HI Folks,
I posted a question over on one of the Sound on Sound forums but it's probably something that software developers might have an opinion on too.
https://www.soundonsound.com/forum/view ... 22#p600122
I'm primarily a musician and music producer but I've worked in software development in the past so don't hold back on the technicalities.
I would have thought that a Neve EQ would be something that would be relatively straightforward to replicate?
Any thoughts?
Cheers!
Hardware vs Software - Why is There a Clear Difference in this Case?
-
- KVRAF
- 16321 posts since 22 Nov, 2000 from Southern California
You are correct that the EQ curves themselves are very straightforward and easy to replicate with other EQ's, and this is what Eric Beam's statement "All Digital Parametric EQ’s are the same" caused an uproar about 10 years ago. What we've since figured out is that it's more about the circuitry and processing around the EQ's themselves that's causing the differences in sound. A Neve 1073 and a Yamaha PM1000 are both inductor-based EQ's but a PM1000 does not sound the same as a 1073, even when you install 1073 inductors into it. For that matter, a modern 1073 does not sound the same as a vintage 1073.
Perhaps an even better illustration of this is a Pultec. That's a passive EQ, meaning the EQ's themselves should be about as natural and color-less as you can get, and yet it is full of color. Thus, if someone wants to emulate a Pultec, the easy job is emulating the EQ curves, the real work and accomplish would be in accurately emulating the tubes and Peerless transformers that make up the rest of it.
Perhaps an even better illustration of this is a Pultec. That's a passive EQ, meaning the EQ's themselves should be about as natural and color-less as you can get, and yet it is full of color. Thus, if someone wants to emulate a Pultec, the easy job is emulating the EQ curves, the real work and accomplish would be in accurately emulating the tubes and Peerless transformers that make up the rest of it.
-
- KVRist
- 323 posts since 19 Jul, 2008
Time-domain models exist for nearly all components in vintage and modern EQ circuits, with a nice selection of models on the spectrum of accurate/complex to inaccurate/simple. These models are also important for industries outside of audio, so IMO circuit simulation is a mostly-solved problem.
The problems encountered by audio DSP developers are usually nonrelated to the possibility of achieving good circuit emulations. They're usually along the lines of "Is it worth spending weeks on circuit research? Is it worth the CPU? Do customers even care about accuracy to the modeled hardware? If so, does it actually benefit profit? Is it worth buying a Mathematica etc license? Is staring at a system of 80 nonlinear PDEs really what I want to be doing all day, or do I want to whip something up that makes me happy?"
The problems encountered by audio DSP developers are usually nonrelated to the possibility of achieving good circuit emulations. They're usually along the lines of "Is it worth spending weeks on circuit research? Is it worth the CPU? Do customers even care about accuracy to the modeled hardware? If so, does it actually benefit profit? Is it worth buying a Mathematica etc license? Is staring at a system of 80 nonlinear PDEs really what I want to be doing all day, or do I want to whip something up that makes me happy?"
VCV Rack, the Eurorack simulator
