Comments

ScottW wrote on 3/22/2005, 6:30 PM
As 2 pass VBR isn't an option for the MC encoder with V4, it doesn't seem like much of an issue, unless you were asking whether it's worth upgrading to V5 just for this capability.
stepfour wrote on 3/22/2005, 6:55 PM
Nope not asking about upgrading. I use standalone MC Encoder 1.04. Just wondering if CBR is better than VBR for shorter videos. This project doesn't have fast motion but does have a lot of panning and zooming along with crossfades and moving text. I guess some f the text could be called fast motion. I'm also wondering if VBR gives any advantage in terms of playability of the burned discs.
B.Verlik wrote on 3/22/2005, 10:12 PM
CBR with a very high bit rate is the best encoding. C = Constant. With a very high bit rate a Variable bit rate can only go down. Not sure of how long the entire video could be, but about 8,000,000 is about as high as you should go for compatibiliy with most players. I believe it's somewhere between 50 minutes and an hour is the longest length the video can be and it may not quite fit if it's close to an hour. I think one pass should be okay with a very high CBR. 2 pass is better for cramming more information on a DVD and making it look better under those conditions. (total amount of space in folder where mpeg2 is stored after render but before making the menu should be 4.38 Gigs)
stepfour wrote on 3/22/2005, 11:04 PM
Thanks. I was thinking along the same lines. The encoding time is not too bad so I might even try one of each and see how some of the composited and cookie cutter sections look.
johnmeyer wrote on 3/22/2005, 11:13 PM
VBR is useful when you encode at very low bitrates (well below 6,000 kbps). At higher rates -- and certainly at 7,000 kbps and above -- it doesn't do much good because the difference between the average and the maximum bitrate isn't enough to make much difference.

Also, doing one pass VBR only lets the encoder change the bitrates over small portions of the video. In particular, it has no way to know that at the end of your masterpiece, you might have twenty minutes of video of a snail crawling across the lawn. With two-pass encoding, it knows all about this, and can therfore use lots of extra bits on the fast scenes.

So, if you are doing a video that is one hour or less, and encoding at an average of 7,000 kbps and above, there isn't much difference in quality between VBR and CBR. When you encode at less than 6,000 kbps average, the quality differences can begin to be noticeable, especially if you use two-pass. By the time you get down to 4,000 (which I generally would not recommend for high quality work), you pretty much HAVE to do two-pass to make the results presentable.
farss wrote on 3/23/2005, 4:09 AM
One trap I've been caught out on with VBR is having the bitrate change too quickly, some players go brain dead at that point, camera iris downs can be a source of this problem as the camera gain goes to max and so does the noise, the encoder was idling along at 500Kb and suddenly shoots upto 8Mb!
If in doubt always check your encoded mpeg-2 file with a bitrate viewer such as: www.tecoltd.com/bitratev.htm.
Bob.
stepfour wrote on 3/23/2005, 6:30 AM
That bitrate viewer is a nice. Thanks for the link. I did analyze a MC 2-Pass VBR and detected a couple of points where the yellow line (bitrate) made a quick jump from 3,200 to over 8,000. That might explain why some of my DVD do make some players go "braid dead," as you aptly put it. Funny thing is, on that clip I made the settings in MC asking that the highest bitrate not exceed 7,500. I'm starting to get a better understanding of what is really going on with VBR encoding. Next I'm going to try a super low bitrate 2-pass just to see what it looks like.

*** Correction ****
I was wrong in the above. I had the max set at 8,000kb
stepfour wrote on 3/24/2005, 4:28 PM
farss, again, thanks for the bitrate viewer. Good tool. It helped me see the huge spikes in bitrate on my VBR encodes. Now I am using a much smaller range between Min and Max. That lets me use 2-Pass encoding but no chance of 4,500KB bitrate spikes being encoded. I tested an encode with 7500Max, 7000Avg and 6500Min and it looks good and plays fine in a portable player that would glitch with prior encodes with 8000Max and 4000Min. I could just use CBR, but believe there is a benefit to having the encoder make 2 passes.