On Sun, 2012-08-12 at 18:09 +0100, Mauro Santos wrote:
On 12-08-2012 17:11, Ralf Mardorf wrote:
On Sun, 2012-08-12 at 18:02 +0200, Tom Gundersen wrote:
The second case, where the total gain should be <0dB, I would have thought intuitively that doing this purely in software (especially on very faint signals) would be less ideal than doing it in hw (you'd be throwing away the resolution, wouldn't you?), but I'll admit that I don't have the experience to talk about that with any authority.
It's a common misconception that keeping the level lower than 0dB would lead to less "resolution". It depends to the sampling rate and bit depth and less to the level control.
Sampling rate would not matter to level discussions since it limits only the maximum frequency that can be properly sampled or reproduced.
Agree, but it is also "resolution".
For the same bit depth a lower playback output level will yield a lower signal-to-noise or signal-to noise + distortion ratio, thus leading to the same effect of having a DAC of less resolution playing at full scale, so in a way you can say that for lower output levels you have less resolution.
But the lower resolution doesn't become audible that easy, if the bit depth is high enough. It's better to keep the level within reason instead of 0dbFS. Even at 48 KHz 16 bit, headroom is better than maximum level. And if you use totally low bit sampling e.g. 2bit for the C64, you need to play with the level. Higher level doesn't mean better sound quality per se.