Riddle me this: synth filters

[Synthoid]

"laser enema" is a good hardcore synth band name.

 

That name makes me cringe.

 

[Dave Weiser]

 

Also the filters, depending on kind and drive parameters, can significantly distort the signal in terms of harmonic and/or inter-modulation distortion. The filter itself usually (in case of the widely known ladder and ota variation) filters it's own distortion components to a degree, so depending on setting it can act as a signal exciter, even with subtle and interesting properties, and in other cases (like the Prophet) it can be adjusted and driven in such a way that it arrives at a particularly "clean" waveform, with interesting slight distortion modulations creating a comely spectrum.

 

 

T

 

This is a great point! Distorted filters have a classic recognizable sound that's commonly found in techno/eletronica like Chemical Bros, Crystal Method, etc.

 

The Kurz does it with a Drive parameter for Resonance on certain filters. (Can be accessed on the DSPCtl page where you also find key-tracking and velocity-tracking.)

 

The PC3 preset DaywalkerBass uses it, with filter freq assigned to the modwheel. http://kurzweil.com/product/pc3k8/audio/#%C2%A0Bass

 

 

[DrSynth]

I don't know about the details of PWM waves, but varying the pulse width varies the harmonic content. I couldn't say whether it introduces new ones or deletes old ones or just changes the relative amplitudes of various ones.

The latter.

 

A pulse wave contains only odd-numbered harmonics. Square waves are pulse waves with 50% duty cycles...or, the ON time is equal to the OFF time. Changing the duty cycle ratio changes the balance of the harmonics, so every pulse wave has it's own harmonic structure/sound.

 

If you get a chance to work with one of the analog synths that have continuously variable waveforms (as opposed to discrete switching, where only fixed values are available) you can really hear what the effect of changing the pulse width does.

 

dB

 

Actually pulse waves have odd and even harmonics. In general the ones missing are related to duty cycle -- a 1/8 width pulse wave is missing every 8th harmonic. 1/2 width missing every 2nd -- leaving odd only, thus the classic square wave. And the relative amplitude of the harmonics present also varies with the pulse width, with the lower harmonics at less relative amplitude to the upper ones as the width gets narrower.

 

Manny

[Dave Weiser]

"laser enema" is a good hardcore synth band name.

 

That name makes me cringe.

 

Stole it from George Carlin.

 

[video:youtube]

 

It's right up front, 18 seconds in.

 

 

 

 

[Odyssian]

Another modular owner here. I've been surprised with how much a single patch can change sonically by substituting a different filter (all other settings staying the same).

 

I just found this on youtube. It's a comparison of different filters. You can hear some differences with the various filters, but I think the manual extreme sweep might be masking some of the differences. Or maybe not... I dunno.

 

http://www.youtube.com/watch?v=4vguV1cBiBM

[Dave Bryce]

Actually pulse waves have odd and even harmonics. In general the ones missing are related to duty cycle -- a 1/8 width pulse wave is missing every 8th harmonic. 1/2 width missing every 2nd -- leaving odd only, thus the classic square wave. And the relative amplitude of the harmonics present also varies with the pulse width, with the lower harmonics at less relative amplitude to the upper ones as the width gets narrower.

 

Thanks for the clarification!

 

dB

[Dr88s]

Think I'll just stick to enjoying synths and stop thinking so much about how they work.

 

Seriously, 3 college level physics courses with grades of 95% or higher and I can barely follow the nuances of this conversation.

 

-sigh-

 

[Bucktunes]

Think I'll just stick to enjoying synths and stop thinking so much about how they work.

 

Seriously, 3 college level physics courses with grades of 95% or higher and I can barely follow the nuances of this conversation.

 

-sigh-

As the old saying goes, be careful what you ask for. You just might get it!

[J. Dan]

This gets into some interesting territory, especially when discussing harmonic content, odd vs even, and what it means to the sound. It makes me think of the Fourier Series and the principles of additive synthesis.

 

http://www.sfu.ca/sonic-studio/handbook/Graphics/Fourier_Series.gif

 

Regardless, I wanted to clarify a bit for non-EE's something that Theo mentioned in passing that I think is worth expanding.

 

OTA's (Operational Transductance Amplifiers) are essentially a current controlled Op-Amp. Many of you are familiar with Op Amps. They are low signal integrated circuits used as preamps, filters, etc. An OTA is basically an Op Amp with current control of the gain. To make a filter requires several stages as well as additional components to convert a control voltage (CV) to current control as well as capacitors to set the cutoff range, and usually some sort of feedback to provide resonance.

 

The CA3080 and LM13600 chips were very commonly used for this. I could be wrong, but I believe the MS20 filter used the 3080. When you get to something like the Curtis chips, they were essentially an application specific chip that included multiple OTA stages with all the other necessary components to have a dedicated Filter chip (or VCA, or VCO, or whatever - the OTA's could be used to make compressors, EG's, filters, VCA's, whatever).

 

I guess the point I'm making is that the goal of technology is to improve the original design. If you consider the original intention was a "perfect" filter, then each evolution made it more perfect. The Ladder Filter was discrete transistors and other components complete with all kinds of imperfections, distortion, drift, etc. Multiple OTA's with supporting components was a slight "improvement" in terms of a consistent perfect filter. Dedicated chips combining even more of the circuit were "better" yet. Now use DSP and digital models, and you can make them almost "perfect". But we like the sound of the imperfections. We like the distortion.

