SWAY - a synth modeled after the Yamaha SY77

[oldmica]

+1

[Artie Fichelle]

I suggest a paid public beta. This gives you cash and plus we can make some presets.
And I bet these presets will be better than what Yamaha did, our common knowledge is much better, and we are seasoned sound designers.

[Uncle E]

+1

[planetearth]

Aren't there any public domain presets you could add?
At least for pure AFM sounds.
Since you are not shipping with the original sample ROM, AWM based ones won't work without the ROM.

There are for sure. But to avoid any possible copyrights stuff, the factory presets will be made for scratch.

Isn't this supposed to import Sysex from the SY77/TG77?

We don't really need a bunch of presets as there are already thousands upon thousands of them available online from the original units

Without the factory ROM-based samples. the SysEx presets won't sound anything like they're supposed to. And there are some people who aren't familiar with AWM synthesis, and they might appreciate a few presets to give them an idea on how to get the best sounds from the synth.

Steve

Sure but that's already been addressed by the developer in this thread but if you want AWM2 sample based Synthesis there are several decades worth of Motif, MODX, and Montage presets floating around so you know what it sounds like.

Realtime Convolution and Modulation Synthesis (RCM) in SY77 was unique but it's the FM engine that was the driver behind that

In theory the AFM parts should sound almost identical to the AFM parts in the SY99 hardware or at least as close as Dexed sounds to a DX7

again we don't really need a ton of presets on a synth like this, the original didn't have many after all

If the Motif, MODX, and Montage patches you're referring to work anything like the Kurzweil K2x00 samplers/ROMplers, or the Roland/Korg/Yamaha synths/ROMplers, then the patches (just like the SysEx) don't contain any actual sample data; they're just "containers" with parameter values that point to on-board/ROM samples.

Of course, there were libraries for the Kurzweil samplers and expansion cards for synths that had sample data, but I'd argue that most of the presets being passed around point to on-board/ROM samples—if only because multi-megabyte libraries weren't as easy to "share" when these samplers and synths were new.

All I'm saying is that if the preset expects a "metallic" sample at a specific location in the ROM and there's a "string" sample there in the emulation, the preset won't sound the way it's supposed to.

I have a Yamaha synth with AWM2-based synthesis, so I'm a little familiar with what it sounds like. But as you implied, the RCM synthesis is what made the SY-77 stand out. All that said, I'm pretty sure whoever is designing presets for this emulation will have some good ideas on how to get the most out of RCM synthesis, regardless of the sample data.

Steve

[IvyBirds]

But as you implied, the RCM synthesis is what made the SY-77 stand out. All that said, I'm pretty sure whoever is designing presets for this emulation will have some good ideas on how to get the most out of RCM synthesis, regardless of the sample data.

Steve

FWIW as a point of clarification I didn't really didn't say that and never would. Without a doubt what made the SY-77 stand out was the Advanced FM Engine (AFM) that offered many significant upgrades to the DX7 engine. That engine is still able to do some things that Yamaha's current FM Synthesis engine (FM-X) can't do even when used inside of HALion7 which has all of the SY waveforms

Having Advanced FM in a Multi-Timbral Workstation with CD Quality Sample Playback was also a great leap forward. Layering Deep FM Synthesis patches with samples was and is pretty cool. Layering a sampled piano on top of an FM one gives you instant David Foster 1980s cheese, copying the D50 by using a sample for the transient and FM for everything else in the timbre is cool. But all of that is just layering. I have created many patches that do the same in my Montage and in HALion7 where samples and FM get layered together

RCM is different than that. On paper it's a great idea, you can use a sample as a modulator, who wouldn't want to do that, but in 35+ years of trying I have never been able to use it to actually make musically useful timbres with any kind of consistency. Turns out that samples almost always have to many harmonics, and those harmonics are in different phases and while that is happening the sample itself tends to have vibrato which is fundamentally altering the pitch and probably didn't start with precision anyway and while pitch is fluctuating there are probably dynamics at work on the sample and that sample is also fluctuating in amplitude.

[EvilDragon]

Advanced FM Engine (AFM) that offered many significant upgrades to the DX7 engine. That engine is still able to do some things that Yamaha's current FM Synthesis engine (FM-X) can't do even when used inside of HALion7 which has all of the SY waveforms

Hmmm, which are those things?

