Do profilers or modeler exists for analog synths? Why not have a Kemper for that Moog or Oberheim?

joegrant413

Hope the subject explains my question. The guitar amp world has had great modelers like Kemper for a relatively long time.

How about in the synth world? I'm not talking about sampling, or buying a very flexible synth that can come close to many vintage sounds. I'm talking about something that captures and models any other synth on-demand across a spectrum of dynamics and pitches. And maybe even a few knobs, including the pitch and mod wheel.

Does it exist? Why not?

Thx

jebni

Isn’t that an iPad with a bunch of virtual analogue synth apps on it? I’m not sure I understand the question.

wim

It's going to be interesting to see how / if people step up to answer this. I tried for quite some time to come up with an explanation of the difference between modeling the effect of an amp on sound traveling through it, and the idea of "capturing" a synth that generates it's own sounds using the almost infinite combination of interacting components in order to model it.

It's one of those things that is quite clear in my head is impossible or massively impractical. But how to explain it?

I do hope some people step up. 👍🏼

Tarekith

I think it's not need as much as it is for guitars because there's already so many companies out there modeling famous synths and effects already.

ecou

What a Kemper does is model the amp, guitar cabinet/speaker captured thru a microphone. The guitar being the sound generator.

How does it do it in very simple terms? It take the wave captured
by the microphone minus the wave created by the guitar. The resulting sound is a impulse that is applied to the original sound.

In the synth world we only care about the generator, filter and amp. The Speaker/ microphone is not part of the sound. In the guitar case the speaker and mic are a very big part of the sound.

Ik multimedia Synthtronik is the closest we have to a Kemper. They start with samples and have a virtual filter amp section .

ecou

@wim said:
It's going to be interesting to see how / if people step up to answer this. I tried for quite some time to come up with an explanation of the difference between modeling the effect of an amp on sound traveling through it, and the idea of "capturing" a synth that generates it's own sounds using the almost infinite combination of interacting components in order to model it.

It's one of those things that is quite clear in my head is impossible or massively impractical. But how to explain it?

I do hope some people step up. 👍🏼

I hope It made sense. I re wrote this a half dozen times.

wim

@ecou said:

@wim said:
It's going to be interesting to see how / if people step up to answer this. I tried for quite some time to come up with an explanation of the difference between modeling the effect of an amp on sound traveling through it, and the idea of "capturing" a synth that generates it's own sounds using the almost infinite combination of interacting components in order to model it.

It's one of those things that is quite clear in my head is impossible or massively impractical. But how to explain it?

I do hope some people step up. 👍🏼

I hope It made sense. I re wrote this a half dozen times.

Better than my half dozen times that I rejected. 😂

jebni

Oh, I didn’t get the “capture on demand” aspect of the question. What @ecou said, I guess.

It does beg the question of what the ontological difference between generating and treating a sound actually is, though. So besides unpacking what synthesis actually does, how well do different amp sim approaches work? And the difference between modelling and profiling/capturing an amp? And what something like the AI business that Tonex uses brings to the table?

pedro

Guitar is more about mixing acoustic with a tone, it makes sense to have separate parts
With synths you can get that tone right at the sound generation stage
I love both approaches

espiegel123

@ecou said:
What a Kemper does is model the amp, guitar cabinet/speaker captured thru a microphone. The guitar being the sound generator.

How does it do it in very simple terms? It take the wave captured
by the microphone minus the wave created by the guitar. The resulting sound is a impulse that is applied to the original sound.

In the synth world we only care about the generator, filter and amp. The Speaker/ microphone is not part of the sound. In the guitar case the speaker and mic are a very big part of the sound.

Ik multimedia Synthtronik is the closest we have to a Kemper. They start with samples and have a virtual filter amp section .

It is not that simple. The capturing of the distortion and non-linearities are not simple impulse responses like one does for cabinets or reverbs or eqs. They may capture impulse responses or sine sweeps as part of the process to build the model...but it isn't like IR capture and playback.

Some profiling (such as with TONEX and THU rigs) capture the response of the amp with one set of dial settings. They model the performance of the amp handling a various levels of signal with those settings. I am not sure to what extent Kemper models the knob response ... which adds a lot more complexity to the process and makes things more processor intensive.

Guitar amps have way fewer controls than synths. The combinatorials are staggering...which is why they don't use profiling techniques. Some virtual synths are actually simulating the actual circuits.

Telefunky

@ecou said:
What a Kemper does is model the amp, guitar cabinet/speaker captured thru a microphone. The guitar being the sound generator.

How does it do it in very simple terms? It take the wave captured
by the microphone minus the wave created by the guitar. The resulting sound is a impulse that is applied to the original sound.

