| Author |
Message |
Tusker
Joined: Feb 03, 2005
Posts: 110
Location: Texas
|
 Posted: Thu Dec 08, 2005 1:05 pm Post subject:
 Help me understand this cool patch!
Subject description: I love this sound. I'm not sure why it sounds the way it does. |
  |
|
Fellas, I need your help understanding this great patch. I downloaded Steve Harris' patch from this page:
http://nm-archives.electro-music.com/010_NordModular/014_Interesting_Threads/Folder/GuitarFX/GuitarFX.htm
I don't know if he is still involved the NM community. I love what this patch does. What I don't understand is why it works the way it does. Can anyone shed some light?
This is my understanding of the modules and what they are doing as well as some of my questions:
1 - Asymeteric detect and asymetric gain work to amplify the waveform when it has a positive value.
2 - DC offset also amplifies the positive, but additionally reduces the negative amplitudes
3 - Waveshaper low pass filters (rounds) the waveform, but also imposes more of a square shape (more odd harmonics?)
4 - DC offset restores the waveform after waveshaping. (Why does it need restoring?) (and why offset before waveshaping?)
5 - LPF takes of some (alias causing?) high frequencies.
If you have read this far, thank you. Could you please explain the sound design strategy from your perspective? On a superficial level I think I understand what the components would generally do, but not how they work together to accomplish the sonic outcome. Thanks in advance,
Jerry |
|
|
Back to top
|
|
 |
Blue Hell
Site Admin
Joined: Apr 03, 2004
Posts: 22130
Location: The Netherlands, Enschede
Audio files: 208
G2 patch files: 319
|
Posted: Thu Dec 08, 2005 5:56 pm Post subject:
Re: Help me understand this cool patch!
Subject description: I love this sound. I'm not sure why it sounds the way it does. |
|
|
Ok, I'll have try.
| Tusker wrote: |
1 - Asymeteric detect and asymetric gain work to amplify the waveform when it has a positive value.
|
Correct, the positive signal is amplified, the negative signal is untouched (and some spikes might be generated in the X-fader due to different time delays in the positive and negative chain).
| Quote: |
2 - DC offset also amplifies the positive, but additionally reduces the negative amplitudes
|
I'd not call that amplification but .. well .. DC shift :-) The effect indeed is less negative going signal and more positive going, but the total amplitude of the signal (difference between topmost and bottommost signal levels) stays the same, it's not amplified.
| Quote: |
3 - Waveshaper low pass filters (rounds) the waveform, but also imposes more of a square shape (more odd harmonics?)
|
The wave shaper does not really low pass filter, but it does change the harmonic contents of the signal passed to it. In fact it makes harmonics that were not present, something a filter can not do (as filtering is a linear operation, wave shaping is non-linear) Because the input signal has been made asymetric more odd harmonics will be generated than even ones.
| Quote: |
4 - DC offset restores the waveform after waveshaping. (Why does it need restoring?) (and why offset before waveshaping?)
|
It restores the average DC level to zero. In general it's not a good idea to have too much DC in a signal as usualy this will cause unexpected clipping (and unwanted distortion due to that) at some later stage. So it's simply a good idea to remove it. Offsetting before waveshaping is the exception of course, as here the asymmetry is the thing to be acheived (for the odd harmonics).
| Quote: |
5 - LPF takes of some (alias causing?) high frequencies.
|
There shouldn't be too much aliasing, but the filter simply removes excessive high 'd think. Aditionally the resonance can be turned up for interesting sweeps & stuff.
A good way to help gain insight in patches lke these could be to temporarily tap the output from some point internally in the curcuit instead of using its intended output only. It might help to use an oscilloscope (or a simulation of it on a computer) to look at the signals.
_________________
Jan |
|
|
Back to top
|
|
|
Tusker
Joined: Feb 03, 2005
Posts: 110
Location: Texas
|
Posted: Fri Dec 09, 2005 3:58 am Post subject:
Re: Help me understand this cool patch!
