On Thu, Jul 12, 2012 at 06:50:29PM +0200, Kostya Shishkov wrote:
> On Thu, Jul 12, 2012 at 06:11:50PM +0200, Diego Biurrun wrote:
> > On Wed, Jul 11, 2012 at 08:00:43AM +0200, Kostya Shishkov wrote:
> > > --- /dev/null
> > > +++ b/libavcodec/mss4.c
> > > @@ -0,0 +1,680 @@
> > > +static inline int get_coeff(GetBitContext *gb, VLC *vlc)
> > > +{
> > > + int val = get_vlc2(gb, vlc->table, vlc->bits, 2);
> > > +
> > > + return get_coeff_bits(gb, val);
> >
> > pointless variable indirection?
>
> return get_coeff_bits(gb, get_vlc2(gb, vlc->table, vlc->bits, 2));
>
> would look too LISPy to me and less clear
Are you saying I was spoiled through functional programming? :)
Stay indirect and inelegant if you must ;-p
> --- /dev/null
> +++ b/libavcodec/mss4.c
> @@ -0,0 +1,681 @@
> +
> + ff_mss34_dct_put(c->imgbuf[i], 8, c->block);
> + out = dst[i] + mb_x * 16;
> + // Since the DCT block is coded as YUV420 and the whole frame as
> YUV444
> + // we need to scale chroma.
YUV444,
> +/* Image mode - the hardest to comprehend MSS4 coding mode.
> + *
> + * In this mode all three 16x16 blocks are coded together with a method
> + * remotely similar to methods employed in MSS1-MSS3.
> + * The idea is that every component has a vector of 1-4 most common symbols
> + * and an escape mode for reading new value from bitstream. Decoding consists
from the bitstream
> + if (frame_type != SKIP_FRAME && !bytestream2_get_bytes_left(&bc)) {
> + av_log(avctx, AV_LOG_ERROR, "Not skip frame without data found\n");
This sounds confusing, maybe
Empty frame that is not a skipframe found.
is better.
Patch is fine from my side, no need to send another iteration.
Diego
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
