from
> the network connection.  Upon further review I noticed that two back-to-back
> ASF packets contained no PTSs.  On the later invalid PTS a gigantic frame
> time is calculated and fed to a sleep call from which it never returns (it
> may in a week or two).  The patch below attempts to address this (works for
> me) by crudely extrapolating the frame time based upon the last frame time
> in this instance.  Please review and comment.  ;-)
>
> [root at localhost mplayer]# svn diff
> Index: libmpdemux/video.c
> ===================================================================
> --- libmpdemux/video.c  (revision 29748)
> +++ libmpdemux/video.c  (working copy)
> @@ -426,6 +426,7 @@
>      fprintf(stderr, "'\n");
>  }
>
> +double last_frame_time = -1.0;
>  int video_read_frame(sh_video_t* sh_video,float* frame_time_ptr,unsigned
> char** start,int force_fps){
>      demux_stream_t *d_video=sh_video->ds;
>      demuxer_t *demuxer=d_video->demuxer;
> @@ -592,6 +593,12 @@
>          double next_pts = ds_get_next_pts(d_video);
>          double d= (next_pts != MP_NOPTS_VALUE) ? next_pts - d_video->pts :
> d_video->pts-pts1;
>          if(d>=0){
> +          if (d > 1.0 && last_frame_time > 0.0)
> +                d = last_frame_time;
> +          else if (d > 1.0)
> +                d = frame_time;
> +          else
> +                last_frame_time = d;
>
> Best Regards,
>
> --
> Rob Krakora
> Senior Software Engineer
> MessageNet Systems
> 101 East Carmel Dr. Suite 105
> Carmel, IN 46032
> (317)566-1677 Ext. 206
> (317)663-0808 Fax
>