Circuit modelling ? How do you know ?

[Local Man]

When I buy a hardware synth, how do I know there are circuits in there and not some little race track with some small little wiggly creatures in there pulling levers and spinning gears and what have you?
I mean really there's no telling no matter what.

[acYm]

in the Settings section of sugarbytes Factory, there's a little window you can peep through and you can see how the sound gets made inside the synth, it's quite neat.

[TheSynthScientist]

Just let your ears be the judge. Simple!

Right, this is the final rule for all plugins.

[Dasheesh]

in the Settings section of sugarbytes Factory, there's a little window you can peep through and you can see how the sound gets made inside the synth, it's quite neat.

That is interesting actually. I like factory.

[Dasheesh]

Just let your ears be the judge. Simple!

Right, this is the final rule for all plugins.

Like.

[Dasheesh]

Wrong thread/

[Ah_Dziz]

With the resources available to everybody these days “circuit modeling” in the sense that they are modeling the components of an electronic system is not hard to get into as the circuits are all available for examination and the math that describes the circuits is not too tricky if math is a thing you do. Whether or not an accurate emulation of a physical system is worth the resources or not is a different matter. There are lots of configurations that don’t sound particularly good at all and the amount of cpu it takes to fully model a real-time electronic system is not negligible. I’d like to see a reaktor type rig that would run spice emulations in a vst shell just for people to play around with and see which bits really make a big difference to the sound. It would be brutal on the cpu but fun and useful for folks to get a better understanding of which parts of the system are noticeably enhanced by fully calculating how an electronic component would respond. Anyhow, I’m not particularly concerned with which products do or don’t really model a board covered in op amps, transistors, capacitors, and such as much as which products let me achieve sounds that are useful to my compositions. Repro is awesome and crazy accurate for pro 1 sounds, but sometimes synth 1 is gonna fit better into my song. I doubt many of these companies really have an incentive to lie about how they make their products anyhow, since almost everything can be demoed and if it sounds crappy it’s still not useful regardless of how accurately the coders modeled it against hardware. I’d like to see more circuit modeling of circuits that would be a pain in the ass to make in real life, but could have potential if sourcing parts, proper circuit layouts, and all the practical parts of building hardware can be ignored. There could be lots of fun hybrid vsts made this way.

[vurt]

Oh I can see that now

"Hi Herr Heckmann, could you just send me the source for Repro-1 please"

I suspect a short answer would follow

Equally, finding a bunch of qualified, respected developers might be even more difficult

we could threaten to tickle his feet.

[andy-cytomic]

For confidence in the accuracy of an analog model it's actually pretty basic: you really need the original hardware that was modelled and then compare the resulting audio to that of the plugin with as many signals as possible (not just basic test signals) and as many knob positions as possible. You should be able to dial in a setting you like on the hardware, and then set the knobs to roughly the same position on the model and get a very similar result, at least one you would be happy using if the hardware was not available. It would be ideal to have two or more hardware units so you can get an idea of inter-hardware tone variation as well, from things like manufacturing and age. Extreme settings of knobs and large signal levels may not be useful musically, but they are often very enlightening as to how much detail has been modelled, and the quality of that modelling.

Of course most people don't have this sort of access to the original hardware, or the skill to carefully match the signal levels and knob positions for the fairest comparison possible.

This approach is scalable, modelling technique agnostic, and also places no value judgements on the results. If done right the end result would be a whole bunch of audio files that everyone can listen to and make up their own mind as to what sounds best for them.

[Stokely]

You get similar arguments with piano plugins...sampled vs modelled.

The bad part is that people pre-form expectations based on the tech.

"Well it's sampled, so it probably won't play quite as well but it will sound great!"
"Well it's modelled, so it will be so very playable, but the sound will be sterile and unnatural!"

Soundwise (putting aside cost, hd space etc), it should only matter: does it play and sound like a piano?

Personally I'm glad different approaches are being taken in these areas, as they hopefully push each other to get ever better.

[zzz00m]

That's a good argument, until you realize that sampling or modeling an acoustic instrument (such as a piano) is an entirely different problem from sampling or modeling electronic synthesizer circuits.

An acoustic piano's sound is made up of much more than the direct hits of hammers on strings. You have the various sounding locations around the piano body, dampers, pedals, the room ambience, etc., that creates the whole listening experience. What the player hears and experiences up close can be different from what a listener in the audience can hear. In that case, you don't even have an engineering schematic and formulas to follow along with to create the algorithms.

In any case, I would agree that the resulting sound is the most important thing. And is it musical?

[Stokely]

Agreed, the analogy only works on a general level

What you have described also enters into the "amp in the room" wars (amps vs modellers) that you see on guitarist forums. Playing in a band with guitarists, I've seen a couple struggle with modellers (even analog ones like the sansamp)-- the sound may be identical but the experience is not.

Anyway, didn't mean to derail. I am fascinated by the idea of virtual circuit modelling. Hats off to these folks doing it, it's like some kind of voodoo to a person like myself who can barely plug in a toaster