We've had some discussions about VMS bit rates in other threads, and I finally found some time to do a little test.

In VMS (Vegas Movie Studio), we have limited options for rendering to MPEG2 (for DVD) from within VMS. Basically, we can chose from two different predefined, "blackbox" templates -- neither of which specifies the target bitrate, etc. However, if rendered in VMS to DV-AVI and passed on to DVD Architect Studio, we then have the option there to specify a bitrate, or to select "fit to disc" when rendering to MPEG2.

The basic question has been "how good is the VMS render to MPEG2, compared to the slower, two-step process (render to DV AVI in VMS, then render to MPEG2 in DVDAS)".

I started with a 3 minute long test project. This is the project I do all my "experimenting" in. Play around with chromakey effects. Experiment with track motion. Learn pan-crop keyframes with some big, still photos, etc. Multiple layers, effects, etc. I should have used just a simple video of my kids running around or something -- it would have been much quicker to render.

OPTION 1: In VMS, click "Make Movie", select "Burn it to DVD".
This option creates two files, "Test-VMS6.mpg" and "Test-VMS6.wav". This took 22 minutes to render just the 124MB mpg file. (Audio file -- 32.9 MB .wav file -- was then rendered very quickly after the video was done.)

OPTION 2: In VMS, click "Make Movie", select "Save it to your hard drive", then select "MainConcept MPEG-2", with the "DVD Architect NTSC Video Stream" template. This also took about 22 minutes to render just the 124MB mpg file. (This option does nothing for audio. Creating the audio file would require another manual operation.)

OPTION 3: In VMS, click "Make Movie", select "Save it to your hard drive", then select "Advanced", "Video for Windows AVI", "NTSC DV Template". This took about 21 minutes to render a 652MB AVI video file, which includes the sound stream. Opened DVDAS, brought the AVI file into the project, then rendered AVI to MPEG-2 using a default bit rate of 8.000 Mbps. Took about 4 minutes 30 seconds to render from AVI to a 178MB VTS_01_1.VOB MPEG-2 file (includes audio).

Using the freeware version of the "Bitrate Viewer" that I just discovered mentioned in another CamcorderInfo.com forum, I determined the following results

(Viewer available at http://www.tecoltd.com/bitratev.htm)

First of all, all three options produced a video stream described as "MPEG-2 MP@ML VBR". I assume this means we get a Variable Bit Rate. Does this also suggest that the results below -- especially the average values, might be highly dependent on the content of the video -- ie how much motion there is, how many scene changes, etc.? I'm thinking, but I'm not sure, that the resulting average bit rate in VBR could end up low because it really didn't need any more data to render a scene with little motion in it (ie slow pans on a still photo).

Also, all three were the same in other, not-understood-by-me aspects: DCT Precision = 9, QuantScale = nonlinear, VBV buffer size = 112, etc.

OPTION 1: "Burn it to DVD"
Peak Bitrate = 7685. Avg Bitrate = 5382.
Peak Q.Level = 20.05. Avg Q.Level = 6.40.

OPTION 2: "Save to hard drive" "MPEG-2" "DVDA NTSC" template
Peak Bitrate = 7685. Avg Bitrate = 5382.
Peak Q.Level = 20.10. Avg Q.Level = 6.40.

OPTION 3: "Save to hard drive" "DV AVI" DVDA render 8.000 Mbps default bitrate
Peak Bitrate = 8097. Avg Bitrate = 6270.
Peak Q.Level = 12.75. Avg Q.Level = 5.42.

Okay, now do any of you experts want to comment? (I'm ready to learn something! You have a willing student!) Does anybody know what "Q.Level" is? I assume lower numbers are better, as Option 3 with the higher bit rate has a lower value.

Also, what does the "MP@ML" part of the stream description mean?

You guys doing pro or semi-pro work with the full Vegas -- what kind of bitrates do you find acceptable?

Oh, and no, I didn't burn all three to a dvd and watch it on a TV to see if I could tell the difference. That would be a sensible thing to do, but would take some more time. And I just spent a bunch of that time typing in this post.

One more thing: the Bitrate Viewer seemed just a little buggy. The number-of-frames count does not reset when you load a new file -- only resets to 0 when you first start the program. Load 3 files and number-of-frames is a cumulative count of all 3. Also, I got slightly different values when reloading the same file. Start the program, load the first file, get X results. Reload same file, get Y results (probably within 1% or so of X results). Reload same file 3rd, 4th time, etc., keep getting Y results. For my tests above, I shut down the program each time and used the first set of results for each file.

All comments welcome...

Tim L

I'm gonna have to think about this for a while. I'll get back to you in a few days.

As far as Q.Level, I was surprised to come up dry in several places I looked. I got one more to try.

Good luck.

Dennis

My explanation in simple terms is that MPEG is a compression technique which does not have individual frames. A frame (frame 1 for example) is compressed. The following frame (frame 2) only has the data which has changed from frame 1, and it goes on to frame 3 etc. So slow pans could actually have a lot of changes if you view one point in the sequence of frames. Trying to guess what the encoder is going to do can be challenging.

