opaqueice;192266 Wrote: > > Suppose for simplicity there are only two types of noise - noise before > amplification N_B, and after amplification, N_A. Then the total signal > going to the speakers is something like > > g*(S + N_B) + N_A, > > where g is the gain, and S the original signal. The final signal to > noise is > > g*S/(g*N_B + N_A) > > So if the noise is mostly from the TP, so that g N_B > N_A, the signal > to noise is largely independent of gain, and attentuating S (and N_B) > doesn't do anything. On the other hand if N_A > gN_B, so the noise is > mostly from after amplification, signal to noise will improve linearly > with gain, and so it's better to attentuate. > > Actually if this simple formula were correct it would always be better > to attenuate, but of course there could be noise components that scale > faster than linearly with gain, not to mention distortion.
This is one of those cases where the maths looks convincing, and its only bacause the proposition was so obviously wrong that I took the trouble to work out why. Noise introduced after the amplifier's gain control has to be irrelevant: for a given output volume, the signal level is whatever it is, and the noise introduced by the amplifier is whatever that is. Nothing earlier in the chain can affect the S or N values, therefore the S/N ratio. The corresponding flaw in the equations is that, for a given output level, g*S is constant. Increasing g will increase the signal level and therefore the output volume. Comparing S/N ratios for different final output volumes is pretty pointless - the SNR will always be better for higher volumes. I think you might have confused yourself by over-simplifying the system, which actually has several points of amplification in it... Ceejay -- ceejay ------------------------------------------------------------------------ ceejay's Profile: http://forums.slimdevices.com/member.php?userid=148 View this thread: http://forums.slimdevices.com/showthread.php?t=34140 _______________________________________________ audiophiles mailing list [email protected] http://lists.slimdevices.com/lists/listinfo/audiophiles
