Re: [PD] analogue clipping
As a fan of the "Dirty French" electronic music of late, this is a great thread! ~Kyle > On 5/9/07, Claude Heiland-Allen <[EMAIL PROTECTED]> wrote: > > Claude Heiland-Allen wrote: > > > Otherwise, I think there are some theorems about waveshaping and the > > > amount of oversampling necessary to avoid aliased frequencies and that > > > nasty digital sound... > > > > http://groups.google.co.uk/group/comp.dsp/msg/9576a7a248851937 > > > > > > > > > > ___ > > PD-list@iem.at mailing list > > UNSUBSCRIBE and account-management -> > > http://lists.puredata.info/listinfo/pd-list > > > -- http://theradioproject.com http://perhapsidid.blogspot.com (()()()(()))()()())( (())(())()((( ))(__ _())(()))___ (((000)))oOO ___ PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Re: [PD] analogue clipping
Claude Heiland-Allen wrote: > Otherwise, I think there are some theorems about waveshaping and the > amount of oversampling necessary to avoid aliased frequencies and that > nasty digital sound... http://groups.google.co.uk/group/comp.dsp/msg/9576a7a248851937 ___ PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Re: [PD] analogue clipping
hi hardoff here is an other kind of waveshaping (i think). the idea was to divide a signal by it self so you would get a rectangle. s / s = 1 and further if divide a signal with a range from the signals own value to 1 you don't get a hard rectangle i try to write the formula s / (slider between s and 1) = some rounded nice looking shape try the attached patch. eni On Die Mai 8 12:18 , "hard off" sent: >what's the best and most cpu efficient way to clip a signal in an >analogue fashion, rather than getting nasty digital distortion? > >___ >PD-list@iem.at mailing list >UNSUBSCRIBE and account-management -> >http://lists.puredata.info/listinfo/pd-list s2rect~_dev.pd Description: Binary data ___ PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Re: [PD] analogue clipping
hard off wrote: > what's the best and most cpu efficient way to clip a signal in an > analogue fashion, rather than getting nasty digital distortion? i have used the function x/abs(x+a) applied to the audio input, and it sounded quite ok. a controls the 'steepness' of the 'distortion-function', good values are between 0.1 and 0.8. bis denn! martin ___ PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Re: [PD] analogue clipping
Hallo! > quick tanh looks good. i put in a dollar and my cpu gave me 96 cents change! I did not follow the whole thread, so maybe this was already said. But I would also do the tanh interpolation with a wavetable. You can simply once use this function to generate the wavetable and then use it and it will be very fast (and you can use any function you want) - so you can also draw a little bit around, which might be interesting too ... LG Georg ___ PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Re: [PD] analogue clipping
quick tanh looks good. i put in a dollar and my cpu gave me 96 cents change! ___ PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Re: [PD] analogue clipping
Hallo, Frank Barknecht hat gesagt: // Frank Barknecht wrote: > > Here's a nice approximation with five multiplies > > http://www.musicdsp.org/archive.php?classid=5#238 > > That's a nice version! Attached is a comparision of both in Pd, looks > (and sounds) good. Oh, and even it uses only four multiplies - or rather three and one division. ;) But then the Pd version also calculates the clipped valuesi outside (-3,3), which are always -1 or 1 anyways, so that eats up the savings again. Ciao -- Frank Barknecht _ __footils.org_ __goto10.org__ ___ PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Re: [PD] analogue clipping
Hallo, [EMAIL PROTECTED] hat gesagt: // [EMAIL PROTECTED] wrote: > Try tanh function mapped to the range of about -5 +5 and a small signal input. > http://mathworld.wolfram.com/HyperbolicTangent.html > > Here's a nice approximation with five multiplies > http://www.musicdsp.org/archive.php?classid=5#238 That's a nice version! Attached is a comparision of both in Pd, looks (and sounds) good. Ciao -- Frank Barknecht _ __footils.org_ __goto10.org__ quick-tanh.pd Description: application/puredata ___ PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Re: [PD] analogue clipping
On 08/05/2007, at 13.11, [EMAIL PROTECTED] wrote: > Here's a nice approximation with five multiplies > http://www.musicdsp.org/archive.php?classid=5#238 Indeed. I had to see it. For anyone interested: http://puredata.info/ Members/stffn/tanh-aprox.png/ ___ PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Re: [PD] analogue clipping
Try tanh function mapped to the range of about -5 +5 and a small signal input. http://mathworld.wolfram.com/HyperbolicTangent.html Here's a nice approximation with five multiplies http://www.musicdsp.org/archive.php?classid=5#238 what's the best and most cpu efficient way to clip a signal in an analogue fashion, rather than getting nasty digital distortion? ___ PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list ___ PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Re: [PD] analogue clipping
hard off wrote: > what's the best and most cpu efficient way to clip a signal in an > analogue fashion, rather than getting nasty digital distortion? I quite like [expr~ tanh($v1)] - but it isn't especially cpu efficient. To hear it in action, just about every sound in this has been through several [expr~ tanh($v1)] :: http://www.archive.org/details/ClaudiusMaximus_-_Incidents_At_Sea Otherwise, I think there are some theorems about waveshaping and the amount of oversampling necessary to avoid aliased frequencies and that nasty digital sound... Claude -- http://claudiusmaximus.goto10.org ___ PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Re: [PD] analogue clipping
On Tue, 2007-05-08 at 19:18 +0900, hard off wrote: > what's the best and most cpu efficient way to clip a signal in an > analogue fashion, rather than getting nasty digital distortion? > Have you tried waveshaping? I.e. use your signal as an index to a function (usually stored in a table)? A good place to start would be a stored sigmoid, and then introduce a few small 'glitches' into it to taste. Jamie ___ PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Re: [PD] analogue clipping
hard off wrote: > what's the best and most cpu efficient way to clip a signal in an > analogue fashion, rather than getting nasty digital distortion? waveshaping mfga.sdr IOhannes ___ PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list