Turns out that samples almost always have to many harmonics

So then you start by filtering the sample before using it for RCM.

[bnz]

Not sure about FM-X (never worked with it), but f.em covers most AFM obscurities to my knowledge: multiple input operators, multiple feedback loops, samples as operator input, noise as operator input, looped envelopes. Maybe I forgot something.

[IvyBirds]

Advanced FM Engine (AFM) that offered many significant upgrades to the DX7 engine. That engine is still able to do some things that Yamaha's current FM Synthesis engine (FM-X) can't do even when used inside of HALion7 which has all of the SY waveforms

Hmmm, which are those things?

Mod Matrix and operator routings are different, on SY77 some things are way more convenient

Beyond that the biggest is on SY77 you get up to 3 feedback loops, while FM-X in HALion7 only has one. That's not an issue with the FS1R style operators as they don't need feedback loops, but if you are using any of theSY77/SY99 waveforms it means you can't reproduce those patches

In Montage of course you don't have the SY waveforms you just have FS1R which has "Spectral Forms," "Skirt," and "Resonance" parameters that negate the need and use of feedback loops on operators

They do however include one feedback loop to maintain backwards compatibility with DX7 patches

Turns out that samples almost always have to many harmonics

So then you start by filtering the sample before using it for RCM.

Even then that's still an issue, especially as you still have issues with vibrato and dynamics that the filter can't really deal with

[EvilDragon]

IIRC there is a custom algorithm mode in HALion FM-X which gives you the full matrix? Or am I misremembering?

[wintoid]

IIRC there is a custom algorithm mode in HALion FM-X which gives you the full matrix? Or am I misremembering?

I think even FM Lab has this.

[EvilDragon]

IF FM Lab has it then it's in there yes. DM Lab just exposes the FM layer with a different GUI.

[IvyBirds]

IIRC there is a custom algorithm mode in HALion FM-X which gives you the full matrix? Or am I misremembering?

Yes you can freely build any algorithm you wish

[EvilDragon]

Yes, so that also means you can have many feedback loops, not just one:

https://www.steinberg.help/r/halion/7.0 ... ops_t.html

It's way way waaaaaaay better than SY77 in this regard.

[IvyBirds]

Yes, so that also means you can have many feedback loops, not just one:

https://www.steinberg.help/r/halion/7.0 ... ops_t.html

It's way way waaaaaaay better than SY77 in this regard.

The point I am trying to make but not doing a good job of I guess is that HALion7 has a great DX7 Emulator, it also has FM-X mode which emulates what Montage does with the addition of TX81Z and SY99 waveforms not in Montage

If you go into DX7 mode you are presented with a single Legacy Feedback Loop tied to an Algorithm, if you go into FM-X you don't

Yamaha has five generations of FM including Montage. We had the DX era with only sine waves in the operators

We had the second generation with TX81Z and others which went from 6 ops to four but also came with additional waveforms which means you could do more as you didn't need as many operators. This meant we lost some things but gained some things when it came to FM

Then we had the third generation which was called AFM as was seen in SY77 and SY99 which had the DNA of the DX7 but borrowed and expanded on the second generation by going back to the first generations 6 operators and having even more waveforms. So we added a lot over the DX7

However Yamaha giveth and Yamaha Taketh Away and we also lost algorithms 19 and 20 where each of those have three modulators and two carriers both modulated by one modulator and there is no direct equivalent in SY77. One could argue that doesn't matter as you can add feedback loops going from a modulator to a carrier which can accomplish the same thing which is true, but it's also not the same workflow wise which means the DX7 is able to do things the SY77 can't namely have Algorithms 19&20

The fourth generation was FS1R which was a radical departure from the first 3 generations in both waveform generators and moving from 6 to 8 operators

With FM-X we get the 5th generation which borrows heavily from all of the previous 4 generations and adds new features but also loses things along the way

That's why I am excited for Sway as it's an emulation of the third generation

[DocAtlas]

Not sure about FM-X (never worked with it), but f.em covers most AFM obscurities to my knowledge: multiple input operators, multiple feedback loops, samples as operator input, noise as operator input, looped envelopes. Maybe I forgot something.

F.em is really good, but one thing missing is the ability to send a sample operator through a filter before using it as a modulator. The filters can't be fed back to earlier points in the matrix. Being able to filter a sample before using it as a modulator is crucial. Or, just sticking to harmonicallly very simple samples.