Login / Register 0 items | $0.00 New @ KVR
JCJR
KVRAF
 
2310 posts since 17 Apr, 2005, from S.E. TN

Postby JCJR; Thu Mar 31, 2016 1:21 pm Re: Moog String Filter

Thanks, I did not know ReaVerb is a convolution plugin either. Great!
camsr
KVRAF
 
6857 posts since 16 Feb, 2005

Postby camsr; Thu Mar 31, 2016 2:03 pm Re: Moog String Filter

Leslie Sanford wrote:
camsr wrote:What's going on with these convolutions? I just got lost there!


You know how you can use an impulse response of a concert hall with a convolution reverb to make it sound as though you're in that same concert hall? It's the same principle. You can use an impulse response of an instrument with a convolution reverb to make it sound as though your synth is being played through the same space. It's a way of giving body to a sound.


But what about the mixing of the frequency response after convolution? Are the instrument IRs flattened first? Or is EQ required?
Image
User avatar
Chris-S
KVRAF
 
2601 posts since 10 Nov, 2013, from Germany

Postby Chris-S; Thu Mar 31, 2016 10:52 pm Re: Moog String Filter

Depends. Convolution already contains a filter curve. EQ is only needed for adjusting.

Basically convolution is a combination of filter and delay (so called linear operations).
What it not can do are non-linear operations (e.g. distortion, compression, pitch shifting).
Elhardt
KVRist
 
43 posts since 11 Oct, 2006

Postby Elhardt; Sat Jun 09, 2018 10:26 pm Re: Moog String Filter

Just came across this 2 year old thread. I should point out that 15 years and more ago, I've done lots of experiments in "string filtering", using everything from 4 short delay lines, to 42 band filter banks. I still have a few audio examples online (though some are gone), and I can tell you what works best or the problems with different methods, but I'd have to hear from people here, otherwise I'm wasting my time commenting on a death thread. It's too bad all that I've done is being forgotten and people are reinventing the wheel. Even on KVR somebody started an entire thread about my Nord Modular violin patch (and I've also emulated an entire symphony orchestra and tons of other stuff too) see link below. If you want more info on the subject just post that you're still hear and interested.

viewtopic.php?f=1&t=149524
Leslie Sanford
KVRAF
 
1595 posts since 3 Dec, 2006

Postby Leslie Sanford; Mon Jun 11, 2018 5:15 am Re: Moog String Filter

Elhardt wrote:Just came across this 2 year old thread. I should point out that 15 years and more ago, I've done lots of experiments in "string filtering", using everything from 4 short delay lines, to 42 band filter banks. I still have a few audio examples online (though some are gone), and I can tell you what works best or the problems with different methods, but I'd have to hear from people here, otherwise I'm wasting my time commenting on a death thread. It's too bad all that I've done is being forgotten and people are reinventing the wheel. Even on KVR somebody started an entire thread about my Nord Modular violin patch (and I've also emulated an entire symphony orchestra and tons of other stuff too) see link below. If you want more info on the subject just post that you're still hear and interested.

viewtopic.php?f=1&t=149524


Thank you for your reply. I remember hearing the Nord Modular violin patch demo. I was very impressed not only by how realistic the patch sounded but also by the idiomatic performance. I reviewed the above thread and found your comment in which you go into some detail describing how the patch was created:

viewtopic.php?p=2159418#p2159418

The description of using oscillator sync modulated by S/H noise and shaped by an envelope for the attack is very interesting. Would you recommend a speed at which the S/H is modulating the pitch of the slaved oscillator?

I find imitating the attack transient of bowed string instruments to be particularly challenging, so that bit of information is very valuable.

Any additional information you would like to share would be greatly appreciated.
JCJR
KVRAF
 
2310 posts since 17 Apr, 2005, from S.E. TN

Postby JCJR; Tue Jun 12, 2018 12:30 pm Re: Moog String Filter

Maybe someday will find the beta pcm freeman string symphonizer samples that I think I recall making after the freeman had the 24 bandpass moog resonator and chaotic analog chorus installed. Didn't yet run across the right tape.