From http://tangentsoft.net/video/mpeg/enc-modes.html :"The quantization (Q) level of an MPEG stream is a measure of the amount of data the encoder threw away to bring the bit rate down. Higher Q values mean the encoder threw away a lot of data, implying a low bit rate, but also low visual quality. A low Q level implies a high bit rate, and high visual fidelity."

From http://www.dvd-software.info/dvd-ar...-glossary/4.php :"MP@ML - Main profile at main level. The common MPEG-2 format used by DVD (along with SP@SL).

That is the proof in the pudding if you are going to end up with DVD's to play on a stand alone DVD player connected to a TV.

Me too, which is why I skipped a lot of areas.

Rich

__________________
FireWire Survey

Rich gave a lot of it. Here's some more info.
You got that exactly right. VBR is a two-pass operation. First the video is analyzed to get a profile of it including where the high motion scenes are. On the second pass, using the profile from the first pass, the video is encoded. The encoding uses the min, avg and max bit rate to provide higher bit rates to the high motion scenes and lower bit rate to the lower motion scenes. If there isn't a lot of motion you could well get lower than the avg because the bit rate simply wasn't needed.Rich described Q.level. Most, if not all (I didn't go through them all), of the other parameters are described in this paper that describes the various parameters used by the MainConcept MPEG encoder.As Rich mentioned, this stands for "Main Profile Main Level". MPEG 2 actually incorporates the MPEG 1 standard. It uses profiles and levels, which are groups of parameters to specify what the base encoding will be. For example, there is a Low Level (LL) that is the MPEG-1 max 352x288 pixels and max 30 fps. It also has a max 4 Mbps bit rate.

Main Level (ML) is the MPEG-2 we usually use with max 720x576, max 30 fps intended for SDTV applications. There's also a High Level (HL) for HDTV.

For profiles, there's an Simple Profile that uses no B frames intended for cable TV, Main Profile (MP) with the resolutions and frame rates we are most familiar with and several others. (My source for this is "Video Demystified, 4th Edition" by Keith Jack.It's been a while since I burned a DVD of DV video so I can't really remember what I used. I have done a few DVDs of VHS source material and I find that min 2 Mbps, avg 2 Mbps, max 4 Mbps VBR gives acceptable results. My goal is to get as much on a DVD as possible here and this seems to be a reasonable tradeoff of video quality vs. capacity.It looks like a useful application. Is there any way to communicate these bugs back to its author?

Good luck.

Dennis

PS: Good links, Rich. Thanks.

Hey Guys,

Thanks for the info and for the links (I did check them out).

Someday if I find the time I may repeat my test with a simpler (and faster to render) and more realistic video clip -- something with more motion in it. (Like I said before, maybe just a clip of my kids running around.) Also, abrupt scene changes would probably be good to include in a test. I'll put the clips on a DVD and try them out on a standalone player and TV.

If the render to MPEG from within VMS looks good enough, it would be more convenient than creating a huge AVI file from VMS and then encoding that to MPEG from within DVDAS (where we have the option for 8.000 Mbps). It looks like there's not a huge difference in bitrates between the 3 options I tried (but 8 Mbps from DVDAS does clearly result in a higher avg bit rate and a lower Q level).

Again, thanks for your input.

Tim L

Hey there...

I've gone through some of Edwards Newsletters in regards to bit-rates and file preperation for DVDA.

I realize your initial thread covers VMS 6.0, and I am using Vegas 7, but figured since it is still dealing with Bit-rates, we could keep this in here (unless a new thread is necessary).

My question: I've had some problems with some DVD's I've produced "freezing up" and acting wonky on stand alone players. I have heard it said that it could be the media... I am using DVD+R, and I think I've seen it said that DVD-R tends to work better a lot.... however... when I am rendering out from Vegas7, I am doing VBR, but do 'not' have the Two-pass option enabled... should I enable this option? would it make playback smoother on DVD players?

thanks,

jG

If you use VBR I think you should use 2-pass. I don't know if it will make playback smoother, at least that's not its primary purpose. It's purpose is to better allocate bit rate across the video. I guess that could affect playback smoothness.

The thing that makes some DVD players hiccup is having MPEG-2 files encoded near the max allowed for DVDs. Cheaper players are often engineered a little sloppily and can't quite decode high bitrate MPEG-2 files fast enough. If you're encoding at high bit rates, try the 2-pass first. If that doesn't work, lower the bit rate a bit.

If your bit rate is already well below the max or lowering it doesn't help, maybe its the media. Oh, another thing--slow down the burn rate (from 8x to 4x or 4x to 2x or even 1x) and see if helps.

Good luck.

Dennis

Thanks Dennis,

I have gone back to my original Vegas files, and am re-exporting them with proper 'max' bitrate so that the video will fit properly onto a disc... and this time around will be enabling the Two-pass.

I suspect it is the media itself, but would not rule out slopy authoring on my behalf : )

My stand alone is an older Sony DVD player, and it seems to get some form of C13 error (it constantly stops playing commercial dvd's even), so I would not rule it out either.

I picked up some DVD-R's today, but they are LightScribe, so I'm not sure how well they will work out either. I'll let you know.

Also, for the past while I have been using 4x burn rate for DVD creation , but could go slower.