Relation between ffvfw and xvid

[Dali Lama]

Well, while I was encoding my second pass in xvid, I stopped to try out the stats file in ffvfw. I know its wrong to use the stats file from my xvid encode to encode into libavcodec. But I wanted to see what happened. It encoded fine, just with a larger filesize. Maybe 50 meg oversized. I used b-frames and single coefficient elimination.

My point is, is this codec similar to xvid? It seems from the info in the docs that it has code from xvid as well as other projects. So, does this improve upon xvid or is it just different. Is it a MPEG-4 compliant stream, etc, etc. Why doesn't the developers from that project just work on xvid or vice versa?

So, basically, my questions are not really encoding orientated, rather I just want to know the relation this codec has to xvid.

Thanks to anyone who can clarify,

Dali

[-h]

There is a video/audio codec (actually it's much more than that) called "ffmpeg" based at http://ffmpeg.sf.net , and it started life as a means to encode audio and video on *nix platforms. Over time it integrated more and more video and audio formats (via the "libavcodec" library), and now it supports quite a bit of the MPEG-4 spec, almost as much as (and in some areas more than) XviD.

ffvfw is an interface to ffmpeg's libavcodec library which lets windows programs use ffmpeg's encoding/decoding capabilities, just as XviD's vfw code lets windows programs use the MPEG-4 functionality of XviD's "xvidcore" library.

As for why there are two different projects which do seemingly the same thing, ffmpeg has a *far* wider scope than XviD, which is only destined to be an MPEG-4 video codec. Check out ffmpeg's sourceforge site if you want to see what it's capable of.

And as for why there are people working on separate projects with seemingly the same goal, the answer is that people will work on whatever they feel like at the time. I never said it was the most efficient way to distribute a workload :)

-h