A few months ago dubbed off a low budget vinyl 45 we recorded sometime around 1978 - 1980 in a pretty nice studio before I modified the freeman. Submitted just as a reference in case anyone interested. There are not many good "sweetened" freeman examples online that I've heard. It is just the three-oscillator freeman string sound sweetened with an old-fashioned acoustic reverb chamber and the multiband channel strip EQ on an old Sphere console. Other keyboards I played were steinway grand, hammond c3 and whiteface arp odyssey.

http://errnum.com/MP3Files/01_SkyHighChatt_IfEverComesTheDay.mp3
Leslie Sanford
KVRAF
 
1595 posts since 3 Dec, 2006

Postby Leslie Sanford; Wed Jun 13, 2018 3:51 am Re: Moog String Filter

JCJR wrote:A few months ago dubbed off a low budget vinyl 45 we recorded sometime around 1978 - 1980 in a pretty nice studio before I modified the freeman. Submitted just as a reference in case anyone interested.


Very cool, JCJR! I enjoyed listening to that. Thanks for sharing.
Leslie Sanford
KVRAF
 
1595 posts since 3 Dec, 2006

Postby Leslie Sanford; Sat Jun 16, 2018 6:13 pm Re: Moog String Filter

An attempt at synthesizing a solo violin sound using V-Station:

https://soundcloud.com/leslie-sanford/v ... v-station/
User avatar
Smashed Transistors
KVRist
 
132 posts since 10 Oct, 2014

Postby Smashed Transistors; Sun Jun 17, 2018 7:01 am Re: Moog String Filter

The Moog String filters was designed based on Max Mathews experiments and findings (see page 129 for some info http://cachescan.bcub.ro/e-book/Adriana ... 13-169.pdf). The filter should have a high density of notches and peaks and they must not emphasize a particular note (so, the notches and peaks should not coincide too much on harmonics).

Synthesis of complex formant filtered sounds, with a high density of spectral static peaks and notches can be obtained with granular synthesis, or synched phase modulation as here :

https://soundcloud.com/thierry-rocheboi ... patch-test

If you add to that some tuned comb filter, you have something that Julius O. Smith calls "commuted waveguide synthesis".

These methods are quite computationally efficient.


Another efficient method is to use "sparse filters" i.e. multitap filters (a short multitap delay). I've just started to experiment with these. I think those filters have a great potential but i have to work on a "design" algorithm (genetic or so) to find the correct taps based on a set of constraints.
See you here and there... Youtube, Google Play, SoundCloud...
User avatar
Smashed Transistors
KVRist
 
132 posts since 10 Oct, 2014

Postby Smashed Transistors; Sun Jun 17, 2018 7:50 am Re: Moog String Filter

See you here and there... Youtube, Google Play, SoundCloud...
mystran
KVRAF
 
4979 posts since 11 Feb, 2006, from Helsinki, Finland

Postby mystran; Sun Jun 17, 2018 8:42 pm Re: Moog String Filter

Smashed Transistors wrote:Synthesis of complex formant filtered sounds, with a high density of spectral static peaks and notches can be obtained with granular synthesis, or synched phase modulation as here :

https://soundcloud.com/thierry-rocheboi ... patch-test


The spectrum of this sounds at least approximately plausible, but the texture leaves much to be desired. The short version of it is that it's way to clean.

The physics of bowing a string are approximately such that the bow hairs will slip slightly in a more or less random fashion as the tension from the string exceeds the static friction. One hair slipping might or might not cause other hairs to slip, but the point is that there are constant "micro-slips" (as I've seen in described in the literature) as you drag the bow. This is what puts the string into vibration, but because it is quite a chaotic process, the sound is necessarily somewhat noisy, with a tell-tale granular texture as long as the bow is applied (and then clears up if you lift the bow while the string is still vibrating, at which point the remaining decay sounds more like a plucked string).

ps. I actually used to play violin for years. :)
Image <- plugins | forum
JCJR
KVRAF
 
2310 posts since 17 Apr, 2005, from S.E. TN

Postby JCJR; Wed Jun 20, 2018 9:14 am Re: Moog String Filter

Leslie the v-station example is interesting.

Smashed Transistors wrote:I don't know if these links have already been posted in this thread.

https://www.keyboardmag.com/miscellaneo ... ion-part-i

https://www.keyboardmag.com/miscellaneo ... on-part-ii

Thanks for the links. I know I must have read those at the time of publication in keyboard mag but do not specifically recall anymore. After the years, "new information". :)

I confuse dates of reading and of projects back then, apparently placing some of the things earlier than it really happened. Try to recollect dates on some of them based on where I was living when reading the info or doing the project.

