Has somebody a stable Ubuntu Studio NATTY 64-bit?

[Karl Giesing]

> Btw. some cards generate a better sound quality at some sample rates,> regarding to their microchips, not regarding to Nyquist issues etc..
True, but in my experience, this is really noticeable only on consumer-level cards. That has to do with the DAC's, and the ones in pro audio cards (even less expensive ones) are far superior, for the most part.
Then again, I've only had a chance to compare a few models, mostly older ones, so what do I know.

> Hm? When using 32-bit float will there be 'loss' too?
The bit depth mostly affects the signal-to-noise ratio. I personally think 24-bit is fine for "raw tracks," but there's no harm in going higher, if your soundcard even supports it.
Keep in mind, most audio software does all its processing by converting everything to 32-bit float internally, and only drops bits at the final stage (of output, rendering, whatever). Nowadays, I'm betting some software actually does 64-bit float, what with the new 64-bit systems, but I don't have one of those.
Also remember that even at 24-bit, you have a 144dB S/N ratio, which is far greater than the human ear can handle.

> I (nearly) always> kept 96 KHz and 48 KHz recordings at those sample rates. But of cause,> burning an audio CD is interesting for me too, I just had many issues> with Linux audio, regarding to the sound quality. For professional> Studios as far as I remember we had a combination of Neve + Sony digital> recorders, that if I remember correctly, were at 48 KHz and perhaps> 16-bit 'non-float', but to be honest, I never heard a CD. CDs and MP3s> always do sound disgusting. When I did jobs I usually had good luck and> could use analog equipment.
Well, I'm not about to open the whole analog/digital debate again, but the basic answer is if your tracks sound good at the higher sampling rates and bit depth, it should be possible to make them sound good on CD's and MP3's as well.
Most pro audio software, when mastering for CD, has the option of choosing dithering. Good dithering can actually make a big difference in the final master, when converting from higher bit depths to 16-bit. As far as downsampling, the results have mostly to do with how the converters (or software) handle frequencies around the Nyquist rate, but I don't know enough theory to make any recommendations. I do know that for ADC's specifically, oversampling is critical. (The Sony converters probably had huge amounts of oversampling, which is why they sounded so good, but I'm just guessing here.)
Linux audio does seem to be about ten years behind proprietary audio... but that's a topic for another day, and probably unsolvable anyway. Having said that, Linux is growing by leaps and bounds, and proprietary audio is becoming unsustainable financially, so that gap will close sooner rather than later.
-Karl.