Wouldn't that make more sense? Doesn't sound all that complex. I really don't use 24 bit anyway. At least you could have the option....
-
Why can't it read both 24 and 16 bit?
-
It was mentioned in another thread that there were technical reasons, that made it appropriate to have 24 bit. The 24 are stored in a lossless compressed format and are being decompressed on the fly, so it will not take that much of yir RAM.
It also was stated that the samples were re-edited to fit the new engine, IIRC.
tele
-
New thinking. One of the ways they're able to load so many samples is to use losless compression. Chances are, 16-bit files wouldn't get you any better performance anyway.
This is just speculation, but it's possible that having to determine whether a sample is 16- or 24-bit and then add zeroes to the 16-bit files would actually hinder the performance. I don't know that, though.
And why on earth would you "not use 24-bit anyway?!" Your DAW and audio hardware are almost certainly capable of it, so why wouldn't you want to eke every bit of sound quality out of it that you can! It doesn't cost anything other than an insignicant amount of disk storage to run sessions at 24 bits and then dither down to 16 at the last stage.
The principle is that pre-Vienna Instruments VSL uses 16-bit samples, but every time you do anything - gain change, reverb, etc. - you generate more bits of precision. Keeping them as long as possible will usually result in a better-sounding mix.
-
Yes, 24-bits buys you more fidelity, especially as your mixing path grows longer, but at the cost of:
50% more memory (for non-VI data at least), CPU load, bandwidth useage, and latency. You're pushing 50% more data through the entire audio rendering pipeline. Remember, the data decompression very likely happens very early in the data path (probably first, actually).
I run everything (Sonar & Gigastudio) on a single DAW machine, so resources are precious to me, especially with a big symphonic mix. But I'll probably end up switching to 24-bits since that's the native resolution of the libraries. However, memory is the big issue on a single machine like mine, and supposedly the compression should help with that.