DivX vs XviD 1 Pass CQ

[Jacquers]

If possible to do this without flamewars, has there been a comparison of DivX vs XviD in 1 pass CQ mode?

I have noticed that as far as filesize goes DivX Q4 = XviD Q3 (roughly). Anyone have comments on the quality? Which DivX Q factor matches up to which XviD factor?

[GodofaGap]

Without giving more settings your question does not make any kind of sense, and it doesn't make a lot to begin with. (A common misunderstanding is that with CQ most people still use offset and ratio for b-frames, while others are of the opinion that this is not CQ anymore).

The biggest problem is, though, that you are asking for anecdotal opinions which do not have any real value. So try for yourself.

[Jacquers]

Ok, basically I want to compare DivX & XviD 1-pass quality based mode. Seems from my own tests that the same value for the quantizer doesnt match up (at least not as far as file size goes): DivX Q 4 isnt equal to XviD Q 4 for the same video source.

I was wondering if maybe someone has done a similar comparison so that I could match the two codec's quatizer settings?

[dimzon]

http://forum.doom9.org/showthread.php?t=112827

[weaver4]

The author of AutoGK (who I respect very much) seems to think that the quality 75% in XviD is equal to the quality of 60% in DivX. Now if I remember correctly that would be XviD quantizer of 2.5 and DivX of 3.5.

In my limited experimentation I have found that the quality is very close for the following:

XviD = 3, DivX = 4
XviD = 2.5, DivX = 3.5

Now before I get blasted with "but what setups are you using" I will say that this testing was done using the default parameter for AutoGk, avi.net and Staxrip; except I have the ess option on for AutoGk to force the h.264 filter. I always encode to a width of 640 and I play them on a 34inch XBR TV.

[Sharktooth]

To force what? ESS option forces h.263 quantization.
Also please note AutoGK and all automated softwares cant be used to compare codecs coz they choose completely different settings depending on some factors for each codec.

[weaver4]

Sharktooth you are right, it is h.263 not h.264.

You can only compare them if you assume that the author made the same trade-offs for speed, quality and filesize. I think that AutoGK, Staxrip and avi.net did a pretty good job of this.

Every time a discussion starts on comparison of Xvid or Divx the question quickly goes to what settings you should use, that conversation goes on for 19,345 messages. By then the poor chap (generally a newbie) is completely lost and he never gets an answer. For 99.5% of these people they will never make better setting than AutoGk, Staxrip or Avi.net anyway. So, I try to answer the question based on the general tools that they will probably use rather than answer it at the Codec level.

[Jacquers]

Thanx weaver4, post #5 was the type of answer I was looking for :)

[DigitAl56K]

Please note that so long as you have a profile other than 'Unconstrained' selected you are in fact using target, not constant quantizer mode in DivX. This mode attempts to maintain the quantizer you specify, but may increase it temporarily if necessary to ensure smooth playback on DivX Certified devices. XVID does not do the same in 1-pass mode as far as I'm aware.