Subject description: I love this sound. I'm not sure why it sounds the way it does. |
|
|
Jan, that is so very helpful. Thank you. I understand the basic strategy, based on your comments.
| Blue Hell wrote: |
The wave shaper does not really low pass filter, but it does change the harmonic contents of the signal passed to it. In fact it makes harmonics that were not present, something a filter can not do (as filtering is a linear operation, wave shaping is non-linear) |
Yup, I misunderstood what the transforming function was doing. Doh.  
| Blue Hell wrote: |
Because the input signal has been made asymetric more odd harmonics will be generated than even ones.
|
Could you expand on this statement? I think it will help me learn something more about single cycle waves and the resulting harmonic patterns. I know the basic spectra of saw, pulse and square. But these are symmetric patterns. Is there a concurrent technique for generating even harmonics for example? (I have read the chebyshev technique Rob discussed, but that appears to additive, not a waveshaping technique.)
| Blue Hell wrote: | | A good way to help gain insight in patches lke these could be to temporarily tap the output from some point internally in the curcuit instead of using its intended output only. It might help to use an oscilloscope (or a simulation of it on a computer) to look at the signals. |
Thank you. Still new to the NM and haven't developed good analytic skills in it yet. Thats a great suggestion. 
Jerry |
|
|
Back to top
|
|
|
Kassen
Janitor
Joined: Jul 06, 2004
Posts: 7678
Location: The Hague, NL
G2 patch files: 3
|
Posted: Fri Dec 09, 2005 2:00 pm Post subject:
Re: Help me understand this cool patch!
Subject description: I love this sound. I'm not sure why it sounds the way it does. |
|
|
[quote="Blue Hell"]
| Tusker wrote: |
4 - DC offset restores the waveform after waveshaping. (Why does it need restoring?) (and why offset before waveshaping?)
|
It restores the average DC level to zero. In general it's not a good idea to have too much DC in a signal as usualy this will cause unexpected clipping (and unwanted distortion due to that) at some later stage. So it's simply a good idea to remove it. Offsetting before waveshaping is the exception of course, as here the asymmetry is the thing to be acheived (for the odd harmonics).
Indeed. Aditionally; DC offset can theorettically cost lots of energy from your amp and could also damage some speakers in some cases (though you probably won't be able to get it through the average DAC).
Because DC offset can be seen as a part of a infinitely low frequency LFO mixed in, it can be taken care of with a HP filter. The advantage of using a hp filter is that you don't need to know what the exact offset is nor in what direction. The disatvantage is that using a hp filter can cause phase issues in some cases that might (rarely) cause trouble.
I recomend using a very low HP filter in all audio feedback loops to avoid DC building up and cloging the whole loop. Generally it's probably wise to HP signals like this too unless you know exactly what the DC is.
It's a bit of a jump in reasoning for many people; it's clear that a infinitely low LFO could look like DC offset but to many people it's not so clear that you can filter out DC offset like it were indeed a infinitely low frequency. Still; you can and very often you should.
When taking care off DC offset in wave files recorded from a live analogue mix it often pays off to use a HP filter over a dedicated "remove dc offset" function because that last option might not catch the variable offset that might be introduced by faders going up and down. In that case it's a good idea to get the best of both worlds and use a FIR (convolution) phase-linear filter around -say- 15Hz or so. That will remove all DC offset without introducing any phase issues around your precious bassline (at the expense of quite a bit of cpu load but it need not be realtime).
_________________
Kassen |
|
|
Back to top
|
|
|
Blue Hell
Site Admin
Joined: Apr 03, 2004
Posts: 22130
Location: The Netherlands, Enschede
Audio files: 208
G2 patch files: 319
|
| Posted: Sat Dec 10, 2005 11:26 am Post subject:
|
|
|
I've got the odd/even harmonics vs. symetric/asymetric wave forms mixed up again ...
See : http://en.wikipedia.org/wiki/Physics_of_music it has a section about Harmonics and non-linearities.
_________________
Jan |
|
|
Back to top
|
|
|
|
Forum index » Clavia Nord Modular » NM Classic (NM1 or G1)
FAQ
Calendar
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
Chat Room
