Post by bobbyt on Feb 20, 2011 7:05:59 GMT -5
So I've been wondering about the XDA's output voltage under the new logarithmic map. Someone in another thread, I believe cfelliot, measured and graphed the linear output.
Under that setup, 80 = 2v, 40 = 1v, 20 = .5, etc--cutting volume number in half meant a 3dB decrease. That meant only about 21dB of range (7 doublings between .5 and 80).
What I'm curious about are the levels under the new log layout. Voltage should be of the exponential format 2^L, where L is the volume level, but obviously it has to be scaled.
A pure inverse log would be 2^(L-79), cutting voltage in half at every increment (80 = 2v, 79 = 1v, 78 = .5v, etc). But that would be too aggressive in the other direction, as you'd be silent for most of the range.
If the volume increments are meant to correspond to dB, that would be 2^[(L-77)/3]. So 80 = 2v, 77 = 1v, 74 = .5v. That seems to make the most sense, and would give 80 dB of range, but I was wondering if anyone knows or has measured.
Under that setup, 80 = 2v, 40 = 1v, 20 = .5, etc--cutting volume number in half meant a 3dB decrease. That meant only about 21dB of range (7 doublings between .5 and 80).
What I'm curious about are the levels under the new log layout. Voltage should be of the exponential format 2^L, where L is the volume level, but obviously it has to be scaled.
A pure inverse log would be 2^(L-79), cutting voltage in half at every increment (80 = 2v, 79 = 1v, 78 = .5v, etc). But that would be too aggressive in the other direction, as you'd be silent for most of the range.
If the volume increments are meant to correspond to dB, that would be 2^[(L-77)/3]. So 80 = 2v, 77 = 1v, 74 = .5v. That seems to make the most sense, and would give 80 dB of range, but I was wondering if anyone knows or has measured.
