to "compand" a signal i'm waveshaping the input with the table opcode and
a hyperbolic tangent function. 

by trial and error i found that a table with these values gives the kind
of results i want:

f1 0 32768 "tanh" -3.5 3.5 0  (i don't want too much distortion)

now i'm looking for a way to do the same but controlling the level of
distortion at k-rate. i found some opcodes like compress, distort,
distort1 and powershape, but i haven't been able to find the proper values
to get the same results as with the table above.

distort is simple and looked very promising for what i want to do. but i'm
either too optimistic about what to expect from the opcode, or it has
some important issues.

any ideas how could i get the effect i'm looking for? perhaps with a
formula with a k-rate factor? i'm afraid i'm not strong at the math...

thanks,


lj

Generally speaking, the higher the amplitude of the signal going into the waveshaper,
the more distortion. What you can try to do is to create a normalising table with GEN 4
http://www.csounds.com/manual/html/GEN04.html
and then use it to normalise the signal as you change the input signal amplitude
to vary the amount of distortion. That's how variable-spectrum 
waveshaping instruments are implemented.

Victor

On 13 Sep 2013, at 22:03, luis jure wrote:

> 
> to "compand" a signal i'm waveshaping the input with the table opcode and
> a hyperbolic tangent function. 
> 
> by trial and error i found that a table with these values gives the kind
> of results i want:
> 
> f1 0 32768 "tanh" -3.5 3.5 0  (i don't want too much distortion)
> 
> now i'm looking for a way to do the same but controlling the level of
> distortion at k-rate. i found some opcodes like compress, distort,
> distort1 and powershape, but i haven't been able to find the proper values
> to get the same results as with the table above.
> 
> distort is simple and looked very promising for what i want to do. but i'm
> either too optimistic about what to expect from the opcode, or it has
> some important issues.
> 
> any ideas how could i get the effect i'm looking for? perhaps with a
> formula with a k-rate factor? i'm afraid i'm not strong at the math...
> 
> thanks,
> 
> 
> lj
> 
> 
> Send bugs reports to the Sourceforge bug trackers
> csound6:
>            https://sourceforge.net/p/csound/tickets/
> csound5:
>            https://sourceforge.net/p/csound/bugs/
> Discussions of bugs and features can be posted here
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"
> 
> 

Dr Victor Lazzarini
Senior Lecturer
Dept. of Music
NUI Maynooth Ireland
tel.: +353 1 708 3545
Victor dot Lazzarini AT nuim dot ie

What about this one? It uses the method I suggested.
(you can play it safely ;)





On 14 Sep 2013, at 00:20, luis jure wrote:

> 
> 
> i think i found a way to get exactly what i want, please see the attached
> orchestra.
> 
> DO NOT SEND TO -odac !!! 
> 
> the orchestra is just to look at the waveform in an audio editor, to check
> that for a linear input, the output goes from linear to distorted depending
> on kdist.
> 
> 
> 
> 
> Send bugs reports to the Sourceforge bug trackers
> csound6:
>            https://sourceforge.net/p/csound/tickets/
> csound5:
>            https://sourceforge.net/p/csound/bugs/
> Discussions of bugs and features can be posted here
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"
> 
> 
> 

Dr Victor Lazzarini
Senior Lecturer
Dept. of Music
NUI Maynooth Ireland
tel.: +353 1 708 3545
Victor dot Lazzarini AT nuim dot ie

on 2013-09-14 at 09:44 luis jure wrote:

> i attach the new version. 

ugh, yet another bug in the UDO...

here is the correct version...


	opcode	varws, a, aki

; requires audio input normalized to 1
; works best with tanh or sigmoid transfer tables
; 0 < kist <= 1

ain, kdist, ifn	xin
kscale	table	kdist*.5, ifn, 1, .5
adist	table	ain*kdist*.5, ifn, 1, .5
	xout	adist/kscale

	endop