MSoundFactory Wishlist

[jmg8]

I thought I would start a thread for feature requests for MSoundFactory, so that we have it all in one place.
It will still take some time before we see MSoundFactory, but now it has its new name I thought it would be a good time to start ideas and suggestions for it. Vojtech might see some great ideas and implement them.
So I will kick off first and share some of my ideas and wishes:

[jmg8]

Poly Effects

For example a distortion sounds completely different when it is "per voice" as opposed to a global/standard effect. These type of effects will probably need to be much simpler as they take huge amounts of CPU.
If you are playing just 4 notes with just 4 voice unison then that is already 16 voices and if there is some release that overlaps the next chord, then we have 32 voices! It basically means there would be 32 versions of that effect running at the same time! CPU killer.
However if Vojtech can make some super low CPU basic FXs then we should be able to run them.
There is already some competition out there with these type of effects in the synth and the results are great, the same sound just cannot be achieved with global/standard effects.

I thought we could name these effects as follows to differentiate them from the standard ones:

MPolyDelay, MPolyVerb, MPolyDistortion etc.

[jmg8]

Exciter/Resonator Physical Modeling

I would love to see a module along these lines. Something with a deep editor like MTurboReverb's algorithm designer, where we could make our own physical models.

Basic application:
Exciter = Finger plucking.
Resonator = Guitar string vibrating.
Model = Guitar body.

This can cover so many basic sounds like, guitar, drums, mallets. But can also make for some really creative sounds too. Imagine a lightbulb being hit by a brush inside a metal singing bowl.

Again there a handful of competitors out there with something similar, but in my opinion, they are all pretty limited as there are only a few default materials to choose from. But the ability to make our own with some form of editor/designer would open up a huge range of sonic possibilities. And of course the exciter doesnt have to be a sound generator, it could also be an audio input.

[andrewernick]

I have some questions before I could put together a reasonable wish list:

Is MSoundFactory going to have the same / similar layout as MXXX? Or is it going to be completely new, more comparable to VCV Rack or Reaktor Blocks kind of way, with infinite module space?

Will MSoundFactory host ALL Melda effect plugins like MXXX does or just specific effects (designed to keep the CPU low next to all the synthesis) ?

I still can't quite wrap my head around this beast

The main reason as to why is because in MXXX, we already have MPowersynth, Oscillator, LFO, Mixer Modules, Crossovers etc etc.

Which already provide plenty of complex synthesis options. For example, open six Oscillators and route them into Mixers to frequency modulate them with each other, you can warp between the waveforms while doing so, use modulation modules ... the synthesis possibilites inside MXXX which is labelled an effect plugin are already endless, so how is Vojtech gonna top this with MSoundFactory?

What I could imagine is that MSoundFactory is going to be the digital equivalent of the modular synthesis world. So that an oscillator module doesn't just offer a sidechain input like in MXXX, but different inputs for a sound source to do different things with the oscillator signal (FM, AM, RM and all that good stuff, in more complex fashion) .

Also the connectivity needs to be expanded. So that you can route a signal to infinite destinations for example.

So the more we know, the more we can wish for

What I'd personally love to see:

1. A Wavetable synth module where we can generate our own wavetables via algorithms (and modulate the algorithm, LOL)
2. A Multisampler with various pitch and time stretching algorithms.
3. A Spectrum Synth like in Metasynth.
4. A Grainsynth.
5. A dedicated FM synth.
6. A Modular approach to connecting the modules.
7. A Sequencer.
8. As mentioned above, physical modelling module.
9. Ability to open MXXX inside MSoundFactory and vice versa.

....

Well, all in all I cannot wait for this thing to come out. In whatever shape or form!

[andrewernick]

Exciter/Resonator Physical Modeling

But the ability to make our own with some form of editor/designer would open up a huge range of sonic possibilities. And of course the exciter doesnt have to be a sound generator, it could also be an audio input.

Yes exactly! An editor is what I yet have to see out there. Although I probably don't know enough about the matter

EDIT: I just realized we are way ahead of time, I think it might take a year or more for MSoundFactory to actually be released But the hype is real, we gotta talk about this!!

[jmg8]

Hehe, yes, probably will be a long way off. But it is so much fun to talk about. Plus our ideas could help shape what the product will become.
From my understanding, yes the layout will be like MXXX.
Vojtech has confirmed a sampler will be in there. Plus other new sound generation modules.
All the current FX should be there also, but obviously not per voice.

[werzel]

My brain is picturing 2 different approaches, a sound design architecture and a synth design architecture.
Sound design, where the qualities and rules that dictate the sound's characteristics are formed and then are all passed on to the synth.
Synth design, where we can insert the 'patch' created previously as a base and apply both typical and atypical (Melda'd) synth processing. OR use the synth design end as a customizable layout/ featured synth.
One of my hopes is that conventional thinking is somewhat ignored. Just because I don't see the need for a feature doesn't mean that 20 others might not make use of it. RM a wavetable? FM an FM series by a harmonic series? Organic filter morphing? MPoly effect insertion at any point?
Already sounds impossible.

[Fourstepper]

I would totally love a dedicated FM synth (the Melda modular power would be strong with this one) and even more I would love a sampler

[MeldaProduction]

Hehe guys we are getting a bit ahead of ourselves, first let me do some magic, then we'll see .

[DarkStar]

MMagic, surely?

[jmg8]

Hehe guys we are getting a bit ahead of ourselves, first let me do some magic, then we'll see .

Haha. I knew you would say that. I am just so excited I can't help talk about it!

[mevla]

MMagic, surely?

MMagicMB perhaps ?

[dmbaer]

Hehe guys we are getting a bit ahead of ourselves, first let me do some magic, then we'll see .

Well, better to make suggestions too early than too late.

Here are several things I’d like to have considered. For starters, we know that Melda always boasts of having the best, most pristine wave generator in the business. For MSF, we could also use a waveform generator that is more oriented to speed than perfection of the waveforms. If I recall correctly, both Z3ta+ 2 and Zebra have extensive capabilities for manipulating partials in real time. So, an speed-oriented oscillator module (Turbo Oscillator?) would be most useful.

For noise generation, look to UVI Falcon and what they’ve done: way more than just a few colors – extremely interesting possibilities there. In fact, just look to Falcon in general for great ideas – I recall they have a Karplus-Strong algorithm for general use that’s very cool. But also note that, although wave table synthesis is nicely supported, Falcon currently supplies no capability go build wave tables – not cool at all.

Here’s a real pie-in-the-sky idea: provide an API, at least on the front (sound generation) end that would allow for community-developed sound generation modules, sort of a semi-open-source initiative … and, no, you may not have some of what I’ve just been smoking.

[_TIP_]

Just a reasonable price, the rest i know is going to be amazing!

[dmbaer]

Oh ... one more thing I forgot in my earlier post: good multi-sample support. This could be well-delivered in the form of SFZ file compatibility. That's a large specification (the SFZ spec, that is) of which nobody supports anywhere close to all of the options. But a useful subset is not too large or complicated, mainly the specification of what files go with what note ranges (and maybe velocity ranges, etc.). There are a lot of SFZ sample sets that have been put together, and more than a handful of synths that can import them, so this is not an obscurity. Unfortunately, the ones that do rarely (If ever) document just what subset of the full SFZ specification they implement, and that's most unfortunate.