Welcome to %s forums

BrainModular Users Forum

Login Register

FM & PM synthesis in HH4

Tell us what you'd like Usine to do
Post Reply
emiliano
Member
Posts: 28
Contact:

FM & PM synthesis in HH4

Unread post by emiliano » 15 May 2020, 13:20

Hi,
I have spent quite a few hours exploring synthesis possibilities in HH4, in particular Frequency Modulation, which I am fond of.

I have found that pure FM synthesis seems to work well in HH4. However this method of synthesis (even with the new FM osc) quickly produces raw, wild sounds, and provokes detuning of the carrier when pushed just a little to far, in particular in the low frequency notes. Besides, feedback in pure FM does not sound all that great. These findings in HH4 are corroborated by my readings on the web, and they seem to be known shortcomings of true FM synthesis.

Now, what I think would be really great is if we could have Phase Modulation, too. As you know, this is the type of synthesis that was actually used in the Yamaha so-called "FM" synths (DX7, TX81Z, etc.). In Usine, both "Analog osc" and "Digital osc" have a "Phase" inlet. I have tried to modulate this parameter with a -1/+1 audio-rate oscillator, and it sounds very bad. So I suspect that the "Phase" inlets are note coded for proper audio-rate modulation, but only for bloc-rate modulation.

I would very much like to have an audio-rate "phase" input in all Usine oscillators. Would that be possible? This way, we could build our own patches inspired from the amazing TX81Z and other gems from the past, but with the extreme modulation power of Usine ! That would be amazing ! :P

Thank you

SylvainT
Site Admin
Posts: 494
Contact:

Unread post by SylvainT » 15 May 2020, 14:12

Hello Emiliano,

Yes, synthesis possibilities in HH4 have grown seriously : modular wiring, FM Osc, Analog Osc.... Many new thinks arrived or will arrived around granular ....
There's more to do for sure.

With Senso, we thought of the FM Osc. To be quite honest, I'm not sure to understand the difference between FM et PM. Maybe it's a good suggestion.
To do FM, Senso had to create a new module for ensure compatibility of the ones existing, as he needed to change some input capacities (pass "negative" frequencies for example).
Yes, if we put the feedback or FM level to hard, it introduces bad behaviours. But other, it is quit good. I made an emulation of the NI FM8 with great results.
For PM, it would certainly the same thing. But I can confirm that you can do real FM with freq inlet or PM with phase inlet, and it is not due to bloc-rate modulation, as those inlets can receive arrays, but because the module avoid some informations that are (a priori) bad in "normal" situations.

We'll still have other limitations to make real FM, synthesis, modular patching .... with Usine, it is due to the bloc-rate and object programming. But this limitations can be pour in front of the capacities of Usine you don't find in other DAW or in the analog wolrd.

Sylvain

emiliano
Member
Posts: 28
Contact:

Unread post by emiliano » 15 May 2020, 14:53

Thank you Sylvain for your reply !

Here is an example of a patch I did while trying to produce Phase Modulation. Unfortunately it does not sound like a 2-OP PM synth like it should. Can you see a mistake in my patch? Please let me know how I can make it better. Besides I would really like to see your emulation of FM8, so I can take inspiration from it.
PM1.jpg
Many thanks !

emiliano
Member
Posts: 28
Contact:

Unread post by emiliano » 15 May 2020, 17:48

Just in case somebody else is interested in FM synthesis in Usine, here is the basic patch that works for me so far, with one carrier and one modulator. Of course to use it really you need to add all sorts of things like envelope for the modulator OSC, mod fader, etc.
FM1.jpg
Shortcomings are :
- feedback does not sound as good as in standard Yamaha PM synthesizers (DX7 etc). So if you add feedback on the modulator Osc, it will very quickly change the pitch of the carrier in a drastic way. This does not happen in DX7.
- sometimes the sound just changes with no reason. It might be due to phase between the two oscillators but I am not sure.