The 24 filter resonator described in the Moog articles (1978) is almost certainly basically the same schematic they sent me on request. I probably called for the schematic sometime before 1980, probably didn't build it until 1981 or 82. Maybe the experimental platform for Moog's experiments were the mods of Jan Hammer's and Neil Diamond's Freeman String Symphonizers, or maybe he built breadboard standalone boxes as well. Dunno.

In the articles he mentioned that the resonances are most audible on pitch bends and vibrato. Freeman had vibrato but not "synth type" pitch bend wheel. It did have a weird feature of a toe switch on the expression pedal that would slew the pitch a semitone low on toe switch press then time-slew back to normal pitch on release. Wasn't musically useful except embarssing cheezy attempts at "country fiddle" tricks. Maybe that was enough for Moog's comments, or maybe he also experimented with analog synths or polymoog. I think polymoog had a wider pitch bend capability but can't recall.

As said earlier, Freeman had transistor key-switching for "not instant" attack and front-panel-controllable release. Also each key had its own diode-resistor-capacitor passive tone shaping filters to convert the square waves to "kinda like a wimpy sawtooth". Or maybe a tone-shaper per octave. Or both. Got the schematics somewhere but can't recall. Lots of parts anyway. It allowed low notes to be mellow like ac bass or cello while allowing higher notes to be brighter ala viola and violin.

Freeman keying was grouped by octave and the resonator was in two banks of 12 filters, a low bank and a high bank. So the freeman was wired to the resonator via 5 wires from the octaves thru 5 pc-mount internal trimmer resistors, so that each octave could be individually panned to drive a proportion of the signal to the low or high resonator bank. With the low octaves "mostly" driving the low filter bank, the middle octaves driving both banks about equally, and the high octaves mainly driving the high filter bank. Adjusted to taste of the ear via screwdriver before buttoning up the instrument.

I always liked "generic broadband strings" instrument setup where low notes sound like bass or cello, morphing into viola and violin as we go up the keyboard. IMO that is still pretty hard to do well with sample mapping. To get "seamless convincing transitions" across the keyboard. It was a long time ago but there were some EMU Proteus and MPS+ keyboard layered patches that worked pretty well, "small ensemble" orchestra on a keyboard where the appropriate fiddle family member plays depending on what note is played on the keyboard. Also some pretty good small ensemble baroque woodwind mappings among bassoon/bass clarined on the bottom, oboes and clarinets on the top.

If someone were to write a resonator plugin, perhaps it could be made MIDI note aware. Either run it on multiple convolution impluses or multiple filter bank settings, or whatever approach used. So that according to MIDI note, the bass, cello, viola or violin resonances would be applied. In the note-overlap between bass and cello, the resonance would gradually morph from one to the other, and then later gradually morph into viola and then violin as one goes up the keyboard.

But with synth feeding such a resonator plugin, it would seem to need to be monophonic to work correctly. Maybe write a MIDI plugin which would accept polyphonic playing and assign the monophonic notes to a bank of synth plugin->resonator plugin chains.

Maybe simpler just to write a poly string synth plugin where each voice has its own private MIDI-note controlled resonator bank? Put it all in one plugin.
kamalmanzukie
KVRist
 
145 posts since 12 May, 2012

Postby kamalmanzukie; Tue Jun 26, 2018 12:36 am Re: Moog String Filter

Smashed Transistors wrote:The Moog String filters was designed based on Max Mathews experiments and findings (see page 129 for some info http://cachescan.bcub.ro/e-book/Adriana ... 13-169.pdf). The filter should have a high density of notches and peaks and they must not emphasize a particular note (so, the notches and peaks should not coincide too much on harmonics).

Synthesis of complex formant filtered sounds, with a high density of spectral static peaks and notches can be obtained with granular synthesis, or synched phase modulation as here :

https://soundcloud.com/thierry-rocheboi ... patch-test

If you add to that some tuned comb filter, you have something that Julius O. Smith calls "commuted waveguide synthesis".

These methods are quite computationally efficient.


Another efficient method is to use "sparse filters" i.e. multitap filters (a short multitap delay). I've just started to experiment with these. I think those filters have a great potential but i have to work on a "design" algorithm (genetic or so) to find the correct taps based on a set of constraints.


now, a commuted waveguide in the simplest terms is basically feeding a delay line with an impulse response instead of white noise correct?
Previous

Moderator: KVR Moderators (Main)

Return to DSP and Plug-in Development