What are the quality vs time ramifications of the VBR and CBR settings in TMPGEnc when encoding an AVI to MPG?

In other words, which gives the best quality, which gives the fastest conversion, and which is the best compromise?

VBR will give better quality, while CBR will be faster. VBR will analyze scenes and distribute bits to where they're needed most, while CBR keeps the bitrate constant, so a scene that's not as hard to encode will be given the same amount of bits as a scene that's very difficult to encode.

i haven't used TMPGEnc before, so i don't know exactly what the time difference would be between the two methods, but i'm guessing the quality wouldn't be the greatest with CBR unless u use a bitrate that's high enough.

There's a number of VBR settings, it's running now so I can't check. but they are something like "manual VBR", "auto VBR", etc.
There's also a 2-pass VBR. You can set average and max bitrates too. And even a target file size setting.

for CBR, max bitrate = min bitrate = avg bitrate. targeting a file size is really only useful for VBR, as the encoder will try to distribute the bits where it sees fit while staying under the file size limit. if TMPGEnc has the option to target a file size while using CBR, it would simply be choosing the constant bitrate that corresponds to the target file size divided by the time of the video. the avg bitrate would be the same in both cases, but the CBR vs VBR ramifications would remain the same as i described above.

There's also a "constant quality" VBR setting. You can set the target quality with a slider (in % of quality). The rather sparce help files suggest that this may be the best choice for quality if file size is not a concern.

it's probably better that their help files are sparce. i've seen their attempted english translations on their website, and it's not pretty. it's like trying to read a bad babelfish translation.

Your best quality will ALWAYS be with 2 pass VBR. But the Motion Search Precision will greatly determine the overall quality and speed. It needs to be on Highest Quality (Very Slow) to get the best out of TMPGEnc.

Just for fun, I set everything at max quality and did "constant quality" rendering of a 90-min AVI. (max bitrate-8000, min bitrate-4000) It took about 30 hours on a P-4 2.8 GHz system. With PCM audio, the DVD files are about 5.8 GB.
Quality is excellent.

A 90 min AVI (divx or xvid) takes me about 1 3/4 hr to 2 1/4hrs on my AMD 2600+ to convert using DVD2SVCD with CCE and thats in VBR mode Multipass 1 (equal to TMPGEncs 2 pass VBR mode) Final size is always just under 4.37gb and always as good as the origianl avi.

Yes, well DVD2SVCD with CCE is not option here on either my AMD or P4 systems. Nothing but crashes on both machines. It's really not worth the trouble.

Use a bitrate calculator to use all the disc space on your DVD and still fit on one disc. Or you can make a rough calculation with a normal calculator.

90 minutes is 90*60 seconds = 5400 s
A DVD-R can hold 4 700 000 000 bytes.
4700 000 000 / 5400 = 870 370 bytes/s = 6963 kbit/s

But this is including video, audio, subtitles and muxing overhead.
Let's say you use 256 kbit/s compressed audio (ac3 or mp2) and lets use 5 % muxing overhead.

0,95*6963 - 256 = 6359 kbit/s.

I guess 6300 kbit/s would be fine. Use that as average video bitrate. If you still get slightly oversized you can always use DVDShrink to squeeze it into a DVD.

You can aso use this bitrate calculator:
http://www.dvdrhelp.com/calc.htm