[FFmpeg-devel] NC4600 camera code

[nicolas martin]

> nicolas martin wrote:
>
>> Hi there,
>>
>> It's been a long time, but I have been working on my demuxer for my
>> cameras.
>>
>> I put all my code in a separate file.
>> However I have  a problem.
>> I've seen that you can have a private structure in the priv_data  
>> field
>> of an AVFormatContext.
>> So I put all my private fields here.
>>
>> But I have to modify the standard get_partial_buffer
>
> This is a sign that you are doing something wrong.
> You should explain why you need a modified version of  
> get_partial_buffer
> and then someone may explain you how avoiding it.
>
>> So I made one myself that takes a pointer on my structure as an
>> argument.
>> But If I want to use standard functions like fill_buffer defined in
>> aviobuf.c, which file should I included ???
>
> fill_buffer is static in aviobuf.c so you can't use it outside
> aviobuf.c. If you really need to implement modified version
> of get_partial_buffer, you have to do it inside aviobuf.c,
> but I'm pretty sure you don't need to do this.

in get_partial_buffer, when you read a new chunk of data, at most  
[size] bytes are copied to the destination buffer.
but the video feed contains header that tells the size of the next  
frame, so it's not at most [size] bytes that should be copied, but the  
difference between how much is to be read until next frame and how  
much has already been copied.

if I let get_partial_buffer copy [size] bytes, then it means, I  
sometimes will have to concatenate a chunk of that buffer, to another  
piece returned by get_partial_buffer. It's twice as much memory that  
should be reserved...

maybe it's the way to go...


>
>
> Aurel
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at mplayerhq.hu
> https://lists.mplayerhq.hu/mailman/listinfo/ffmpeg-devel