Constrained VBR & Unconstrained VBR

[Craig Seeman]

Basically Constrained sets an upper limit to the variable bit rate. In certain parts of a VBR encode, a very complex frame by frame change can create a very high bit rate peak. Constrained prevents that from happening by setting an upper limit.

[Charles Simonson]

Also, with most Windows Media encoders, the peak bit rate of a constrained VBR encode refers to the max average a bit rate can achieve over the specified buffer time zone. So, if your peak is set to 18Mbps, and the buffer is set to 5 seconds, then for that 5 seconds, the peak is allowed to go over 18Mb, to even say 22Mb if necessary, as long as the average for any five seconds including that 22Mb frames aren’t over 18Mbps total. If the encoder you are using uses this method, and you need tight control over the max bit rate that can be applied, you need to either set the buffer time lower or the peak bit rate lower and closer to the average bit rate than you normally would with other codecs.

[Ben Waggoner]

I’m sure I covered this in there somewhere…

Anyway, I recommend you stay away from non-peak constrained VBR. There’s pretty much always a maximum peak you’d want to use in a particular application, even if it is crazy high. There’s no downside to specifying a peak.