Jaime,
File size is not necessarily a good indication of image quality. In fact, if you use the same bitrate settings with CBR and 2-pass VBR, the 2-pass VBR should produce a smaller file at the same image quality. If you use 2-pass VBR, an average bitrate of 6 or so and a max of 7.8, you're not going to get any better quality out of Compressor regardless of file size. It would probably take higher-res source than DV to improve it. Watch out for ultra-high bitrates too. You can cause problems with those, as there is a max total bitrate for DVD.

Also, are you encoding a 24p DVD? If so, your file size will be considerably smaller than a normal NTSC encode. Just worry about how it looks. Just because you have all that space on the DVD doesn't mean you have to use it for your feature.

Hi Jaime,

As I was never able to figure out how to use compressor, I can only speak for QT settings, which I'm using (export->using QuickTime conversion). I found on the web a very useful document here:

http://www.kenstone.net/fcp_homepage...tudio_pro.html

Great thanks to Ken for the document. You might like to try and see if that works differently. Although compressor is using the QT engine, I think.

Presently, I'm working on the long footage but mine consists of parts not longer than 1 hour each (which I still yet to do). I do not have an explanation of why that is hapenning to you. One thing I can think of at the moment is that you may already be getting the maximum quality. Congratulations! Further, I read that normal 4.7GB DVD would hold around 120 min of video, which is about the limit for those DVDs.

By the way, we are talking about normal MPEG-2 encoding here. I know nothing about latest encoding process from HDV.

Cheers,

Zach - Yes, it's going to be a 24p DVD. But I still don't understand. What's the point of buying a Dual-Layer DVD burner if I'm only going to use 1 more GB of space than a regular DVD-R offers? Should I try to return it? Plus, I've read reports on Apple's discussion boards about people being able to get 7.5GB m2v files from compressor, and also of people getting "spectacular" results with other encoding programs (Encore, BitVice...) but I can't get a measly 1hr 42minute video to look good on a dual-layer DVD?!

Something's not right. At the moment, the footage is being exported directly from FCP 4.5 to Compressor at maximum setting for everything. 9.0 average bitrate and 9.0 maximum, 2-pass VBR, best motion estimation, 23.976fps, 16:9 aspect ratio, no filters, GOP set to defaults (12, I think) just to see what happens. If anything, it should look the best it can, right? Even if it doesn't fit on a DVD, or even a Dual-Layer DVD, the m2v file that is being created with the above settings SHOULD be the best quality possible, because every setting is at its highest limit. Or am I missing something here? 9 average bitrate + 9 max bitrate = 9 mbps for every frame. Even if it won't play back well on DVDs, or cause them to crash, this SHOULD BE the highest quality m2v file possible with Compressor 1.2.1. It should finish encoding in the next few hours, and I'll post the results here.

This is getting very annoying. I don't expect it to look as good as the playback from FCP on an NTSC monitor, but it better damn well look at least as good as the test I did with iDVD!!!!! (The only reason I'm not using iDVD is because I need to put subtitles on the thing, and it's for dual-layer dvd, but the quality I got with it was REALLY good.)

Valery - I'll try exporting with the method you describe as well, and post the results here.

Jamie,

Zach is correct about file size as an indicator of quality. A sa point of interest, 2 pass VBR will give you a smaller file size that 1 pass and VBR gives smaller file sizes than CBR. CBR sets the bit rate the same for every frame, wasting space for frames that don't benefit from less compression.

I've found that there is much more going in in encoding mpeg2 than simply clicking the right settings in the software, much more than I understand and have the time to learn. I'v spent a fair amount of time comparing Compressor and BitVice aslo. My results favor BitVice even though it's slower. I'm using BitVice for G3 macs which doesn't take advantage of G4, G5 or duals but still uses the same encoding algorithms as the current release, 1.6.

Another thing to consider is that the hollywood feature movies on DL DVDs usually include many bonus features that take up as much space as the feature movie, so I'm not thinking that your file size is that small. As a matter of fact, I ripped a DVD to my hard drive last week that was from a dual layer disc and all the files including menus, main feature, and bonus features totaled 5.1 gigs.

One more thing to consider. If you like the iDVD encode and still have the project file, control click or right click if you have the mouse for it and choose Show Package contents. Go to the resources folder and you can get the mpeg2 and aiff files it created. The aiff files will need to be converted to AC3 to ensure better compatibility with set top players, but this might be something to help you out until you can get an encode from Compressor that you are happy with.

I'm not sure it's possible to get the same bitrate on a frame-by-frame basis with Compressor. NTSC has to have GOPs of 4-18 frames, which means you'll definitely have a minimum of 3 P and/or B frames per GOP, and those have to be smaller than I frames. I-frame only MPEG-2 is for authoring, not delivery (or at least not delivery in NTSC).

But I don't think it matters anyway. If your encode with Compressor looks worse than iDVD, something must be seriously wrong. I think you should export a Quicktime reference of 2 or 3 minutes of footage from your project and do a batch of encodes at different bitrates to compare. That should be much easier than constantly encoding the whole file. I doubt it will be 100% faithful to the encode you'll get with the whole file, but it should be close enough for comparison. I think iDVD uses CBR at a bitrate of 6 or 6.5 Mbps, so if your iDVD encode looked good then try that in Compressor.