From: hermann m. <bru...@we...> - 2013-08-24 07:06:17
|
Am 23.08.2013 19:32, schrieb Andreas Degert: > I have made a new tar archive with some corrections. The plot buttons > should work now (especially without blocking the main program), and > I've added a hanning window after the cepstrum (in function mps) which > smoothes the spectrum and removes the "noise". Not sure how much it > changes the audible result. It looks like the process gone be more stable with the hanning window. I compare guitar tracks with my cocolele, were the spectrum surly didn't match. As well I didn't play, by any means, what is mean to match the "original file". With the version before, the IR-calculation breaks, but now, I get a IR-file, and, it is usable. That mean, it colorize my sound indeed in the direction of the example sound. Just it is a bit wired with the gain influence of the IR-file, some times, I get a file which lead to silence in the overall output, then I need to increase the gain of the example file to get a usable IR-file. I wonder, it seems it didn't help, when I decrease the output volume (FastMeter) of guitarix, before I generate the IR. |