Subject: Re: [linux-audio-dev] Dithering idiots
From: Paul Davis (pbd_AT_op.net)
Date: Mon Jun 17 2002 - 13:55:22 EEST
>I'm know this has been asked and answered before in some manner but I'm
>currently thick and low on ideas and none of my searches are working.
>
>When generating a graphical preview of an audio file at various different
>resolutions you obviously get slightly different pictures of the same
>envelope depending on how you extract and interpret the sample information
>at the dithering resolution you require.
>
>I'm pretty sure I was playing around with some kind of interpolation
>algorithm a while ago that (I assumed) would give this consistent view
>at the various resolutions I required. I just never got it to work
>and now can't remember how it worked, what it did or where it was. I'm
>pretty sure it was on a website someone linked on LAD or ALSA but I
>could be wrong.. What I do remember is that it had separate attack and
>release paramters and, hmm, that's where I run out of steam.
>
>Any clues much appreciated.
i don't quite get it. when you render a waveform on the screen, you're
normally not doing anything like interpolation. you're subsampling the
amplitude waveform to a given frames-per-pixel density. i don't see
how you can end up with different pictures of the envelope, except
that the location of a given peak will shift slightly back-and-forth
as you move in and out. the problem with trying to eliminate this
motion is that its actually strictly correct. the true location of the
peak amplitude will, at various zoom levels, frequently be between
pixels, and hence its natural for it to shift left and right between
the neighbouring pixels as you zoom in. is it just that you want to
avoid this tiny motion?
i don't think that any processing on the waveform will avoid this,
since you will still end up with situations where the natural location
of the signal peak is between pixels.
--p
This archive was generated by hypermail 2b28 : Mon Jun 17 2002 - 14:02:01 EEST