I’ve been getting really, really into physical modelling with my eurorack. The results are often dissatisfying, mystifying or otherwise confounding, hence why I love it. I think with other patching styles I rarely had a specific goal in mind (don’t get me wrong, I love aimless patching), but saying something like “today I’m going to make a cello-feeling instrument” has really lit a fire in me. I feel so involved in the process! For the first time, I really feel like I’m doing synthesis in the same way someone might do alchemy (or magic [or science]). Reading through this thread and listening to the examples people have posted was really the inciting incident for this new obsession, and has been a huge help in wrapping my head around /waves hands/ all this. So thanks!
Yesterday I spent the entire afternoon trying to coax a bowed string-feeling patch out of my rack. I’m pretty chuffed with the results, so I thought I would share a recording and a detailed patch breakdown. I’m going to refer to modules generically in the breakdown a la namelessness, but I’ll include a list at the end if you really, really want to know. Here is the recording:
Expression
The patch is played using a joystick and a floor pedal.
The floor pedal is operating a switch with 5v running to one of two outputs. One output isn’t patched, while the other is used to gate both sections of a dual function generator which sends out two envelopes: one for the exciter and one for the delay/body feedback path [see below]. The switch operates momentarily at first, then latches for most of the recording.
The joystick controls a number of parameters using its x and y bipolar outputs.
x is sent to the v/8 input of a body filter, the FM input of a room filter, and through some pretty heavy slewing and attenuation to the v/8 input of a delay line.
y is controlling both of the aforementioned envelopes total respective durations. The particular inputs used are scaled to v/8 but an ‘all’ input or multing the signal to rise and fall will also work. y also controls a crossfading circuit on a pitched delay [see below].
Exciter
The exciter section is a bit more involved than what I’ve done in the past. It feels pretty crucial for a bowed instrument, so I paid a lot of attention to it. In fact its probably the section that received the most tweaks attention across the entire patch.
First, noise is sent into a distortion module. I’m not actually totally clear on whether there is also saturation happening, so try different stuff out! Also, I made some alterations during recording and I can’t remember if the core cello sound used straight white noise, pink noise or extremely low-pass filtered noise, so again, try stuff out!
The distorted noise is then sent to a VCA and enveloped with a pretty short log rise/expo fall function sent by the one of the function generator segments mentioned above. The VCA I used was a four-quadrant multiplier, but in the patch it was just acting as a regular ol’ linear vca with no offset and the envelope is pretty highly attenuated. This is patched into a matrix mixer and then sent to the L input of a stereo delay module. Honestly, I don’t even know if this bit did the much in the final result and might just be an artifact of the patching process.
The bit that does a lot though is this: the distortion I used has an open feedback path. If yours doesn’t, you can make one using a matrix mixer or some other crafty mixing solution. The feedback path is running out to a limiter, then to another 4QM. Here I did use some negative offset and attenuverted the same envelope as before. So basically the exciter signal is being inverted and offset so that it always passes through, with the envelope providing an initial little jump in amplitude to simulate the bows first brush with the strings. The inverted exciter is then sent to the R input of a stereo delay and the R output is sent back into the feedback path of the distortion, as well as being multed to the FM input of the body filter. I found using negative feedback made things sound a little warmer and a little more muted, which I liked, but the differences were pretty subtle. However, this feedback path really felt like the secret sauce of the whole patch, making the overall sound much richer and “scratchier” like what you would expect from bowing a string.
String and Body
The string was taken from the L output of the same stereo delay mentioned in the exciter section. This signal is passed into a low-pass filter (actually I used a state-variable filter and played around with band-pass settings as well). Before moving on to the body, I’ll also mention that the pitch, besides being controlled by the slewed x signal from the joy stick, is also being very slightly FMed by a slowish sine wave lfo whose own FM input is being fed a free-running S&H. The effect of this is a very subtle, non-repeating vibrato. I like it a lot.
Like I mentioned before, the body filter is tracking along with the delay line (although not 1:1), and is being FMed by the delay line via the exciter feedback path. The filter’s resonance is set a bit after noon and is being modulated by noise. Again, play around with the noise you use. Looking at it right now, it looks like I’m using high-pass filtered noise, but I might have changed it after the recording I took. Ooooops.
From the output of the filter we go into a VCA, with the signal being enveloped by the other function generator segment. The response curve on this VCA is variable, but I used something between linear and exponential. The filter I used also has variable gain and I had to play around a lot with the ration between the filter’s gain and the VCA’s before I got it right. The envelope I used here was linear and had about equal proportions of attack and decay, set quite soft, though they would lengthen and shorten with the joystick’s y position.
Finally, the signal was sent through not one but two channels of limiting. Again, I don’t think this was strictly necessary. At one point I had some reverb in the feedback path without realizing it and everything was blowing up so I tried adding the second channel of limiting before I found my mistake.
Out of the limiter and into the matrix mixer, the signal is being sent in equal measure back in to the L input of the delay and out to my room section.
Room
To create a space for the sound, I used another stereo pitched delay, some reverb and a stereo lowpass filter. The reverb is built in to the delay I used but you can roll your own.
First, one channel of the matrix mixer (the one with just the feedback loop, sans the exciter) is sent into the L input of the delay. The delay is set super fast and pitched up in…4ths? I don’t know, I’m not really trained in that way. The y output of the joystick controls the dry/wet control of the delay. This basically creates a copy of the original signal that harmonizes with it, although not super dynamically. You could make it dynamic though with some sequencing or other clever expression modules.
There’s more reverb than you might think. I find placing the filter after the reverb gives a much more intimate feeling to the patch, but the means that a lot of the reverb gets dampened. Its funny, this final delay and reverb section almost feel like they’re still part of the body, I guess because I’ve never heard an instrument outside of an acoustic/reverberant/resonant space like an anechoic chamber.
Finally, the stereo filter I used is sending -12db/8 outputs to a mixer, while -24db/8 outputs are self patched: L is patched to one FM input (this filter has two), R is patched to a stereo skew parameter. And lastly, the x output of the joystick is being sent to the filter’s other FM input where it is inverted and drastically attenuated so that higher pitches mean the filter’s cutoff chokes up just a little bit - lower notes ring longer and higher notes ring shorter.
Kit, if you're really dying to know
1xDoepfer 138s Stereo Mixer
1xInstruo Neoni
2xInstruo Vinca
1xInstruo tanh[3]
1xIntellijel Planar[2]
1xIntellijel Duatt
1xMake Noise Mimeophon
1xMake Noise Wogglebug
1xMutable Instruments Beads
1xRitual Electronics Altar
1xRitual Electronics Miasma
1xRitual Electronics Anima
1xRitual Electronics Pointeuse
1xRitual Electronics Krach
1xToppobrillo Sport Modulator
1xWorng Electronics Parallax
1xWorng Electronics Soundstage
I know this breakdown is kind of a lot, but I figured I would write it all out both as a memory aid for myself and to try and give a detailed explanation of what these kinds of patches look like. I’m pretty new to synthesis and even newer to PM, and up until super recently it wasn’t entirely clear to me that eg you need a VCA in your feedback path or that there are a lot of ways of going about the same thing, as in the case of my exciter. Hope this is interesting/helpful to someone, and I would love to hear any feedback (heh) you might have on the patch!
Oh also please forgive the incredibly sloppy playing in the recording. Hitting the right note with an unquantized joystick is so, so hard.
edit: haha, when i uploaded this soundcloud recommended i tag it as “classical” and my phone listened to it and thought it was the orchestral cover of “If I Had a Heart” by Fever Ray that the show Vikings uses as a theme song. I’ve got someone fooled…well, some software at least 