[Tusker]

Think I'll just stick to enjoying synths and stop thinking so much about how they work.

+1

 

If you buy a synth with a pleasing character, you just twist knobs and it sounds good. Or you can add all kinds of enhancement to a more "vanilla" synth. The second approach is more educational, but ... how much time do you have to spare?

 

 

"I spent several man-hours making my Nord Modular sound much moogier. Free running oscillators and a 4-pole filter (modeled on the Rolland 100's filter) were already provided. In addition to the typical signal path, several overlays were done...

 

1) To emulate Model D's oscillator jitter, modulate oscillators slightly with white noise

2) To model oscillator drift, buss the above noise in a mixer module with slow lfo's (with different phases).

3) Model the prefilter mixer overdrive with a soft clipping module.

4) To make the above non-linear, create feedback loops (prefilter back to clipping module, and post filter back to clipping module). This second loop also makes the filter's resonance characteristics slightly moogier. (e.g. lower resonance at lower frequencies.)

5) Since the standard dsp filter/VCA connection doesn't scream enough, add a very mild smooth clipping module with a rectifying-circuit post filter (to me, rectifying means asymmetrically clipping the positive and negative portions of the waveform by different amounts). This gets you closer to the VCA rasp of the Mini.

6) Since the Moog waveforms are not pure, round the waveforms before they go into the pre-filter overdrive but with a slightly resonant keytracked BPF in parallel, to emphasize the fundamental frequency of each oscillator. Alternatively you could add just a bit of sine wave, but dsp oscillators have this habit of losing phase coherence depending on your computing bandwidth so it's safer to fine-tune the character of the synth with amplification/attenuation of the requisite frequencies of the oscillators you have.

7) Lastly, there is a very gentle roll off of high frequencies (about 1db/octave) which adds a kind of roundness. I added a circuit at the VCA stage to emulate that."

[SteveCoscia]

Informative and fun thread. I learned much. Thanks.

 

Lots of smart folks here - what a unique forum.

[Synthoid]

[video:youtube]

 

Thanks, Dave.

[Synthoid]

Informative and fun thread. I learned much. Thanks.

 

Thanks from me as well. Always something new to learn here... and a few good laughs as well.

 

And for Tom, a good ale:

 

http://1.bp.blogspot.com/-INH1f3xB-Uc/T0A1TMTixsI/AAAAAAAAAuw/mNBE2o7YcZA/s400/Dogfish_Head_90_Minute_IPA_200.jpg

 

 

 

[David Emm]

Think I'll just stick to enjoying synths and stop thinking so much about how they work.

Seriously, 3 college level physics courses with grades of 95% or higher and I can barely follow the nuances of this conversation.

-sigh-

 

Heh, likewise, but no one is a master of all disciplines. I feel a bit outclassed here on the tech side at times. I comprehend the underlying theory well enough to know which filter(s) to use and how to trim them for the work at hand, but the more contentious arguments escape me. I had my hardware years, were I learned general technique and where in a chain to put the noise gate. Now, I just drop things into a Logic channel strip and sculpt. I can readily hear the unique n' magical fatness of a Moog LP filter, but these days, you can build virtually any OTHER filter in software and come close to a nice Moog-ish one as well. I manage several things readily by stacking a bit, as we sometimes do with multiple compressors. So the tonal debates can seem a little hazy when we all have a rack of lively tools lined up like pool cues. Its a cockeyed privilege to have had it become so reflexive, I almost feel like an Arteest.

 

 

[Analogaddict]

[...] we like the sound of the imperfections. We like the distortion.

 

Amen! When I studied music history, the teacher said something like "distortion is something every 'natural' musical culture strives for". This is a very interesting thread..!

[The Real MC]

Heck, every traditional instrument has distortion. Even the human voice. Group them in quartets or entire orchestras and they are further distorted.

 

It just seems a lot easier to pull off pleasant sounding distortion with organic materials and nature than with silicon junctions and electrons.

[Theo Verelst]

This gets into some interesting territory, ....

 

I've made an applet that runs on most fully-fledged webbrowsers, which can compute frequency component estimates with an FFT, and has some common example waves to analyze:

 

Wave Laboratory page

 

Wait till the "Waveform lab" and icon appear on the left pane, click the folder icon, then click the icon next to "Theory" and them "View Applet".

 

Then click "Common Waveforms" for instance, and click on the icon next to Sawtooth, perfect, which will put a saw formula in the "Signal Wavefrom expression" in the right pane, and the applet wills tart to compute and show the waveform, and spectrum. Idem for more complicated waves.

 

T

 

[JeffLearman]

WAY cool, Theo! thanks for sharing that!

 

Folks, when you select a waveform link, click "plot signal" and then "plot spectrum". If you don't plot the signal first, the spectrum doesn't change. This is perfectly reasonable for a work in progress!

 

Theo, can you please add a triangle?

[J. Dan]

This is an awesome excercise to understand harmonic content in what you're hearing.

 

Every sound has a fundamental (the lowest sine wave that makes up the sound) and a series of harmonics (a bunch if other sine waves at other frequencies), this can really help visualize what's going on and explain the often discussed even vs odd harmonics.

[Theo Verelst]

On my wavelab page, you could make triangular wave by using a sign ( sgn() ) function to multiply a saw of double frequency.

 

T