Some MSF FM Questions

[pone]

MPS and by extension MSF as it’s not monophonic is my favourite FM synth actually (wish it had min. phase freeform eq and convolution though and why not freeform phase as well).
The main trick is eq/filtering and oversampling.
Modulation matrixes or algorithm lists are fun to play with but for getting quick, precise, not ‘dx’ sounds it already has everything.

[Chandlerhimself]

In some cases a simpler implementation is good, but in this case I think making it too simple defeats the purpose. I think 6 operators are nessicary and I wouldn’t be mad at having 8. With less than 6, things done in the DX, FM8, Sytrus, etc won’t be possible. For simple uses I’d prefer to use the merger, but for something more complex I think something without limitations is a better way to go. I’d try to surpass the capabilities of the most powerful FM synths currently available.

[JunSev]

In some cases a simpler implementation is good, but in this case I think making it too simple defeats the purpose. I think 6 operators are nessicary and I wouldn’t be mad at having 8. With less than 6, things done in the DX, FM8, Sytrus, etc won’t be possible. For simple uses I’d prefer to use the merger, but for something more complex I think something without limitations is a better way to go. I’d try to surpass the capabilities of the most powerful FM synths currently available.

Exactly, you know I'm not even beta tester or anything, but this deserve recognition. is important to bring something innovative to the table and demonstrate more capabilities and the powers of the synth by itself, what sense would make to have another synth that just do the same that others already do?.

just wish the best for MSF, even if I don't know if some day I will afford to buy it

[dmbaer]

I think 6 operators are nessicary and I wouldn’t be mad at having 8. With less than 6, things done in the DX, FM8, Sytrus, etc won’t be possible.

This is not correct. Of the 32 DX7 algorithms, all but three (the ones with one carrier and five modulators piled on top) can be decomposed into parallel stacks of four or fewer operators. These stacks are completely independent of each other. With just eleven unique four-or-fewer-operator algorithms, all but three of the 32 DX7 algorithms can be trivially duplicated using two or more parallel four-or-fewer-operator algorithms. Hope this is making sense.

Let me make a clarification, however. I'm not sure what people mean when they say an FM grid would be a good solution. If they are talking about using the main generator grid for this purpose, that is a horrible solution, IMO, because it would be deeply confusing. If on the other hand we're talking about an FM module that has a subgrid for FM routing and modulation (sort of like the modular module having a subgrid), that would be a different matter entirely and an entirely adequate solution. It just comes down to what Vojtech has time to implement. Once again: perfect is the enemy of good.

[Chandlerhimself]

This is not correct. Of the 32 DX7 algorithms, all but three (the ones with one carrier and five modulators piled on top) can be decomposed into parallel stacks of four or fewer operators. These stacks are completely independent of each other. With just eleven unique four-or-fewer-operator algorithms, all but three of the 32 DX7 algorithms can be trivially duplicated using two or more parallel four-or-fewer-operator algorithms. Hope this is making sense.

Let me make a clarification, however. I'm not sure what people mean when they say an FM grid would be a good solution. If they are talking about using the main generator grid for this purpose, that is a horrible solution, IMO, because it would be deeply confusing. If on the other hand we're talking about an FM module that has a subgrid for FM routing and modulation (sort of like the modular module having a subgrid), that would be a different matter entirely and an entirely adequate solution. It just comes down to what Vojtech has time to implement. Once again: perfect is the enemy of good.

I understand that point, but the problem is that it would be missing 3 algorithms from the DX7, which is already limited. FM8, Sytrus, etc have way more algorithms than the DX7, so making something that can only do 87% of what a DX7 can do would be way behind other FM software synths. FM8 has about 60 algorithm presets and there are more that are possible that they didn't include. The same can be said for Sytrus and many other modern FM softsynths. IMO a DX7 isn't a good reference for FM softsynths because I imagine many of the people want modern, harsh, EDM, FM sounds instead of the 80s style DX sounds. By limiting things you might alienate that user base.

To be clear, I'm definitely not talking about the main generator, I agree that would be confusing and a mess. Before I was talking about a grid like FM8, but the more I think about it, something like the modular sub-grid, would work better. Make it 6x6, have only 6(or operators available and allow them to be positioned freely just like the normal grid. Have them labeled A-F and then have other modules available labeled Afb, etc for feedback. The only other thing available might be filters. Allowing filters between operators might be interesting and with all the filters melda has available it might be an interesting and unique feature. On top of that having the osc shapers such as PWM, Sync, etc might be unique too. The only problem I see is either the levels need to be displayed on the grid next to the operator or all the operators need to be displayed in a row with their levels visible next to the grid. Of course there should be presets with all 32 DX7 algorithms included, so people don't have to setup everything from scratch. Hopefully this could keep someone like you happy and also allow more experimental FM without anyone feeling over/underwhelmed.

[werzel]

Filters or other warping abilities between operators would definitely be something that I would be interested in playing with! Good idea Chandler. Parallel chains where FM could be introduced at any point could also prove interesting, eg. creating a pad as a mod source to FM with, or using a reverb impulse as an operator, or or or...I better stop.

[operator]

One amazing FM synth is u-he Bazille.. i would recommend to look for inspiration there aswell... especially due to it s modular nature.. some inspiration could be collected in the u-he Zebra with its FM modules aswell, not the most flexible one but still good creative input...

[pone]

Filters or other warping abilities between operators would definitely be something that I would be interested in playing with! Good idea Chandler. Parallel chains where FM could be introduced at any point could also prove interesting, eg. creating a pad as a mod source to FM with, or using a reverb impulse as an operator, or or or...I better stop.

In short, filtering and eq (mostly a slope) and keytracking of the input signals is what makes FM controllable and convenient (importantly for the purpose of controlling modulator amplitudes) and that is already inside MSF, I suggest you try it. You don't have to mess with a matrix, it removes so much randomness out of FM, don't have to make feedback, you take/add exactly the (small bandwidth) wave what you want in a melda oscilator and then add fm to it to add and change the partials.

[werzel]

Unfortunately, I don't have the time to dedicate to being a beta tester, so I'll have to guess st what's possible based on conversations until the release.

[pone]

It's also possible in MPS just stuck in monophony

[dmbaer]

Well, this has been an interesting discussion but I've got nothing more to say other than this parting thought: Vojtech does not have unlimited bandwidth. If given a choice between, say, a killer-good re-synthesis capability and adequate FM vs. the opposite, my vote goes unquestionably to the former.

[Chandlerhimself]

Well, this has been an interesting discussion but I've got nothing more to say other than this parting thought: Vojtech does not have unlimited bandwidth. If given a choice between, say, a killer-good re-synthesis capability and adequate FM vs. the opposite, my vote goes unquestionably to the former.

I agree that Vojtech doesn't have unlimited time, but IMO this doesn't have to be completed anytime soon. He said he wants to continually update MSF and add new features. There is no need to rush features out and create things that don't meet the needs of users. Right now MSF has FM synthesis and although it isn't the easiest to use, it does work. I'd rather wait 2 years for amazing FM implementation, rather than have something soon that will immediately have people asking for more features. I think its better if features aren't rushed and everything is feature packed and of high quality.

[MeldaProduction]

I'd definitely want to make some FM matrix module, but it would be quite simpler from what you can already do using mergers, that's the point about modularity of course. But for quick results it should be cool and theoretically it may provide better audio quality and lower CPU consumption, we'll see.