I'm still on DVDA 1.0c, but I have found that setting the "DC Coeficient" (Main Concept MPEG2 Advanced Video properties) to 10 instead of the default 9, helps considrably with motion artifacts. I think that setting defines the precision that MC uses to calculate motion rendering. Might work for you.
Hope this helps.
Ken

I'm still on DVDA 1.0c, but I have found that setting the "DC Coeficient" (Main Concept MPEG2 Advanced Video properties) to 10 instead of the default 9, helps considrably with motion artifacts. I think that setting defines the precision that MC uses to calculate motion rendering. Might work for you.
Hope this helps.
Ken

John, thanks. I should have clarified. I'm seing the same stutter on several DVD players (set top) in addition to the mpg file and the DVD being played back on the computer drive, although they are not as noticable (progressive vs interlaced?)

The photo pans were done both in Vegas and in the Canopus product Imaginate.

I thought at first it might be a field order type of thing, but normally that shows up even in slower motion stuff.

I'm just curious if a 2 pass VBR is better in general for motion than CBR, even beyond the stutter question.

And Ken, thanks for your suggestion as well. I'll try that too.

Tom

2-pass should help.

VBR should help. Set the Maximum a lot higher than your average and those extra bits should only kick in on your fast action.

I'm seeing the same stutter on several DVD players (set top) ...

Thanks. That clarifies that.

I still think there is something other than CBR vs. VBR or, for that matter, anything else to do with bitrate.

Low bitrate should not create any visual artifact that looks like "stutter" (although see my comment below).

If you have ever done a really low bitrate encode -- for SVCD, or for DVD at 2,000,000 average or lower, you will see what happens when there aren't "enough" bits for a good encode. As the bitrate gets lower, you start to see "mosquito noise" around high contrast, fast moving objects. You also start to see noise and blockiness, especially in large smooth colored scenes, like a bright blue sky. Any scene with smoke will get filled with really ugly square blocks.

However, the video should not "stutter," and the motion should still be smooth.

If you can post a short segment that shows the problem, I'd be happy to take a look and see if I can come up with an idea of what is causing the problem. I'd try to guess now, but the problem is that what you describe as "stutter," might be totally different than what I am thinking stutter means. To me, that is something I usually associate with having field order reversed, or possibly from dropping frames. Field order reversal is by far the most likely culprit. You need to make sure that encode the MPEG-2 file using the same field order as your source material. NTSC DV is lower field first. HDV is upper field first.

I never use VBR, only a CBR of 8,000,000 and the results are about as good as it gets with the Vegas Main Concept encoder. VBR is only for one purpose - to save space. Instead I put a limit of around an hour on a disc. More than that and I use two discs. Some here insist they have to have everything on one disc so they go the Russian roulette dual layer route. That usually costs more than two discs and is not as compatible. A better encoder is my next move. I ordered Procoder 3.

John

I have the same problem with my wedding videos. Mostly at the sections where everything is very fast pace, zooming, panning..the clip is normaly about 4 minutes but the entire project is about 1h45minutes.

Using the CBR didn't help in that 4 minute clip...my next step is to try using VBR set the maxium to 8,000,000 and the Average to about 5,500,000 with Two-pass on.

If that doesn't work I'll try the DC coefficient to 10bit.

Using the CBR didn't help in that 4 minute clip...my next step is to try using VBR set the maxium to 8,000,000 and the Average to about 5,500,000 with Two-pass on.

If that doesn't work I'll try the DC coefficient to 10bit.
==========
Yeah I forgot to mention that I use 10-bit all the time.

However, I don't understand what you will gain from setting the maximum VBR to 8,000,000 over a CBR of 8,000,000 other than saving space. That is all VBR gets you - more space.

John

Well, it's not low bitrate, I was using 7000 and 8000 CBR on the encodes, and the DVD today was only 15 minutes long.

As I mentioned, I thought of field order and dropped frames, but the appearance of the video is not what I've seen before in field order problems. I'm going to do a little more experimenting this weekend. I'm beginning to have a flicker of suspicion (no pun intended) as to what might be the cause. It might be something really unrelated, like a hard drive or other hardware problem, but the DV files look fine.

John, a folowup on the UF/LF thing. Recording in DV is lower field, true, but I've also read that it's best to encode mpg in UFF for DVD playback. I believe I've tried flipping the fields around before without any problems, but I wil doublecheck my settings.

