The difference between Moog emulations and algorithmic reverbs:Numanoid wrote:But we usually take for granted that plugins become "better" as they are newer, due to them for example being better tailored toward usage on newer computers with more power.valhallasound wrote:Algorithmic reverbs don't necessarily become obsolete with the passage of time.
I mean, the Moog emulations keep coming, but why would anybody need another Moog emulation, when there are already so many about being emulated from the same hardware
- Moogs are highly nonlinear analog circuits. Lots of distortion, and feedback in lots of places in the circuit that moves fairly close to the speed of light. The newer plugin models are taking advantage of the increase in computer power to use more oversampling, as well as "zero feedback" filter algorithms that will get better and better with the number of cycles you throw a them. It is amazing how the original Minimoog filter/VCA used a few dozen transistors, while modeling this sound with any degree of realism requires using tens of millions of transistors in your CPU.
- Algorithmic reverbs are designed from the ground up to be digital. The same techniques used in the 1970s and 1980s will work on today's computers. There might be differences in the sound between a fixed point processor used in older hardware, and the floating point processing in modern CPUs, but this makes less of a difference when porting a 1990s algorithm that worked on fairly high precision fixed point hardware.
Can algorithmic reverbs benefit from throwing more cycles at things? Sure. You can use better delay interpolation, more allpasses, more delay taps, modulate more things, and so on. However, for many older topologies, there is a sweet spot for a certain complexity. Adding more allpass delays to the Blackhole algorithm won't necessarily make it a BETTER reverb, so much as a DIFFERENT reverb. Adding output taps to a Lexicon-type allpass loop will increase the initial echo density, but it won't necessarily make it a better sounding reverb overall - it just changes the sound to something different.
ValhallaShimmer can be used as an example of this. The Big Stereo mode can be used to cascade a whole bunch of instances in series. If the Mix is set to 100% wet, and Feedback is set to 0%, then cascading N Shimmer instances is identical to a hardwired algorithm that uses N times as many allpass delays. This produces cool sounds that take forever to fade in and fade out. Which is great, if you don't mind waiting around a few seconds for things to start coming out of a reverb. For most musical uses, a single instance of ValhallaShimmer is about as "slow" as you'd want a reverb to be.
A short summary of the above: the latest generation of virtual analog algorithms are designed from the ground up to work better with more resources. The majority of digital reverb algorithms are designed with a fixed amount of resources in mind, and will have a particular sound, that doesn't necessarily improve by throwing more cycles at it. Throw more cycles at Blackhole or ValhallaShimmer, and you'll get a sound that isn't the Blackhole or ValhallaShimmer sound that people are expecting to hear.
Sean Costello