H.264 vs. MPEG-4 retains detail better?
 

To my eyes, MPEG-2 seems to retain detail better than H.264 when copying SD DVDs even at equivalent bit rates. Since MPEG-4 is a precursor to H.264, will it retail greater detail, albeit at a higher bit rate?

Correct.

H.264 has an internal softening filter that is on by default. This is to help it with lower artifacts, from lower bandwidth/bitrate, but sometimes the outcomes are what you have observed. So the idea that H.264 is "better" is a myth -- it's just different. It's the main reason we only use MPEG-2 for our VHS conversion projects to Blu-ray. The minor artifacts are better than the wholesale softening that can occur.

I think the filter can by disabled, but there's still other nuisances in H.264 that do not exist in MPEG-2. That makes a KISS workflow needlessly complex.

To end users, MPEG-4 is mostly just a container. (Remember that MPEG-2 is a container AND a format/codec.) But MPEG-4 is also sort of like a toolkit for creating codecs or formats. Unlike MPEG-2, MPEG-4 is not format itself.

It's complicated.

MPEG-4:
- MP4 as AVC/H.264 = MPEG-4 part 10
- AVI as Divx/Xvid = MPEG-4 part 2

I meant MPEG-4 part 2 (think Divx), commonly referred to as MPEG-4

Divx (or open source Xvid) just isn't as good as MPEG-2. It valued size over quality. It was replaced by H.264 by almost everybody, from the hobby users to the equipment manufacturers.

So again, what you see is not surprising.

so to choose the better of two options, will Divx retain detail more like MPEG-2 or is it worse than H.264 in that sense?

Xvid/Divx will retain more noise, as it's really a mix of MPEG-2 and H.264. It's an earlier era of video compression from 10+ years ago.
H.264 will blur if not given enough bitrate.

What do you mean by "copying" DVDs? You're compressing a retail movie down to a smaller size using both MPEG-2 and H.264?

MPEG-4 Part 2 is a different compression algorithm from MPEG-4 Part 10 (H.264). The H.264 standard is newer, having come out in 2003 vs. 1999. Really the only thing they share is the name "MPEG-4", they are completely different video compression formats. Part 2 is the original video codec developed for the standard, while Part 10/AVC was jointly developed by ITU and MPEG and later added to the MPEG-4 family of codecs. If you have a fast machine, encode with H.264. Support is universal now, and the codecs are mature. Much better compression can be achieved over the old Part 2 codec while retaining the same quality.

Also to clear things up, MPEG-4 Part 2 video can be in a MP4 container (defined in MPEG-4 Part 14). It wasn't really intended for use in AVI containers. That was popularized by DivX, which was originally a hacked codec to allow Microsoft's old MPEG-4 encoder to store video in containers besides the proprietary ASF format (AVI doesn't officially support features required by the MPEG-4 codec). MP4 didn't come until 2003. Before that, MPEG-4 Part 2 video was mostly distributed in Quicktime MOV and Microsoft ASF containers.

http://www.digitalfaq.com/forum/atta...5&d=1410792538

I see this very evidently with H.264 Blu-Rays, and the codec does an awesome job with HD material. To my eyes, perhaps due to converting the already compressed MPEG-2 stream, the psychovisiual settings seem to noticeably soften image detail and result in a bit of banding at times and blur edges. I have played with H.264 settings in Handbrake and TMPG and haven't found the sweet spot yet.

No question its a great codec, but for SD, my eyes tell me MPEG-2 has better PQ at high bit rates than H.264 with identical bit rates. Maybe my settings are off, but since part 2 is closer to MPEG-2, I was thinking PQ might be better, but using the MPEG-4 part 2 encoder in Handbrake, I was not satisfied at all. I am probably sticking to x264 for now.