Sorry, that‘s a bit over simplified...
The Kemper method sends a defined measuring signal through amp and cabinet.
This captured signal is compared to the original and analyzed to derive about 50 parameters of a virtual amp sim.
Kemper himself once stated that this „profiling“ is not entirely IR/convolution based.

But it‘s a different process from „capturing“ a virtual analog synth:
the amp is profiled in only 1 state (the current knob position on the amp) precisely.
Opposed to this a synth is in motion (so to say), knobs are tweaked permanently.

As mentioned above it‘s way easier to simply model the synth‘s parts with enough precision to get a convincing replica.

jebni

@espiegel123 This is sort of the point I was trying to make about the different approaches amp sims take. In my mind, the more authentic and responsive we need it to be, the more they’d need to actually emulate the circuit, it would seem. Kemper have this Liquid Profiling thing (“you can turn the knobs!”) that I assume adds an actual modelling aspect to their approach… right?

So in the end, even at this very low level of complexity compared to all the variables in a synth, if we were wanting to fully grasp its glory, we’d still “want” to approach an amp like a virtual analogue synth and a not a capture. Correct?

ecou

@wim said:

@ecou said:

@wim said:
It's going to be interesting to see how / if people step up to answer this. I tried for quite some time to come up with an explanation of the difference between modeling the effect of an amp on sound traveling through it, and the idea of "capturing" a synth that generates it's own sounds using the almost infinite combination of interacting components in order to model it.

It's one of those things that is quite clear in my head is impossible or massively impractical. But how to explain it?

I do hope some people step up. 👍🏼

I hope It made sense. I re wrote this a half dozen times.

Better than my half dozen times that I rejected. 😂

👍

michael_m

For iPad I don’t think so, but if I understand the “on-demand” aspect correctly, then perhaps Synplant from Sonic Charge fits the bill.

wim

@Telefunky said:
As mentioned above it‘s way easier to simply model the synth‘s parts with enough precision to get a convincing replica.

That's pretty much it in a nutshell.

jebni

What would be a good demonstration of the limits of “capturing”? Tonex allows you to capture a pedal. So, can it capture a Hologram Microcosm? 🥳🥳🥳

wim

I've been struggling to come up with an analogy. Here goes nothin' ...

The effect of an amp is like the effect of running water through a coffee filter filled with your favorite coffee grounds. An amp simulator is like adding instant coffee mix to water. Someone analyzed the taste of real coffee and came up with a mixture of real and artificial ingredients to model it. Thankfully in this case the result can be a lot more convincing than instant coffee. 😬

Capturing and modeling an arbitrary synth would be like saying "Make me an instant turkey dinner" out of artificial ingredients, then the next day saying "Now lets make some instant turkey sandwiches" and the next "Gee I sure wish I had some instant turkey and cheese omelette powder". Sure, there are a few common components that could be modeled, such as instant turkey loaf, but the variations of how they're used is almost infinite.

Now, imagine someone asks for instant breakfast, lunch, and dinner mixes that taste like the ones served at a particular restaurant. That's the challenge the OP question seems the most like to me.

I think I have too much time on my hands this morning. 😂

espiegel123

@jebni said:
What would be a good demonstration of the limits of “capturing”? Tonex allows you to capture a pedal. So, can it capture a Hologram Microcosm? 🥳🥳🥳

Tonex is an example of profiling that captures all knobs at one setting. You need a separate capture for each knob setting. When in models a pedal, it is only modeling the sound with the dials in one position.

GuitarML (an open source profiling system) allows one knob to be dynamically modeled. So, you might profile an amp or distortion pedal with all but one knob in fixed positions and profile the impact of one knob through its whole range.

joegrant413

Thx for all the thoughtful replies!

I forgot about Synplant and haven't tried it yet.

McD

A synthesizer is a model of a physical instrument in many ways.

It starts with something that vibrates/oscillates:

• a string under tension
• a column of air in a pipe or tube
• a skin stretched across a hoop

Then there’s typically a physical enclosure that reinforces resonance of that oscillation:
a wooden body place beneath a point of vibration to capture the vibration and feed it back to the oscillating media

The interaction between the oscialltor and its physical design creates a filtering or amyplifying of specific overtones (multiples of the base vibrating frequency).

All these physical attributes are emulated in analog synths and modeled in digital synths.

Any of these synths can be input into electric guitar style stomp boxes and amps to add typically non-physical effects like
distortion, echo, reverb and chorus/phase/flange conversions.

I think the modeling of digital synths is pretty close to an answer for your request. Tera Pro with all it’s potential add-ons can
re-create most physical synths for you.

BroCoast

@jebni said:
What would be a good demonstration of the limits of “capturing”? Tonex allows you to capture a pedal. So, can it capture a Hologram Microcosm? 🥳🥳🥳

Tonex can only capture drive pedals, not something like that.

It’s reasonably accurate for something like a fuzz pedal, i.e if it has a gated character it will capture that.