squarepusher Squarepusher and FM Synthesis

FM Synthesis is a big part of Tom Jenkinson's synth/compositional vocabulary, and as such he is an astounding virtuoso when it comes to programming with the method. The point of this thread will be to discuss the various dimensions of his technique and hopefully come to a few conclusions about synth model used, workflow, sequencing, and misc DSP.

There are multiple examples of him in action, so I'll just go ahead and start with one I'm sure we're all familiar with: 'The Modern Bass Guitar'.

[youtubehd]os_Jt_k-E-U[/youtubehd]

I'd say the fun starts around 0:50 into the track, for those not familiar with the technique.

---------

I think the biggest mystery for me, and something that I spend a great deal of time pondering is Tom's workflow. Compositions of such depth of modulation, and on such a small scale, must require some sort of easy way to get ideas out that doesn't bog the composer down in the minutiae of sequencing a second's worth of audio. Otherwise the entire process would be infinitely exponentially more grueling than I'm sure it already is. This applies to both the FM techniques of the above track as well as his implementations of granular processes.

As someone who primarily works in Logic I find it very difficult and tedious to A) make discreet, subdivided, steps of modulation at regular intervals of time. like you would find on a 16 step sequencer and B) accurately recall a fine, minute, range of values.

For example: If I had a range of values from 20 - 20,000 in a single automation lane, it would be nigh impossible to automate a clean step from 15,000 to 15,001. Now's that's and extreme example but the point is that the resolution of the sequencing cannot be displayed in any easy manner, nor can it be entered manually with a keypad.

Because it is typically quite difficult to accurately and repeatedly sequenced such precise values I'm inclined to think that Tom would rely on a tracker sequencer for such tracks. Now, trackers are not something that I am intimately familiar with, so I wonder if any of the more seasoned users could provide any insight that might suggest what processes Tom is using. Do you hear things that do not come easily to trackers, thus implying that he is using another method?

For those of us who've had success creating similar sequences in a more graphic sequencer (Logic, Cubase, etc) perhaps you'd be so kind as to share a little bit of your workflow.

----------

As for the synthesizers he's using I can think of at least two likely examples:

A) He's using the The Yamaha FS1R, something which he's confirmed.

B) As an avowed user of Reaktor, he has built his own synth; in which case we'll never know.

Now does anyone have any experience with the FS1R? I wonder how easy is to program and are there an idiosyncratic features that should be considered?

I'll leave it at that for now.

i feel like in reaktor this wouldn't be very hard to do if you assigned say the velocity to a bunch of different fm synth parameters. when i listen to this fm synth wankery i think of reaktor but i could be wrong

from what i've read he has a lot of experience with adding real-time expressive capabilities to synth or synth like sounds, he talks a lot about programming vocoder and bass guitar step filters on the eventide orville. I don't think the orville would be very good for doing fm synthesis though

Tom's sequencer is a Yamaha QY700 ever since Go Plastic. before that, boss dr660.

In that recent interview in that magazine, he said he had made Eventide patches for turning his bass guitar signal into squarewaves, letting him play synth patches that way. Once that's all set up he should have a quick way to make expressive synth sounds.

he has 6 control pedals midi'd up to effects and /synths and has them modulating each other

some preform dual trigger/parameter functions simultaneously so i think he basically[pun tended] sets up his controllers[fat] as he wants them and just picks up the bass and shreds

if only i were able to tender a pun

methix we could be mistaking some other FX as deicated FM synthesis here

  On 1/27/2011 at 7:57 AM, Wall Bird said:

I think the biggest mystery for me, and something that I spend a great deal of time pondering is Tom's workflow. Compositions of such depth of modulation, and on such a small scale, must require some sort of easy way to get ideas out that doesn't bog the composer down in the minutiae of sequencing a second's worth of audio. Otherwise the entire process would be infinitely exponentially more grueling than I'm sure it already is. This applies to both the FM techniques of the above track as well as his implementations of granular processes.

 

he probably just has a shit ton of templates that he saves as he goes along and builds off those

Well Hello Everything was the "recorded to 8-track analogue tape" album right? Hence the Vac Tracks and tape edits and stuff?

So I imagine he played his live rig live onto a track while pedal noodling furiously and then punched in a couple times or did a couple different takes. It sounds like the patch changes significantly partway through the track, from that FM squeal-y sound into a more acid-y sound.

It doesn't seem like a massively complicated patch as opposed to a fairly simple one with a lot of modulation. It's not as if it's a painstakingly modeled 6-operator bassoon sound or something. It's just a couple sine waves freaking the fuck out and warping all over the place. I think Reaktor is probably the most likely culprit since other synths (TX-81Z for example) don't necessarily freak out as kindly to rapid modulation, nor do they really appreciate being bombarded by a ton of controller changes at once, at least in my experience.

