[FFmpeg-devel] [PATCH v3] QP tables API

[Michael Niedermayer]

On Sat, Apr 18, 2020 at 12:14:06PM +0200, Anton Khirnov wrote:
> Hi,
> a new iteration of the QP tables/video encoding parameters patches. The
> set now passes FATE and could be pushed if there are no further
> comments.
> 
> Changes from last iteration include:
> - type is now passed to the constructor, allowing for type-specific
>   per-block data to be added later
> - QP type is now int32_t
> - TYPE_H263 (equivalent to FF_QSCALE_TYPE_MPEG1) is gone, since it's
>   semantically equivalent to TYPE_MPEG2 divided by two. So we just
>   multiply by two on export and do not clutter the API with too many types.
> 
> 4/9 on its own breaks FATE (since the decoders start exporting the data
> in a different format), which is fixed by 5/9 and 6/9. We might want to
> merge those so that every commit passes FATE. Alternatively we could
> consider a compatibility layer, but given the obscurity of this feature
> that does not seem to be worth the effort.
> 

> Some filters (codecview, fspp, pp7, spp, uspp), and mjpegdec, are not
> converted yet. Since those have no tests, this does not break FATE.
> I may convert them later or someone else is welcome to do it.

hmm, i guess ill look into adding tests for these (if its easy and
doesnt have unexpected issues on any platforms), thats clearly
the first step so we will know if a convertion is non buggy

thx

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Concerning the gods, I have no means of knowing whether they exist or not
or of what sort they may be, because of the obscurity of the subject, and
the brevity of human life -- Protagoras
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 181 bytes
Desc: not available
URL: <https://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20200418/57417d46/attachment.sig>