[XviD-devel] Interview with Gej, mentioning XviD

Radek Czyz syskin at ihug.com.au
Mon Mar 22 13:24:21 CET 2004


Christoph Lampert wrote:

> I know you know this, but maybe somebody else is interested, too: 
> 
> In theory, distorsion D of R+lambda*D is already quadratic (it's 
> sum-of-squared quantization error). The rate R includes bits for 
> vector as well as bits for coding of quantized coefficients. 

As an electrical engineer (more or less), I prefer to see 
sum-of-squared-stuff as energy (just linear energy). But sure, there was 
one square involved.

> To speed this us, it is only approximated by  R + lambda*D with 
> R = vector bits and D=SAD (and different lambda), and there, the 
> SAD somehow compensates that the coefficient bits aren't taken into 
> account. Usually, SAD doesn't tell anything about PSNR, because it's 
> calculated before the reconstructed images (from quant/dequantized DCT 
> coefficient) is subtracted from current. 

Okay yes, but I was talking about VHQ mode decison and ME. I'm also 
thinking how to repeat this step for trellis quantization, but I'm not 
even trying to implement that one.

> If you find a combination that works better, hey, that would be great. The 
> problem might be how to "measure" is, because PSNR won't help there. 

That's right.... The minimum PSNR (across all frames) would rise, but it 
would rather be a side-effect - the idea wouldn't really work on whole 
frames but on blocks - and I haven't seen "minimum block PSNR in given 
frame" yet.


More information about the XviD-devel mailing list