Supposedly the FS1R is a beast

i work mostly in logic too, but when i want more meticulous control of parameters i'll usually try to set up a max patch to generate the sort of sounds i want, then bounce a few takes and cut them up in logic.

there's definitely a lot going on in that synth - it sort of sounds like fm tones are being brought out by high-speed cutoff mods on multiple filters

just got home and had a listen and I have to say it's nothing to do with FM but sounds like a bass guitar synth with a high LFO amplitude modulation on the filter with the LFO rate modulated by the volume of the notes played

  On 1/27/2011 at 8:56 PM, soundwave said:

just got home and had a listen and I have to say it's nothing to do with FM but sounds like a bass guitar synth with a high LFO amplitude modulation on the filter with the LFO rate modulated by the volume of the notes played

you don't think cutoff is being modulated in a similar way, giving the impression of fm?

it's funny how people imagine things more complicated than they are. The guy obviously works with a computer sequencer quite a bit and he uses reaktor. Someone here fire up a reaktor fm synth patch, assign the velocity to about 5-6 of the modulation paramaters and try doing a jazz like FM synth keyboard solo. You will find it sounds almost identical to many squarepusher 'fm workout' moments.

i think what's happening in that video is some of it starts out as bass guitar but then it eventually turns into an FM synth patch

edit2: the FS1R is one of the most difficult and annoying to program fm synths available, it is the most powerful one on the market but is a bitch to program.

  On 1/27/2011 at 9:08 PM, Boxus said:

  On 1/27/2011 at 8:56 PM, soundwave said:

just got home and had a listen and I have to say it's nothing to do with FM but sounds like a bass guitar synth with a high LFO amplitude modulation on the filter with the LFO rate modulated by the volume of the notes played

you don't think cutoff is being modulated in a similar way, giving the impression of fm?

you got it man the filter cutoff

I reckon an analogue bass synth pedal modified or put though another synth with an envelope follower or something for the LFO rate

  On 1/27/2011 at 9:11 PM, Awepittance said:

it's funny how people imagine things more complicated than they are. The guy obviously works with a computer sequencer quite a bit and he uses reaktor. Someone here fire up a reaktor fm synth patch, assign the velocity to about 5-6 of the modulation paramaters and try doing a jazz like FM synth keyboard solo. You will find it sounds almost identical to many squarepusher 'fm workout' moments.

That's pretty interesting. I'd never thought about that. I suppose I've been too obsessed with having every parameter under my control and understanding. It never occurred to me to simply let go and let prearranged systems (with a relative unpredictability that comes from assigning multiple parameters to velocity) work as they will and try to harness the outcome in a musically satisfying way. I'd like to read it again, but I recall him explaining in a few different places around the time that 'Ultravisitor' was released that he is very open to the idea of following a machine's suggested methods of functioning. He reiterates that he is open to going with the flow, so to speak, and simply accepting it's inherent properties instead of trying to fit it into a predetermined conceptual box. I believe he says it's about having a two-way relationship with the machines you are using.

I do believe that we're making progress.

Here's how I'd do it with any kind of modular setup:

F2V = Frequency to voltage

EF = Envelope follower

Guitar into F2V

F2V to VCO1

F2V to VCO2

VCO2 into VCA1

Pedal1 to VCA1 control

VCA1 to VCO1

Pedal2 to VCO2 pitch

VCO1 into VCA2

Guitar into EF

EF to VCA2 control

That's what it sounds like to me. He's getting the bass to control two "VCO"s, one which FMs the other but through a "VCA" so the FM amount can be controlled with the pedal. The pitch of the modulating "VCO" can also be pedal controlled to produce those 'wobbles' in pitch you hear at times (where the VCO is in LFO territory).

He's also chopping up whatever he plays a hundred times over so the whole "VCA2" part probably doesn't matter so much, but for practical performance usage it has to be there.

I think the amplitude of the modulating VCO/LFO is controlled by the amplitude of the note played i.e. mellow notes a low warble with slaps and twangs sending into the higher frequencies as it's far too fast for a pedal

good thread. good reading

i think we need to move on to actual reaktor ens.

that would be freak out awesome

im working on mine... based on what ive read here...

but be prepared to wait!

i love sp's fm sounds, like those in tundra 4... or in direct to mental

I'm pretty sure the sounds used as the example in "The Modern Bass Guitar' are not FM synthesis. Trying to make them using FM synthesis is going to get you nowhere fast.

  On 3/1/2011 at 8:20 AM, MrSparkle666 said:

I'm pretty sure the sounds used as the example in "The Modern Bass Guitar' are not FM synthesis. Trying to make them using FM synthesis is going to get you nowhere fast.

it certainly doesnt sound like your regular digital FM tho could well be analogue cross mod or something

The only synthesis i do is fm, so im very familiar with the sound. I dont really hear that much fm going on in his music. I do know he uses this tho.

http://www.vintagesynth.com/yamaha/tx81z.php

Other than a few Fm sounds hear and there, id say he is just a fx freak. ring mod all day!

It could very well be the sound of a sample looping at a very high rate and then frequency modulated at an audio rate. I'm definitely hearing a lot of unusual sidebands being produced, which are common in the aforementioned technique as well as FM.

  On 3/1/2011 at 7:24 PM, Wall Bird said:

It could very well be the sound of a sample looping at a very high rate.

Bingo. Now you are headed in the right direction.