Snippage from the DUG:

Chris Townsend
Administrator
Member # 9519

posted January 29, 2002 08:30 PM                  
-
I've been meaning to post on this topic. Sorry for the delay. Anyway, the sound quality of the sample rate conversion in HD (PT5.3) has been improved considerably, and optimized for the new higher sample rates. The improvement mainly occurs when using certain pull-up pull-down rates. For example, converting from 44,144 to 48,000. Converting from other more standard rates such as, 44.1k to 48k has also been improved in many cases. Using the "best" or "tweak head" settings you can now do sample rate conversion with pretty close to a 144dB signal noise ratio and less than 0.001 db pass band ripple. Also, there is pretty much no difference in quality when converting from common rate such as 44.1k to 48k or 44.1k to 96k compared to 44.1k to 88.2k or 48k to 96k. Obviously I think SRC is a very viable option, although I'm certainly not arguing against using tape if you want that sound. And I almost forgot, the algorithm is on average about 10 times faster than before.

Also, I'd like to add that there certainly is at least a theoretical benefit to using higher sample rates even if the delivery format is using a lower rate. Lowering aliasing distortion when using plugins such compressors, is certainly one benefit. Also, converters do their own SRC internally, but there are limits on the quality considering latency and CPU requirements. With software SRC you could theoretically get much higher quality results, by devoting serious processing power and/or time to the problem.

-

Chris Townsend
DSP/DAE Engineer
Digidesign