Page 1 of 1
Does anyone actually know how grain table synthesis works?
Posted: Thu Jul 03, 2014 4:20 pm
by Add9
So Propellerhead apparently invented the synthesis technique as a kind of hybrid between wavetable and granular synthesis specifically for Malstrom... and I know of no other synths that use this method.
But like... how does it work? Is it like some big secret or something? I can't find a single academic paper explaining the technology, whereas for any other synthesis technique there are tons of references. I'm getting into signal processing nowadays and I kind of want to experiment with the grantable idea, but according to Propellerhead a lot of it has to do with the "pre-processing" by an "extremely academic and complex method." And then they don't describe the method or provide any links
Anyway I'm probably going to fuck around with the idea anyway but I was just wondering if anyone had any idea how it might work (or where you could look to research how it works). Obviously Malstrom is a beastly synth so if you could create your own grain tables... the world would pretty much be your oyster.
Re: Does anyone actually know how grain table synthesis work
Posted: Fri Jul 04, 2014 12:35 am
by NinjaEdit
Did you try the manual?
Re: Does anyone actually know how grain table synthesis work
Posted: Fri Jul 04, 2014 8:18 am
by Add9
just looked at it, they actually do give a better explanation in the manual so it was helpful, thanks! they're still a bit coy about the pre-processing but I guess I can't expect them to spill every single bean
Re: Does anyone actually know how grain table synthesis work
Posted: Fri Jul 04, 2014 11:17 am
by Samuel_L_Damnson
Its the same as granular but it only allows the use of preset sounds (wave forms) not ur own sample
Re: Does anyone actually know how grain table synthesis work
Posted: Fri Jul 04, 2014 5:09 pm
by Crimsonghost
If I'm not mistaken, Hamburgs Nuklear is similar in design. I think they call it something different though. I do however remember they had a pretty in depth manual so that might e worth taking a look at.
Or I could just be talking out my ass.
Who cares. I'm getting drunk.
Re: Does anyone actually know how grain table synthesis work
Posted: Sat Jul 05, 2014 11:47 am
by Simulant
It's just marketing nonsense, designed to sell synths. I've seen tons of synths claim to have invented new types of synthesis, but usually they're just using an old form with a few tweaks.
Here are a few:
Tone2 Rayblaster - impulse modelling synthesis. What the hell? Google the phrase and you just get Tone2 results.
Hamburg Audio Nuklear - Pulse train synthesis. Your guess is as good as mine. But it has 4 pulsar train oscillators. I'm not sure if it has deflector shields though.
Tone2 - Nemesis - NeoFM synthesis. An FM synth, with NeoFM. Sounds pretty cool, but sounds pretty average.
Tone2 - Gladiator 2 - HCM synthesis. Yep, another Tone2 one, this time with HCM™. I can't even be bothered to look into what this does, I think these guys just make up names for fun.
Sonic Lab Cosmosf - Advanced stochastic synthesizer. Not a clue.
There's probably tons I've missed, but you get the idea. The main groups of synthesis are subtractive, additive, FM/PM, granular, spectral and physical modelling. Most synths are a variation on these. If there's any other types then I'd like to hear about them.
Re: Does anyone actually know how grain table synthesis work
Posted: Sat Jul 05, 2014 11:54 am
by NinjaEdit
AM falls into the FM/PM group.
Stochastic synthesis would have something to do with statistics.
Re: Does anyone actually know how grain table synthesis work
Posted: Sun Jul 06, 2014 9:48 am
by Add9
Haha, NeoFM synthesis! Or, "regular FM synthesis, but we get to give it a different name because sometimes people say FM when they really mean PM"
Seriously this is how they describe it in the manual:
NeoFM:
"Our exclusive and new approach to FM synthesis. The modulator oscillator 'M' changes the frequency of the carrier oscillator 'C'. Special 90dB 0-phase filters damp the oscillators before FM is applied, providing 2 additional dimensions of sound manipulation."
I guess I'm going to go ahead and trademark NeoAdditive synthesis then, where you add a bunch of sine waves together.
Re: Does anyone actually know how grain table synthesis work
Posted: Wed Jul 09, 2014 9:16 am
by Dub_Fiend
So, putting a couple of filters before the oscillators means they can "trademark" a "new" version of FM synthesis? So so silly.
Re: Does anyone actually know how grain table synthesis work
Posted: Mon Jul 14, 2014 12:06 pm
by koncide
This is my understanding of it, from my own experience as a massive reason fan as well as information from the reason help file.
It works like a sampler I guess. You have a "waveform" of so many individual sections, or "grains". If you play a note in the synth, if will run through the grains, the speed at which it does so is determined by the motion knob. If you set the motion knob to 0, it will just repeat 1 grain, and if you set it to max, it will run through the grains very quickly. Each grain has a slightly different sound. Most of the waveforms in Malstrom are evolving waveforms for this reason. For some waveforms, like the sine, the index slider has no effect since it's just one grain.
You can get really creative with it. Try setting the motion to 0 and then manually adjusting the index slider (this being what controls whereabouts in the grain table the waveform starts from). By doing this you are manually controlling the movement of the grains. Or you can set an LFO to control the index, moving the grains back and forth, or in steps.
Malstrom is one of the most versatile synths I have ever used. I only wish it had some built in effects. In fact, some built in effects tied to the grain-system, for example, a delay where you can control the delay-grains, reversing, glitching them out, would be sick.
Re: Does anyone actually know how grain table synthesis work
Posted: Tue Jul 15, 2014 2:28 pm
by samurai
I'm probably (obviously...) biased here but I think Malstrom is such a great synth because of the quality of the samples used. I just find the quality of the in-built wavetables to be ridiculously musical.