[FFmpeg-devel] [RFC] Implement support for interplay MVE 0x06, 0x0F and 0x10

Paul B Mahol onemda at gmail.com
Sat Jun 17 19:45:46 EEST 2017


On 6/17/17, Hein-Pieter van Braam <hp at tmm.cx> wrote:
> Hi all,
>
> I seemed to have done something wrong when fixing the patchcheck
> problems originally. Apparently I didn't build between making my
> 'fixes' and creating the patch. The previous patch doesn't compile.
>
> I fixed this in this version, my apologies for the noise.
>
> I've also fixed reading garbage data from the 'last frame' on the very
> first frame.

Have you made sure that new code does not cause overreads?

Use bytestream2* for that.

>
> - HP
>
> On Sat, 2017-06-17 at 15:08 +0200, Hein-Pieter van Braam wrote:
>> Hi all,
>>
>> This patch implements support for 3 previously unknown Interplay MVE
>> opcodes. These opcodes together implement support for 2 additional
>> video frame formats.
>>
>> This is my first time trying to contribute to ffmpeg so I expect this
>> code to not be entirely up to snuff, I'm interested in getting this
>> merged so any comments are welcome and I'll make any necessary
>> changes.
>> I have ran patchcheck over it and fix most of the issues it found.
>>
>> I'm working with Multimedia Mike to get these opcodes documented on
>> the
>> multimedia.cx wiki.
>>
>> You may notice something strange going on with decoding opcode 10
>> movies, the reason for that is that the codec appears to need access
>> to
>> a block since before the last time it was changed, this is not
>> necessarily the last displayed frame. To implement this I decode to
>> two
>> extra AVFrame's, swapping them after each decode, and copying only
>> changed blocks from the current decoding frame to the final display
>> frame. I think that my implementation is probably more convoluted
>> than
>> it needs to be, any suggestions on that front would be most welcome
>> also.
>>
>> There's a bug in FFmpeg master currently that prevents the MVE
>> decoder
>> from signaling that the end of the file has been reached. I've made
>> no
>> attempt to fix that in this patch. I'll create a separate patch to
>> fix
>> this issue.
>>
>> - HP
>> _______________________________________________
>> ffmpeg-devel mailing list
>> ffmpeg-devel at ffmpeg.org
>> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel


More information about the ffmpeg-devel mailing list