ok, cool. i’ll elaborate…
one thing i would say is that there is no universal “quality” metric here. (IMO, marketing phrases like “studio grade” are pretty meaningless when applied to well-known algorithms. it’s like saying “high quality algebra.”)
all we can do is identify qualities that we find appealing, for whatever reason or situation, and seek or manifest those qualities. which sometimes discussion like this can help us do.
for example. you find “pools” more useful or inspiring than “blackhole.” the latter is a substantially more complex reverberation model, but that evidently doesn’t make it “higher quality.” it seems clear that your positive response to “pools” is about sound design choices, so that’s why i encourage the investigation of those choices, as a creative problem. (i don’t think you need to be a “coder” to look at a supercollider synthdef, which lists high-level DSP building blocks and doesn’t require you to understand how any of them work individually. but curiousity helps.) the model is modulated FDN + granular pitch shift + saturation and filtering. given that, we can think of other ways to recombine, reconfigure, and extend those elements that might also be (differently) inspiring.
supercollider and puredata are audio programming languages. your hunch seems correct to me: these environments can implement the same math operations that are used on embedded devices, typically with somewhat reduced efficiency but also with more computational capacity than most pedals, and they can produce the same effects. (we aren’t really living in an era where sample-conversion artifacts are ubiquitous and characterful; in fact we now are in the crazy position of working to emulate those “low quality” characteristics of yore.)
taking one example, a partitioned-convolution reverb in supercolllider is very simple to make, demo’d here:
https://sccode.org/1-5am
we can easily add a convolution reverb engine to norns, and we could probably gather a nice set of contributed impulse responses. in the 90’s, this would have been a standard of “quality” (viz., the depth/fidelity of the model) that was hardly imaginable in a portable form factor; now it is just another option we have. for example, you might want IR processing for early reflections, followed by FDN for longer tails, plus some filters and saturation. but architecture is one thing, and arriving at configurations that “work” is, again, an endless creative endeavor that is basically non-prescriptive.
i’ll put this another way: we can provide the technical components of a “high quality” reverb engine pretty easily on a platform like norns (i would start with something like described above: convoluation + FDN + nonlinearities.) but these engines have many, many parameters which interact in highly complex ways, and a big part of appeal of the classic digital reverb units derives from the large amount of effort and artistry that went into their preset and parameter design. IMHO an equivalent wealth of creative options will only be acheived, in this particular corner of open source music stuff, by pooling efforts together.
this leads into some other things that make your question (both forms) a little frustrating to me (no dis intended, just my feedback):
-
pragmatic: it’s not much to go on, describing something “like product X.” for example, the “speculuar tempus” has like 40 reverb and delay algorithms. many of them seem very simple to me, with few parameters or points of coloration (pingpong delay and the like,) while others are full of complex modulations and harder to analyze. but many of the features that add value have nothing to do with DSP per se: the preset selection, and UI/control components like “auto swell.” (i don’t know about that product’s tech stack or design approach; i do know that strymon devices like bluesky tend to use SHARCs and circuit emulation techniques.) so again if there is some specific quality or parameterization that appeals, it’s worth putting in the effort to identify what that is.
-
social: designing reverbs is partly science and largely art. most people who put a reverb on the market have devoted a great deal of engineering and creative effort into their work, which deserves respect and support if it is work that you appreciate, rely on, are inspired by &c. on the other side, you have hardware platforms like norns-shield and fates, and software platforms like pd and supercollider, that are largely volunteer-driven. it rubs me a bit wrong to expect volunteer coders to put in time to “clone” the work of our friends and mentors in the industry… i dunno… just something to keep in mind.
