xvid - mpeg2, quality and bitrate

[Xenoide]

ok, I'm not quite sure if this question will make much sense to you pros, but I asked myself, if there is some kind of a 'crossover bitrate', where you could say: "from here on MPEG2 is recommended and below you should go with xvid".

Or could we say, xvid would be alwys superior at the same bitrate, even in the higher ranges (DVD, etc...)?

leaving all the player and compatibility aspects, and just talking about simple video files, what is your opinion about this?


thanks a lot for some ideas, even if it should sound totally stupid to you, I'm still learning... ;)

[Teegedeck]

I guess in theory there's nothing MPEG-2 can do that MPEG-4 could not do as well or better. Of course there are various implementations of MPEG-2 and 4 encoders that vary in quality a lot. Especially commercial MPEG-2 encoding apps have matured well, and hardware-players are ubiquitous. So if you have DVD-R as a medium you could as well use MPEG-2 for archiving your DV videos. It's pretty safe to say though, that MPEG-2 cannot compete as far as quality/filesize efficiency is concerned.

At DVD-bitrates it would be interesting to compare XviD (using high-bitrate matrices like Andreas' 78er) to a commercial MPEG-2 software like CCE.

[Xenoide]

well, I had this idea as I saw the latest codec shootout on the main site. The original pictures from the DVD also showed encoder flaws in certain situations and I was wondering if xvid could do better there.

However I forgot one little but crushing point: where to get better originals ;)

DVDs won't get better if you re-encode them with xvid, no matter how high the bitrate gets, with DV-cams it's the same, "upcoding" won't help.

So I guess these thoughts will stay theory until someone invents a good software, that will dither and interpolate video material to restore lost content and data.

When it's up to downsize videos and size/quality is more important than compatibility to standalone players, MPEG4 would be preferred I think, otherwise there is not much sense to switch from the original codec, until the standards change.


But maybe I'll do a test somewhen, just for personal interest with a rendered scene stored without compression including many "codec-killer" elements. Then it could be possible to judge by the quality of MPEG2 and xvid at high bitrates...

[Teegedeck]

I'm sorry but we cannot have another one of those 'this is better than that' discussions without proof here. While it is pretty safe to say that in theory MPEG-4 has enough merits to do as well a MPEG-2 at DVD-bitrates potentially, it is not at all safe to negate that chance.

Let me explain; now that you said that a higher DC precision makes MPEG-2 superior to MPEG-4, someone else could say that quarterpel motion-search precision makes MPEG-2 superior to MPEG-4. And so on and so on. If person A argues that his car has better wheels and person B argues that his car is more streamlined, we don't come any closer to learning whose car is faster, if you know what I mean.

Fact is that we can't foretell how decisive a single factor might be for picture quality when all other pros and cons come in - someone would have to conduct a test, there's no getting around it. In the end we'd have to see how well implementations of the standards would compete.