480p/24 Mpeg-2 frame rate issues

[Bexley]

I wasn't sure which forum this properly belonged in since it touches on several topics, so mods please move if it belongs somewhere else.

I have a lot of old homemade DVDs that are taking up a lot of shelf space, and I'm trying to consolidate them down by transferring the material to BD-R. Ideally, I do not want the quality loss of re-encoding with x264 because this is all rare or unique material (like home movies and shows that are not commercially available) that is difficult to replace, and in many cases the original source (mostly VHS and DV) is long gone. I want to use the Mpeg-2/Ac3 streams directly from the DVDs if possible.

MultiAVCHD is my tool of choice, but I've also had this issue with Adobe CS5.5 as well. The 480i material like home movies shot on VHS does just fine, but my problem is with the 480p material. When I import the streams into MultiAVCHD and author, the video plays back very jerky, almost like the 3:2 pulldown isn't being recognized by the player. If it were interlaced, I would think the field order was wrong. It's that kind of motion.

I did some homework and it appears that BD doesn't really support 480p and requires a fake interlace flag. As a test, I re-encoded some material using these (http://www.x264bluray.com/home/480p-ntsc) settings and they play perfectly. So my question is why the 3:2 pulldown flag isn't being recognized in my Mpeg-2 streams, but it is in my x264 streams with fake interlace? And can I fix it without having to re-encode all of my material?

Also, I have a small amount of 480p material that I have already encoded a while back as raw x264 stream without the fake interlace flag, and I no longer have the original source. I stupidly threw it away before I realized there was a problem - I simply assumed that MultiAVCHD would handle it properly during the encode and moved on. Can this material be fixed without having to re-encode?

TIA for any help.

[TheSkiller]

You should be able to reuse the original MPEG2 video and AC3 audio from your DVDs on Blu-ray. :)


To be DVD compliant a video that uses soft-pulldown (RFF flags) must be flagged as interlaced, even though the video is progressive (as a side note, the video absolutely needs to be progressive for pulldown to work). I suppose it is no different for Blu-ray.
It can be encoded using ZigZag scan though, so there's no additional encoding quality loss, it's just the flags.
You can check and change the MPEG2 elementary stream (m2v or mpv) flags using Restream (http://www.videohelp.com/tools/Restream).

So yeah, there's a good chance the problems you are experiencing come from the use of pulldown on a video flagged as progressive. Maybe you were just lucky you haven't yet come across a DVD-player that feels bothered by this.

[Bexley]

So if I'm understanding you correctly, my problem is caused by the flags used by my Mpeg-2 encoder? Almost all of the streams I'm working with were encoded with TmpegEnc Plus 2.5 with the 3:2 pulldown option. I've never had a problem using any of these streams in any DVD player in the last 10 years. Is it possible that BD players are simply different than DVD players in this regard?

If that's the case, which particular flags should I be looking for in Restream? When I open an m2v file in Restream, I see what looks to me like conflicting information - under Picture Coding Extension, the "Frametype Progressive" box is checked, but the "Progressive Sequence" box in the next section is not. If I uncheck the "Frametype Progressive" box and save this will theoretically fix my problem?

[Guest]

Can you please post an unprocessed source sample? Then we don't have to play guessing games.

[Bexley]

Restream seems to have solved my problem with the Mpeg-2 streams. I unchecked the progressive setting and ran a test, and it looks great. After all these years, I never had a clue that my streams were flagged incorrectly. It appears to bother BD players, but not DVD players. This is 99% of my issues fixed. Thanks!

So next question - is there a utility that will do the same with a raw x264 stream header? I have a handful that I encoded with MultiAVCHD directly from source to x264, and it apparently did not add the fake interlace flag like it should have. Can these headers be fixed the same way?

[Bexley]

Upon further fiddling around I think there is actually a combination of things going on here, including player compatibility. It appears that fixing the Mpeg-2 header is not really the solution after all, but in the course of looking further it led me to what the real problem is. Even when I fix the streams, I have a player that has problems with the streams and another that does not (different brands). So I started looking at the auto-detected properties of the stream in MultiAVCHD to see if it's picking up the flag and I noticed that all the streams were still listed as 23.976 progressive. I manually changed the stream properties to 29.970 and interlaced TFF and burned a test disc, and now it works on both players. So it seems the issue was with MultiAVCHD all along.

However, this trick did not work with my x264 streams that I am still trying to fix. Those get even worse if I try to force 29.970 on them. So still looking for a solution there.

[Guest]

However, this trick did not work with my x264 streams that I am still trying to fix. Those get even worse if I try to force 29.970 on them. So still looking for a solution there. Your chances of getting useful help will be greatly increased if you post a link to an unprocessed source sample that we can use to reproduce your issue. Is your issue with these x264 streams also that MultiAVCHD does not detect the stream parameters correctly?