1)Is it just me or do modular software synths have a richer(for want of a better adjective) sound? I have been gearing up to do some recording auditioning various synths mostly VSTs. I am especially interested in FM synthesis, and it just seems that if I implement similar circuits in VCV rack( or Cherry audio nucleus), for example using the Dexter plugin as compared to the Dexed VST, that the modular circuit sound fuller. I have tried a number of synths. Exact is another FM-synth, just doesn’t sound as full.
If so why? I have a couple of hypothesis here. Maybe VST circuits are designed more for keyboard polysynth harmonies, eg they want to emulate strings and pianos and form traditional chords. However, I feel like for example Dexter or multi-oscilator configurations in cherry audio have the same basic circuits. So, my other hypothesis is that the VSTs are actually only simulating FM-synthesis, and approximating it with wavetables. I think some VST synths this is highly likely to be the case, as Exact pretty much draws the wave design. Maybe it has a FM synth algorithm to figure out the basic wave pattern(s), but then it populates the wave table with that. The other hypothesis, is that the Modular software has to be much more precise to accommodate more types of synthesis so used 64bit floating point algorithms, resulting in a better sound, though tends to consume large amounts of CPU.
Ableton have their own built in FM instrument called Operator. It’s a small but extremely capable synth. Probably heard on thousands of dance records.
u-he have their own take at FM/PD with Bazille and to some extent ACE. Both provide audio rate modulation, and afaik Bazille is one of the most original synths out there.
I think of the free VSTs I tried the u-he Tyrell synth was the closest in sound quality, but it is difficult to compare as it is a subtractive synth, which I didn’t really test in VCV rack. The other u-he I tried was the Cheese comb filter, a good synth. I did try to emulate delay comb filters, and I feel that is pretty easy to do digitally.
A question that is near and dear to my heart. First answer: No! VCV sounds “perfect”. VCV itself doesn’t even change audio going though it in many cases.
But the modules! The are thousands, from hundreds of different developers. Some are very high quality, some are very low quality.
I don’t know if you are old enough to remember the dx-7? The original (depending on your taste) sounded terrible, the dx-7 II sounded really good. Many differences, but both did essentially “the same thing”.
You will have to ask the developer of Dexter about that module in particular. In VCV probably the most popular FM operator is FM-OP from Bogaudio. It uses 8X over-sampling (I believe) to get rid of unwanted FM artifacts. That is the current one does. The original one had no over-sampling of multi-operator patches, but I think that was an accidental omission.
Unfortunately, with VCV modules you don’t really have any reviews to go by, so it’s difficult to know which are going to sound good. In addition there is some personal taste. Some people say they like the artifacts that come out when a module does not remove them. Most people probably don’t like that. And plenty don’t notice or don’t care.
I have found that the modules you see used over and over in people’s videos are also ones that tend to sound good (to me).
If you are deeply concerned about this, you should listen critically to a module before deciding to commit to using it in a patch. It is also easy to see most kinds of digital artifacts on a spectrum analyzer, and there are several good (free) ones.
The signal path in VCV is 32 bit float. It’s possible some modules use 64 internally (as it is sort of more natural in C++), but that would not make much difference in quality or CPU usage. A lot of VCOs in VCV do use 32-bit floating point vectors to make 4 voices take no more CPU than one voice, and that does limit you to 32-bits.
I doubt it? FM (actually almost always phase modulation) is very easy to do. As a dev it’s difficult for me to imagine that someone would implement it all with tables. You would need a ton of them for different C:M ratios, different FM depths, etc…
I love Dexed VST, but it sounds pretty boring without effects, I did try other Vsts that do FM too, and imo these perform way better dry. They sound just like FM-op or Kitchen sink here in VCV. I have a Digitone too, but imo it has quite a unique sound compared to some other FM stuff. Btw Sunvox will soon get a new FM5 module, i’ll be curious to try it and see what i can get with it .
Dexed is super cool, it opens DX7 patches, you can build crazy cool sounds on it, it’s complex . But i hate how it sounds dry, i can’t stand it without at least some delay or reverb. I don’t have that “problem” with any other of my Free vsts.
Still waiting on some refinements from @Squinky since I would like to replace the FM-Ops with Kitchen Sinks (because of build-in folding) - but it works well, and you can get a ton of sounds from it, which are actually very close to the digitone.
On.Topic: I agree with @Olival_Clanaro - the better the effects, especially reverb, the better the sounds of a VST. I never use any build-in effects, they are often not up to stand-alone hardware or software units. And a good delay or chorus or reverb (!) has a lot to do with a good overall sound. We have the luxury of plateau, which is a very versatile and usable reverb in VCV. The more modules you find for your style of music, the better the combination of them will be in VCV. In a VST you are limited to the choices of the developers. Nice sounding VCO but not so great filter? Not a problem in VCV.
Are you sure this isn’t some basic issue like volume differences? In modular systems it’s a convention the oscillator modules output at full volume, but in VST plugins the developers might have reduced the output level by default to accommodate playing chords, for example. (The host applications also often have a default pan law that can reduce the output levels.)
There is no technical reason as such why VST plugins and modular synthesis modules would sound any different. Both can run the exact same code.
They can run the same code, but usually they don’t. You can easily measure the differences, and people post all kinds of comparisons in other communities.
The original poster here didn’t post any sound examples or measurements. And I have no agenda for or against either VST plugins or modular synthesis modules. If there are differences, it would certainly be interesting to have some real evidence.
Oh, for sure. The claim I usually hear is “vcv doesn’t sound as good as X”. It is true this post is more specific. But I hear the more general one all the time.
The original poster’s point was anyway that “modular synthesis” sounds better than VST plugins. But no context was provided for that claim. No sound examples, no measurements, nothing that could actually be reproduced by people in order to test it.
64 bits makes absolutely no difference. I have studied this quite a bit. Vcv sample rate is fully adjustable.it’s true that cherry (I think it’s that one?) runs everything at 96k, which can make bad plugins sound better, but is generally not consider a great solution. There are videos about that issue.
since I would like to replace the FM-Ops with Kitchen Sinks (because of build-in folding) - but it works well, and you can get a ton of sounds from it, which are actually very close to the digitone.