TMPGEnc settings...there's too many, my head hurts.

[reboot]

Captured .avi is 829 meg (849, 896k).
GSpot reports it as 512x272, 23.97fps. Audio mpeg1-layer3.
I think this means that it's in PAL format.
I'm running it through TMPGenc now, CQ@100, and the output seems decent quality, however...
I set it to encode at 29.97fps, assuming tmpgenc will handle that for NTSC format to burn to SVCD.
Given the large aspect ratio (512x272) and i'm converting at 1:1, is this the best way to preserve the quality? Output is the same aspect ratio, which is what I want.
Would 2 pass VBR be better?
Would CB work? If so, what bitrate?
Should I reduce the output size, maybe by half?
Will that reduce the finished quality once burned to SVCD?
I know I have to cut it. Should I cut before re-encoding, after?
Is there something I should be doing with VirtualDub before hand? (I don't think so, as the quality of the original is excellent).
I'm just getting tired of waiting for TMPGEnc's 2+ hours of work, to end up with an SVCD that's about half quality of the original avi, when played on my standalone.
What would be the best way (TMPGEnc settings) to preserve MAX quality, regardless of size or other considerations. I don't care if a capture takes up 4 CDR's instead of two, as long as it's as clean as possible.

[falberni]

I use the NTSC DVD template. It sets every variable up for you.

[ZippyP.]

Try following a guide, here's one.

[ZippyP.]

To answer your questions.......

Given the large aspect ratio (512x272) and i'm converting at 1:1, is this the best way to preserve the quality?

That is not the aspect ratio, that is the resolution. SVCD is 480x480, anything else is not to spec and may not play on a DVD player. Use an SVCD template (like in the guide) and don't change the res.

Would 2 pass VBR be better?
Would CB work? If so, what bitrate?
Should I reduce the output size, maybe by half?
Will that reduce the finished quality once burned to SVCD?

CBR at the SVCD maximum bitrate of 2500 kbps is the best quality you can get, but you will only get about 40 minutes on a CD. Lowering the bitrate reduces quality but you get more on a disk. VBR (either 2-pass or CQ) will allow more minutes on a disk due to a lower average bitrate while keeping quality.

I know I have to cut it. Should I cut before re-encoding, after?

If you use the source range function in TMPGEnc then you can specify which part of the clip to encode, that way you don't have to cut it later. Cutting sometimes causes A/V sync problems.

Is there something I should be doing with VirtualDub before hand? (I don't think so, as the quality of the original is excellent).
I'm just getting tired of waiting for TMPGEnc's 2+ hours of work, to end up with an SVCD that's about half quality of the original avi, when played on my standalone.

For captured footage many people like to use filtering to reduce the noise in the video. Although you may not see it, it's there, and it eats up bitrate reducing quality. Vdub has filters (frameserve it to TMPGEnc) or you can use the built-in TMPGEnc filters. Fiiltering will add to your encoding time however.

What would be the best way (TMPGEnc settings) to preserve MAX quality, regardless of size or other considerations. I don't care if a capture takes up 4 CDR's instead of two, as long as it's as clean as possible.

CBR at the maximum of 2500 kbps with some filtering.

[reboot]

I knew I had the wrong terminology...sorry.
OK, resolution. I know that SVCD requires 480x480, and TMPGEnc does that nicely, adding the top and bottom black bits as needed to keep the original full width.
I have been using the default SVCD (NTSC) template, then tweaking the bitrate to CQ 100 (2500). So I guess I got that part right, however, you mention that I can get more on a disk with VBR (2 pass), or CQ-VBR.
I like the idea of CQ-VBR, so my next question about this part is, what should I set min and max quality and bitrate for this?
As for templates. I understand (correct me if wrong) that my video IS in PAL format, and I should be using the SVCD (NTSC Film) template, not just the SVCD NTSC template?

Maybe I've been lucky. Haven't had any audio sync problems, other than one, but that was my fault in my VirtualDub settings. I haven't figured out how to use the source range bit, nor frameserving, but will dig into the guides later.
I still need to learn how to do menu's and chapters as well Found a great link to a guide for that.

Thank you very much for your patient answers, and any more tips would be appreciated.

p.s. Thanks for the tips on filtering. I have successfully used VDub and TMPGEnc both to get much cleaner output video.

[bugster]

You say your source video is 23.976fps. That is NTSCfilm framerate. So use that template.

[reboot]

OK, got that, thanks.
Decided to mess with Mainconcept's MPEG Encoder.
Much faster than TMPGEnc, and once I found all the settings I needed to change for aspect ratio (16x9 original avi to 480x480 SVCD, with the black bars so it looks right), it's doing a great job.

Next question

I have an avi in which the audio is perfectly in sync for the first 5 minutes, then jumps WAY out of whack. Seems the audio is about a minute ahead of the video. This is a sudden change, not a gradual one. I read the guides on fixing out of sync audio, and they don't seem to apply, as the first 5 minutes is perfect.
Would VIrtualDub's "drop frames when behind" fix this, or is there another (better) way (if any)?

[ZippyP.]

I have an avi in which the audio is perfectly in sync for the first 5 minutes, then jumps WAY out of whack..... This is a sudden change, not a gradual one.

Probably bad frames, link below.

[reboot]

Thanks Zippy. Got VDub MP3, and will run the video thru it later today.
Of all the versions of vdub I have, I thought one or the other would "see" the bad frames if there were any. The only version I hadn't tried is this one.
I'll be back later with the results

[reboot]

Vdubmp3 doesn't find any bad frames, yet the original avi is out of sync not quite as I described above, but the audio is AHEAD by almost exactly 10 seconds. This happens at frame 7081. I hear the audio, and 10 seconds later the video appears. I managed to extract the full .wav and see if my thinking is correct here:
I separate video and audio.
I remove 10 seconds from the video, then put them back together in VirtualDub.
Things should be correct then, although I am going to have to play with it a bit to get the exact frames to delete from the video.
Would it be easier to add 10 seconds to the audio? It's mostly background music, and 10 seconds copied from another section of the file could be spliced in, with nobody the wiser I think.

[reboot]

Update:
With a LOT of fussing around, I managed to get the video in sync with audio, mostly by editing out chunks of video, until it fit properly.
That's one way.
Then I tried the other way. I edited in chunks of background music copied from another section of the .wav, and edited in the appropriate amount, in the appropriate places (a few k here, a few k there).
The finished product is better this way, and only noticable to me (I think).
Thanks again for all your help
Now, on to other questions