Re: [LAD] (Modular) Synth and Clipping

From: John Rigg <ladev6@email-addr-hidden>
Date: Sat Jun 15 2013 - 19:36:28 EEST

On Sat, Jun 15, 2013 at 02:56:13PM +0100, Aurélien Leblond wrote:
> - in some cases (or let say modules of a synth), clipping is
> implemented more to copie what an analogue system would do than a
> mandatory part of the algorithm... Let's take an example: 2 sin waves
> mixed together of amplitude -1/1 will just have an amplitude of -2/2
> (as long as they are in phase)... A digital mixer without clipping
> would be able to cope with that, but an analogue one wouldn't... and
> that's why the analogue system would clip the signal......right?

An analogue system with enough headroom wouldn't clip. A DAC will
clip in those circumstances.

> - What method of clipping is used will give a "personality" to the
> module: hard clipping, soft clipping, the method used for soft
> clipping, etc...right?

You'd also need to avoid generating harmonics above fs/2 to prevent
aliasing, unless that's part of the "personality" you're going for.

> - Hard clipping is something of the digital world - it doesn't exist
> in the analogue world... right?

Wrong. An analogue amplifier with lots of negative feedback, eg. a
typical op amp or power amp, will usually hard clip.

> - Soft clipping will deform any waves of amplitude -1/1 even if it
> doesn't exceed the accepted threshold, because just before reaching
> the threshold the algorithm will take over and softly make the signal
> reach the maximum amplitude and keep it there until the original
> signal goes back under a set threshold.....right?

Yes, although clipping threshold can be any level you like (within
reason) in floating point.

> - Is there a preferred stage for clipping? In the case of a filter,
> should we clip before filtering, after or both? Or are all these
> options valid and that's what will give an additional personality to
> the filter?

That depends on whether you want to modify the frequency response
before or after the clipping harmonics are added. Both can be valid
options.

John
_______________________________________________
Linux-audio-dev mailing list
Linux-audio-dev@email-addr-hidden
http://lists.linuxaudio.org/listinfo/linux-audio-dev
Received on Sat Jun 15 20:15:02 2013

This archive was generated by hypermail 2.1.8 : Sat Jun 15 2013 - 20:15:02 EEST