What are the parameter settings in the resampling process that get changed? Could you outline what the differences would be between Good and Best? Does this alter the internal sampling rate that the Vegas project uses or does that remain constant?

Here's a brief outline of how resample quality parameters change as the type of resample is changed:

PREVIEW: Low quality resample.

GOOD: Low quality resample, fast/poor anti-alias/imaging filter.

BEST: High quality resample, slow/good anti-alias/imaging filter.

Vegas' internal sampling rate is always the project sample rate (or the render sampling rate when the project rate is overridden). One of the first steps when media is read into Vegas is to convert that media to the project sample rate.

There is a definite quality improvement when going to BEST. One good way to try this out is to create a log frequency sweep (in Forge 7.0) and use Vegas to resample, then use a spectrum analyzer tool on the rendered file. as the sweep progresses you can see a curved line showing the primary frequency. At lower qualities you can see a web of aliasing/imaging harmonic distortion artifacts. Even at BEST there is a trace of aliasing distortion at the very high frequencies.

Interestingly, some people find the harmonic distortion artifacts at PREVIEW quality desirable. A lot of people complimented ACID4 for having a much cleaner high end, but it turned out this was due to my accidentally setting the internal resample quality to one similar to PREVIEW in Vegas. While I admit to liking clear high end, I prefer to achieve it without adding distortion by using a good EQ.

So if I read this right, if my source is 16/48K and it all stays that way then the setting is irrelevant. If I copy cda into the T/L it is significant but again once it's resampled it's again irrelevant.
But in my instances what about going from the 16/48K on the T/L out to ac3?
Will I get a better ac3 encode at Best?

Matt,
Thanks for the explanation. I was probably digging a little deeper than you can probably disclose for proprietary reasons.

I was looking for the differences between "fast/poor" and "slow/good". I've encountered in the past when transferring 48Khz DAT Tapes to CD at 44.1Khz, that the real-time sample rate conversion was less than desireable on some hardware devices. Then I have had better success with other hardware (ie TC Finalizer) resampling techniques. In the end I always got the best results by recording into Sound Forge first at 48K, then resampling there using the "Best" setting, and then going to the CD recorder. A little longer of a process, but better results. I just tell my clients, "Be patient, perfection takes time :-)"

I guess I have a similar question as farss asked. If ALL the files within the Vegas project are at 44.1Khz and my Vegas project is also at 44.1Khz, would I see any difference between having the GOOD setting as opposed to the BEST setting. Would there be a performance hit, by having it set to BEST? Would there be any advantage to having it set to Best instead of GOOD in this scenario? My track files always match the Vegas project sampling rate. If for some chance they are different, I always open them in Sound Forge and do the resampling offline, because I just haven't had good success with real-time resampling.

It doesn't seem like from your explanation that there would be any performance hit when playing the files back, because it seems the resample is done when you drag a file into the project. So the only down side is that it would take longer to open a file which is at a different sampling rate than the Vegas project sampling rate.

Rednroll,
setting the audio resample to 'Best' sure blew the ac3 encode times out of the water. And the wierd think is I set it back to 'Good' and it stayed slow. Closed and reopened VV and then it ran at its old speed, same as 'Good'