Bit rates, etc when rendering

[Stephen Mann]

If your DVD is short – under 70-minutes, then you won’t see any difference from VBR and CBR except for the file size.

If your video is a lot of action scenes with few static shots, then there is not much difference from VBR and CBR. If your video has a lot of static shots, like a talking head with a plain background (where the encoder has to do less work than encoding the video of, say, a football game), then VBR will give you a smaller MPEG file.

So, if CBR fits on your DVD, leave it alone. VBR won’t buy you anything.

Steve Mann
MannMade Digital Video
http://www.mmdv.com

[Mark Prebonich]

We appreciate your insights Stephen. So, it sounds then that the only thing that VBR can buy you is disk space. I also wonder, like the OP, if there are any situations that VBR would be expected to offer improved quality. It sounds as if Mike Kujbida may use VBR 2-pass if there is video of marginal quality.

Since all of my videos are shorter (less than 30 minutes), I render CBR 8,000,000. I also set quality to ‘best’ in properties. Given my project length and computer specs, I haven’t seen much difference in rendering times.

Mike, I think that there are some other good questions in the original post that I would be interested in as well. Thanks.

-Mark

[Mike Kujbida]

Sorry guys but between regular work and free-lance work the past few days, I’m still getting caught up responding to a variety of questions on various Vegas forums, this one included 🙂

Mark, the bulk of my work is under 1 hr. so I use the same render settings that you do and I’ve always been very happy with the final quality.

On the CBR vs. VBR issue, I’m fortunate to have a 1/2″ 3 CCD camera with a really good lens at my workplace.
For example, I shot a play recently that came in a 2.5 hr.
I was positive that it would have to be a 2-disc set but, after doing some tests with the settings from the bitrate calculator I use and doing it as a 2-pass VBR, I only used a single disc and the quality still looked more than acceptable.
This was a professionally lit stage play with very few dark scenes.
Most of the time I was at either f4.0 or f5.6.

If the video is under 70 min., then it’s CBR at 8,000,000.
If it’s over that length, then I use the bitrate calculator I always talk about and that is https://www.johncline.com/bitcalc110.zip.
The only mod I’ve made to it is to set the Safety margin to 5% from the default 1%.
This way, I’m guaranteed that the render will fit.
If the video is over 90 min. long (no matter what the footage looks like), I’ll do it as a 2-pass VBR.
If the quality is less than optimal (i.e. anything shot under less than well-lit conditions), then it’s a 2-pass VBR for anything over 70 min.

To answer Stewart’s question about what 2-pass does, here’s an excellent explanation courtesy of John Cline:

By doing a 2-pass encode, the encoder analyzes the file on the first pass looking for scenes which need the extra bits and those that don’t, on the second pass it does the actual encode using the bitrate allocation strategy it calculated on the first pass.
2-pass encoding also ensures that the encoder more accurately hits the average bitrate target for which it was set.
The only downside to 2-pass encoding is that it takes about twice as long to encode a file.

And this one courtesy of Terry Esslinger:

As you lower a CBR the quality degenerates.
But if you could use a variable bit rate (VBR) and use higher rate where there is more action (motion) and thus more information is needed and a lower bit rate where very little action is taking place, the quality degradation would be much less.
A two pass VBR is like a ‘smart’ VBR.
It looks ahead (actually reads the material twice) so that it knows where it can make the changes and how much it needs to change so that it can save the bits it would normally use in low motion areas and use them in high motion areas.
This double read is why it takes longer.

On Stewart’s question about the various settings in the Video tab of the Custom render options.
If it’s mostly video, I use the Good mode.
If it’s mostly stills, I use Best mode.
It was explained to me several years ago (Douglas Spotted Eagle?), that the image quality of stills is better by using Best mode.
I can’t remember the rationale behind this but, if it’s good enough for Spot, it’s good enough for me 🙂
The other setting I change is to slide the Video Quality slider to it’s maximum setting of 31.

Two final thoughts on this issue.

#1. Even though the bitrate calculator may say so, I never let my maximum bitrate exceed 8,000,000.
I was told (once again, a long time ago) that while higher bitrates are OK for replicated DVDs (i.e. Hollywood masters), burned DVDs (what we use) are better off with lower maximum bitrates.

#2. ALWAYS use good brand name media.
I can’t stress this enough.
I’ve had people come to me with DVDs that are less than 6 months old that are unreadable because they bought the 100-pack that was on sale at the local big-box store for $10.00
I try to use only Taiyo-Yuden.
If I can’t get them, then I’ll use Verbatim.
I refuse to use anything else as it’s just not worth it.
We all put a lot of time and care into making the best video we can.
Why would we take the chance of it not lasting for more than 6 months when, for a few pennies more, we can use the best media possible.
It only makes sense.

[Stephen Mann]

VBR is best when you have moments of lots of activity interrupted by long periods of relatively sedate scenes. It really comes down to how much work the encoder/decoder have to do. Those sedate scenes will be perfectly good at 5Mbps, but those sudden action scenes really want to be encoded at 9 Mbps. If you encode CBR at 8Mbps, then the sedate scenes have bandwidth to spare while the action scenes lose some quality.

Steve Mann
MannMade Digital Video
http://www.mmdv.com