The only thing remotely different is that a lot of the DV files are ones edited in Canopus EDIUS and composited in Vegas. I wonder if this might be a cause of the wierdness - mixing the Canopus "wrapper" with the Vegas "wrapper". I also note that some of my mpg files were rendered to mpg directly from the Edius timeline using their generic or Procoder mpg encoder. I can't recall if they showed the same problem or not. The very quick panning montages I only do occasionally for a hard hitting video, so I'm not sure at this point.

I'll play and make some notes this weekend +/- family obligations. This only came up again this morning as I had a deadline gettting a DVD out for a presentation this evening and I noticed it again. But, I do have time to fix the issue before any of them get distributed.

Thanks for all of the responses. I'll let you know if I come up with anything.

Tom

I'd vote for field order issues. Do a test burn with the field order reversed, and see how it compares.

The correct field order should look flawless. The wrong field order will be really awful on fast pans, a kind of "shudder."

The only time I've seen a DVD player stumble was when I set my min to 0 on an encode. For a fraction of a second following the black of a title, one of my players stumbled, I'm guessing because it took a while to spin up the disk to the higher bitrate. But I don't think that's what's happening here.

As for VBR bitrate, I've successfully gotten 2:20 on a single-layer disk by being very careful and by using CinemaCraft, which is an excellent encoder and which also allows one to inspect GOPs and manually tweak bitrate for segments, if necessary. For a project last year involving 3 hours of material I did my first double-layer burn. The result was perfect once I figured out the brand of disk to buy (Verbatim +R DL) and the burning software (CopyToDVD).

That is all VBR gets you - more space ...

Just to avoid confusion, VBR vs. CBR will not affect space. The ONLY thing that affects the size of the MPEG-2 file is the average bitrate.

If you set it to, for example, 6,000,000 bps (average), you will get the same file size whether you encode in VBR or CBR. What's more -- and this is the one that surprises most people at first -- you will get the same size whether you encode at 720x480, 320x240 or any other resolution.

The key in understanding this is to look at the name of the setting: Bits Per Second. What this says is that the encoder is going to use that many bits for every single second of video. When encoding a higher-resolution video (e.g., 720x480 instead of 320x240), fewer bits are available to create the intermediate "B" and "P" frames, so you get more motion artifacts (e.g., "mosquito noise"), but the video is sharper. It is an interesting tradeoff, and an especially important one to understand when encoding for the web.

It also doesn't matter what frame rate you use. If you encode the video at 10, 15, 29.97, or 30 fps, the size of the file will still be determined ONLY by the average bit rate setting.

Since I read a lot of posts over at doom9.org and similar forums, I know that many people claim that they get smaller files when they reduce noise. It is a common claim. What I figured out is that a lot of people use encoders that have a "constant quality" setting (TMPGEnc is one of these). As you reduce spurious noise in a video signal, a smart encoder can get the same "quality" by using fewer bits and therefore you will get a smaller file size. However, the MainConcept has no such setting.

Finally, as far as the "DC Coefficient" is concerned, I have tried for years to get a good explanation of what this does and when you want to use it. I often set it to 10 bits, but I don't think I have ever been able to detect a difference. FWIW, here is the explanation from the help file for the MainConcept external MPEG-2 encoder:

"Specifies the effective precision of the DC coefficients in intra-coded macroblocks. 10-bits usually achieves quality saturation. 11-bits can be used if the quantization is very low (the bitrate is quite high compared to the frame size/rate). See ISO/IEC 13818-2 section 6.3.10 for more information). This option is only valid for MPEG-2"

Does that help. No, I guess it doesn't ...

[Edit]

P.S. I just did a short encode using Vegas 7.0d from a NTSC DV file to MPEG-2, using CBR and VBR. Guess what? The file sizes were not the same!! Am I going to re-write everything above? No. Why not? Because the fault lies in Vegas, not in my understanding. To prove this, I re-did the VBR using 2-pass VBR and guess what (I like saying "guess what")? This time, the file sizes match. Now, I only encoded about five seconds of video. My guess is that the one-pass encode would have produced more accurate size results if I did a longer stretch of video where the local optimization it does (as opposed to the global optimization done during 2-pass) would have averaged out more accurately.

I don't have any problems with motion and I use VBR all the time. I tested this by doing a whole bunch of camera whips back and forth just like kids do on horizontal sidings of a house, cars in motion, people in motion with my Z1 in full auto. Absolutely no stuttering, juddering, pixelation. I use VBR for the standard discs (Rendered in NTSC Widescreen) and CBR for the HDV renders because it's the default and it always worked fine.
JJK

I looked up DC coefficient in the CinemaCraft manual, and all it says is:

Intrablock DC precision

"...Specifies the bit precision of the DC coefficient of intra-blocks. 10 is recommended for DVD-Video..."


So I guess I'll always use 10.