[x264-devel] Re: Test results of trellis,brdo,and mixed_refs

[Tuukka Toivonen]

Loren Merrit scribbled once:

>What version of x264? svn-386 should somewhat improve brdo at low
>bitrates, and nofast_pskip might do more. (This doesn't fix whatever bug
>is causing brdo to increase bitrate, but it reduces the frequency at
>which it occurs.)

svn-385. :)

>What are the clips? I have some, but more testcases that break brdo would
>help. (The fluctuations in trellis on test1 are also interesting.)

Self captured from TV with BTTV. Quite noisy, hqdn3d-denoised.
Quite typical TV programs, some deinterlaced, no sports, no news =)

On Thu, 15 Dec 2005, Jeff Clagg wrote:

>chroma_qp_offset=8?! I don't think this is likely to be used in any real
>encodes.

Yes, probably true. It's quite funny though. No matter what I set
chroma_qp_offset, U & V PSNRs tend to be still higher than luma
PSNRs. Some visible color artifacts appear, though, and the bit-rate
reduction isn't that much compared to more sane chroma qp offset.

(I still routinely use chroma_qp_offset=5 or so)

Possibly because I smooth the chroma before encoding with -vf 
pp=lb:n,unsharp=c3x3:-1.2 or something.

Btw, it would be nice if there would be a special greyscale
encoding mode like with ffmpeg. I'll submit a patch in 2007
if nobody hasn't done it until then ;)

-- 
This is the x264-devel mailing-list
To unsubscribe, go to: http://developers.videolan.org/lists.html