Digital software replicating an analog circuit

[Tim C]

I was reading where a software developer claimed that "an audio plug-in can accurately "replicate" the process of the analog voltage as it travels through a circuit path".*

This wording to me seems very over reaching and I am curious if it's true?

I would agree with emulate,... but replicate?

 

[FvM]

Emulate, replicate, simulate, model... these are just words. In case of digital audio effects you'll ask for the subjective percepted quality. Today's digital effects, either dedicated devices or software plug-ins achieve outstanding performance and can be hardly distinguished from the original analog sound. It's mainly a matter of good design, the technicque is basically available. In so far "replicate" might hit the point.

Another question is which analog effects have been yet accurately modelled.

 

[Tim C]

I hear what you're saying with the semantics and agree that the quality of effects and dsp processing are emulated extremely well these days. Compression is one effect I still haven't heard convincingly emulated. There are good but not great emulations. Guitar amp sims can sound pretty good but still don't cut it when compared to the real thing. EQ's can be modeled very good and especially brick wall limiting is a digital process that is even better than it's analog counterpart because of the ability to look ahead.

Technically. I'm wondering how or if it's possible to write the code to replicate the process of an analog voltage as it travels through a circuit path?

 

[BradtheRad]

There is technical language and there is marketing language.

A fast enough computer and fast enough software is able to generate pleasing sound effects, just like analog circuitry can.

Link to a thread which brought up a similar question, namely whether (and how) software can emulate the distinctive Moog sound.

https://www.edaboard.com/threads/259953/

 

[FvM]

Technically. I'm wondering how or if it's possible to write the code to replicate the process of an analog voltage as it travels through a circuit path?

Digital circuit simulators can basically model the behaviour of analog circuits with an accuracy sufficient to make it sound identically. The computation effort won't be accpetable for real-time emulation, however. I presume that usual effect emulators are performing a considerably simplified functional simulation. This means that the emulation isn't a blunt copy of an existing model rather than an intelligent redesign. That's why artistic skill of the instrument/software designer plays a role and results can be quite different.

I also keep my previous comment that many historical analog effects probably never have been exactly modelled yet.

 

[Tim C]

Digital circuit simulators can basically model the behavior of analog circuits with an accuracy sufficient to make it sound identically. The computation effort won't be acceptable for real-time emulation, however. I presume that usual effect emulators are performing a considerably simplified functional simulation. This means that the emulation isn't a blunt copy of an existing model rather than an intelligent redesign. That's why artistic skill of the instrument/software designer plays a role and results can be quite different.

I also keep my previous comment that many historical analog effects probably never have been exactly modeled yet.

Thx. This is what I was getting at and pretty much how I've always understood it works.

When a manufacturer claims that they can replicate the process of the analog voltage as it travels through a circuit path. it just sounds like snake oil.. It's untrue or else many analog circuits would have been exactly modeled already.

I use the software and the analog counterpart daily with the highest quality ad/da conversion and monitoring and am aware of the non linear traits the the modeled software can't get right yet and have no problem accepting that, but when someone says they can get true analog sound through software by coping the voltage as it travels down the circuit through code, I'm calling BS.

Would still love to hear any other views related to the technical side...