Rockbox

The volume is not "translated" correctly if you use %?pv. The current formula, obtained from the source code, is:

(Number of slots in conditional - 3) * (volume - MAXVOLUME)
----------------------------------------------------------- + 2
MAXVOLUME - MINVOLUME


Since any level above 0dB is forced to use the last slot in the conditional, 0dB is forced to use the next-to-last and MINVOLUME is forced to use slot 1, the formula is not right. For example, the iAudioX5 has MAXVOLUME 6, MINVOLUME -73, so the formula will be:

(N-3) * (vol + 73)
------------------ +2
79

Thus, if N=80 (for 1:1 correspondence between different dB values including 0 and slots), the max slot used will be:

77 * (72)
--------- + 2 = 72, and it should be 78
79

If we take into account the "forced" slots, then N=75, and the max slot used will be:

72 * (72)
--------- + 2 = 67, and it should be 73
79

A correct formula can be this:

(N-3) * (vol - MIN - 1)
----------------------- + 2
-1 - MIN

Any other formula in the form y=mx+b can be used to map the dB values to the slots in the conditional, as long as m is calculated as "Number of slots in the conditional - 3) / (real dB steps to map, ignoring the forced values that won't be EVER mapped by this function)". The one presented above is the first I've thought of, and of course it can be improved so it is computed faster. I have a quick'n'dirty test that shows the mappings for an iAUDIO, written in shell script (ouch!).

Please contact me if you want a patch instead of this.
Raúl Núñez de Arenas Coronado