So it would still be great to have a working "Phase" inlet that could properly be modulated at audio-rate, so we could apply PM instead of FM.

SylvainT
Site Admin
Posts: 494
Contact:

Unread post by SylvainT » 15 May 2020, 20:07

ST-FM-Synth.wkp
(1.76 MiB) Downloaded 242 times
Here's the FM synth I made.
It's a work in progress, so not finished yet, no memory., random to tweak ...
Clic on the wheel to setup the synth
A bit complex, but works fine. If you're familiar with FM8, you'll understand quickly
You can modulate with audio input, use MIDI Input with ADSR or no input, make random sounds .....

Sylvain

SylvainT
Site Admin
Posts: 494
Contact:

Unread post by SylvainT » 15 May 2020, 20:10

So it would still be great to have a working "Phase" inlet that could properly be modulated at audio-rate, so we could apply PM instead of FM.
We are working on many other things, but I'll suggest to Senso. But if possible, it will be a new module. Not immediately for sure.

emiliano
Member
Posts: 28
Contact:

Unread post by emiliano » 15 May 2020, 21:25

Thank you very much Sylvain. Well done !
At first look, just a couple of questions:

What is the role of the "On activation" module wired to the sync inlet of the oscillator? (in fact, what is the use of the sync inlet?)

And also, I notice you used a Stepped point data generator in decibels (-80/0) to drive a volume module. May I ask why this preference over data (0/1) to amplitude inlet of the oscillator? (Just curious - I tried both but I am not sure which is the best method for me).

SylvainT
Site Admin
Posts: 494
Contact:

Unread post by SylvainT » 15 May 2020, 22:19

Sync inlet allows you to restart phase of the oscillator. Not sure it is relevant.
dB :lol: I come from the world of sound, we speak in dB, never between 0 and 1, ahah.
For sure, it's ok with data, and I can certainly win 0.01% of CPU.

It's is just FM4, due to CPU consumption. In fact, object programming like Usine are very powerful, but it consumes a lot of CPU. Easy to use for simple synth, difficult for heavy synth. Maybe you have a look on the Poly Mini Mogg I patched some years ago, a kind of additive and substractive synthesis, based on the Minimoog phylosophy. I think that this patch is rather well done, the sound is great, the possibilities awesome, but it's not really CPU friendly. :oops:

For this synth, the FM side with complex "ADSR" is pretty cool, it works fine with audio input, I like very much the rhythm side also.
The main problem is to have relevant random experiences, it's not always the case. Until now, it's a pity to achieve that on complex patches like this. We'll thinking about that.

emiliano
Member
Posts: 28
Contact:

Unread post by emiliano » 17 May 2020, 11:34

Thank you for your explanations, Sylvain.

Yes I noticed too that patches can become CPU heavy. Yet now I really want to try and make music without any VST at all because I'm bored with all these hanging notes which always occur whatever I do, so that means getting rid of all my Midi and Audio VSTs and thinking more in terms of trigger and ADSR directly with the Usine tools. And the same about the VST synths : I'm trying to do without them, as I like the idea of being able to modulate straight into the parameters of the oscillators & envelopes, and also having each voice of a synth with different parameters (including tuning etc...). We'll see if my PC allows that !

Regarding your FM4, it looks like it has a lot of potential indeed !

User avatar
oli_lab
Member
Posts: 1263
Location: Brittany, France
Contact:

Unread post by oli_lab » 28 Aug 2020, 12:12

Hi !
I saw that you're using the de-mux module I made, have you try the phase modulation V2 module ?
http://oli-lab.org

Win11 Ryzen9/32GB RAM - RME MADIFACE - SSL alpha link 4-16 - OSC capable interfaces

follow OLI_LAB adventures on Mastodon
@olivar_premier@mastodon.social

Post Reply

Who is online

Users browsing this forum: No registered users and 14 guests