[PD] waveshaping with a lookup table (idiot question)

Michael Berkowski michael at berkowski.net
Mon Nov 21 23:23:02 CET 2005


Greetings,

A 4096 point table works well for limiting the prominence of artifacts 
when waveshaping with tabread4~.  With a linear table that size, you 
will still hear a significant amount of distortion on a sine tone, but 
it will become much less noticeable with more complex input signals.

Keep in mind though, that if your hope is to be able to draw a shaping 
function into the table with a mouse, you will have reduced the table 
size to the resolution of your monitor, throwing away lots of data.  
For this purpose, I've been thinking about building an abstraction to 
copy a hand-drawn table into a larger array, upsampling it with 
interpolation.

Best, Michael Berkowski
http://www.berkowski.net/

> Message: 1
> Date: Mon, 21 Nov 2005 22:19:08 +0100
> From: Frank Barknecht <fbar at footils.org>
> Subject: Re: [PD] waveshaping with a lookup table (idiot question)
> To: pd-list at iem.at
> Message-ID: <20051121211908.GB12197 at fliwatut.scifi>
> Content-Type: text/plain; charset=us-ascii
>
> Hallo,
> james holden hat gesagt: // james holden wrote:
>
>> what i've noticed though is with a small table (100 points) i get a
>> distortion that sounds like bit-depth reduction. it goes with a larger
>> table. surely though, if tabread4~ interpolates this shouldn't happen?
>
> Even with interpolation you can get these kinds of artifacts, so I'd
> recommend using a bigger table. RAM is cheap. Generally it's starts to
> be fun at 512 and above.
>
> Ciao
> --
> Frank Barknecht                 _ ______footils.org_ __goto10.org__






More information about the Pd-list mailing list