Quote:
Originally Posted by hackenslash  But with reasonably efficient Nyquist filters all aliasing can be taken care of during mastering, making it a moot point for the end user, and since the upper limit of human hearing can't hear anywhere near 44.1 KHz, 88.2 seems overkill on an end-user playback device. Somewhere in the 60s, maybe, would be better. I can see why you chose 88.2, at least, as it is a common sample rate. For the end user, 44.1 is a reasonable compromise, as there aren't many people who can hear up into the 20s, so there aren't really any gains in the real world. |
+1
64k/24b would be great and pretty much enough for end-user delivery formats. Even 48k allows for less steep aa/reconstruction filtering and results in more open/natural top end. There is no reason what so ever to deliver 30k audio - I guess the non-audiophools agree on that - and the benefits of 24 bit are common wisdom by now.
Dan Lavry's article on higher sample rates also suggest 60k-ish as an optimum, and since 32k is a somewhat standard lo-fi samplerate any way... How hard can it be to have converters and audio apps run at double that rate?
Using a lossless, open source encoder such as flac would result in substantial data reduction which makes it even more viable.
I'm all for flac encoded 64/24 as the new hi-fi standard, with 48/16 as the compromise for situations where storage/bandwidth is at a premium. For lo-fi applications the 48/16 bit data could still be ogg/mp3 encoded.