[Ffmpeg-devel] Odd Missing Video Results

[Myron L Stewart]

What does that mean do not top-post?

- Myron L Stewart


> From: M?ns Rullg?rd <mru at inprovide.com>
> Reply-To: FFMpeg development discussions and patches
> <ffmpeg-devel at mplayerhq.hu>
> Date: Wed, 07 Jun 2006 00:33:27 +0100
> To: FFMpeg development discussions and patches <ffmpeg-devel at mplayerhq.hu>
> Subject: Re: [Ffmpeg-devel] Odd Missing Video Results
> 
> Please don't top-post.
> 
> Myron L Stewart <mavmann at ingeniak.com> writes:
> 
>>> Myron Stewart <mavmann at ingeniak.com> writes:
>>> 
>>>> Hello,
>>>> 
>>>> I have been using the ffmpeg library for a project and for some
>>>> odd reason seeking past the the start on these 2 particular files
>>>> is causing the video to be missing. The file size seems to
>>>> indicated video was added, but when you attempt to play the video
>>>> in for example mediaplayer, quicktime, or vlc, there is only
>>>> audio. One thing I did notice is that both of these files have
>>>> swapped audio channels where the audio is on channel 0 and the
>>>> video is on channel 1.
>>>> 
>>>> I also did a test with ffmpeg.exe using the ss option to seek
>>>> ahead and get the same problem. For some reason the ffmpeg library
>>>> thinks that the first stream is a video stream instead of an audio
>>>> stream. Here are the 2 files:
>>> 
>>> The video elementary stream in those files has no repeated sequence
>>> headers.  Strictly according to the standard, this is allowed.  It is
>>> not generally a good idea though, as you have seen.  I'd suggest
>>> avoiding TMPGEnc if it creates such streams.
>> 
>> I did not create them. I got them from my client. The project I am
>> working on has to be able to work with just about any mpeg stream. I
> 
> This is the first time I've come across MPEG streams like that.
> Whatever program created those files must surely have some option to
> create sensible output.  If your client is not willing to investigate
> that, I'd say he's a waste of time.
> 
>> can see that the mpeg is poorly created, but it has been difficult
>> convincing my client on this, especially when avidemux can work with
> 
> What can avidemux do with the files?
> 
>> the files. I am not sure what they are doing. At least I am not the
>> only person who can see the way these are done are not correct.
>> Thanks for the info. I'll pass this along to them. I am still trying
>> to see if there is a way around this using ffmpeg.
> 
> The only way to decode streams like that is to read the sequence
> header at the start before doing anything else.  When that is done, it
> is possible to seek to any position (well, to any I-frame) in the
> stream and start decoding.
> 
> -- 
> M?ns Rullg?rd
> mru at inprovide.com
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at mplayerhq.hu
> http://lists.mplayerhq.hu/mailman/listinfo/ffmpeg-devel