[ardour-dev] Advice on FFT Analysis ... How to Read it, and What Can it Tell You?

[Sampo Savolainen]

On Sat, 2006-11-04 at 20:21 +0000, John Rigg wrote:
> On Sat, Nov 04, 2006 at 12:48:06PM +1100, Shayne O'Connor wrote:
> > i'm gettin ready to master some tracks, so i'm trying to get used to the
> > idea/purpose of FFT Analysis ... here's a composite graph of a short
> > region of all my tracks from a session:
> > 
> > http://s52.photobucket.com/albums/g11/machinehasnoagenda/screenshots/?action=view&current=BassGuitar.png
> > 
> > 
> > ... and here's the guitar and bass alone:
> > 
> > http://s52.photobucket.com/albums/g11/machinehasnoagenda/screenshots/?action=view&current=BassGuitar.png

It seems you posted the same picture twice.

> Your graph looks like the frequency scale is linear, which is fairly
> useless. You need a logarithmic scale (preferably on both axes),
> otherwise the low frequencies are so squashed together that you
> can't see any useful information.

The scale is not linear at all, but it isn't slow enough on the bass
side. I will look into this.

The tool is perfectly usable in the sense that it shouldn't contain any
bugs and all features are in (except what to do on tracks with more than
one stream, suggestions are welcome). Though, the amount of features on
the other hand is very lacking. It has waited for discussion on what
people would like this tool to do.

For this discussion to be as fruitful as possible, a few ground rules
should be made clear to everyone:

 - The window will only display composite analysis. For real time
   analysis there is far better software written by far more
   knowledgeable people than me. Japa for example.

 - Also note that currently the tool does not apply plugins to the 
   data it analyzes. In other words, the analysis is done purely on 
   the data on the selected areas. Changing this is a PITA at the 
   moment though.

Thanks to Shayne for bringing this up! 

 Sampo