[LAD] detailed spectrum for wide freq range

Andrew Gaydenko a at gaydenko.com
Fri Mar 12 12:42:27 UTC 2010


On Friday 12 March 2010 15:20:33 fons at kokkinizita.net wrote:
> Do you want a logarithmic *scale* or a logarithmic *analyser* ?
> In other words what exactly are you trying to do ?

> How accurately you can measure certain things (or if
> they will be visible at all) depends on the analyser
> bandwidth and the levell of its sidelobes. For Jaaa
> the BW is constant over the F range, and all sidelobes
> are below -80 dB (and decreasing as you move away from
> the center frequency). For Japa BW is function of F and
> sidelobes are below -40 dB.

The idea (under discussing at some of russian DIY audio forum) is to use 
multitone signal, compare it before and after audio amplifier and (having such 
*multiple* results for *different* amplifiers) to try get some - at least 
partially useful - correlation between results and subjective audio 
perceptions (in particular, "clearance" - don't know an appropriate English 
term; it's about how easy different instruments can be "separated" in brain).

I'm not sure it is a right way, but would like to take participation in the 
experiment (especially with my last SE-invention :-)).

Fellows (using appropriate software analyzers under MS win) show very clear 
spectrum plots for original multitone signal (ten -20db tones in 20Hz - 20KHz 
range). Having single-boot Linux, I'd want to get something the same :-)


Andrew



More information about the Linux-audio-dev mailing list