Sorry guys, but… Can anyone explain briefly how a convolution reverb works? Is something like sampling a space?
There’s a decent 30-second visual explanation of how convolution works, in one of the Altiverb 7 demo videos (1:26 minutes in)
So for every step in a waveform, you add a “tail” of sound, and this is added together over time. If you watch at 7:30 in the video above, you get some examples of how you can use non-reverb impulse responses to convolve audio more in a sound-design manner.
You made some very good points pichenettes, about the non-tweakability of convolution effects. Another popular sound design method that suffers a bit from this, is spectral synthesis, as you can make some really interesting sounds that are non-typical for any eurorack, but once you try to modulate some parameters quickly (especially pitch with wide spectrums), the illusion starts to fall apart. Like demonstrated in this video of GRM Evolution (1:40 minutes in)
Thanks for the links
Good point about the modulations. A drawback of frequency domain effects (FFT
> spectral modification> IFFT) - be it convolution or the kind of spectral shaping done by the GRM tools (contrast, warp…) is that parameters need to be constant during the processing of a window. So parameters are “quantized” at the same rate as the window rate. And some effects need loooong windows to offer a good spectral resolution. So in the end, they are fine for “dial a setting with knobs / load a preset, then let the audio run through it” ; but not great for modulating things with CVs or just “playing” things with controllers. Increasing the overlap rate is computationally costly and is kind of a shitty workaround - it just smears everything. It’s no worse than sending the output in a big reverb to diffuse steps
I have a shitload of abject frequency domain processing code running on STM32F4s. Enough material for stupid Easter Eggs in half a dozen modules
I’m still figuring out if you’re one of those extremely rare insanely productive genius coders, or if you’re just an “average” productive developer who has managed to sidestep from a boring reliable career in enterprise software development into a field where most people writing code have no idea what they’ve been doing.
OK, so now we know that NM1 does (or did) frequency domain processing using FFTs. And that frequency domain processing in a modular context is good in theory. And that the gap between theory and practice is small in theory but large in practice.
the truth is he is both…insane genius productive and managed to step into a field of clueless coders (which clearly includes me as head of cluelessness…)
those kind of programmers/coders/engineers/builders are kind of my heroes. I’m the kind of guy who stuides Literature at university but secretly wishes to be good at maths to do this kind of stuff
I see Olivier more as an artist. One of the rare who succeeds in making a living from his art.
yep, that’s what I mean by “those kind of…”! A kind of art that allows people to produce art from it!
IMHO he’s a designer first and foremost. Calling him an artist stretches the definition of “art” beyond its breaking point.
imo there is no breaking point for that… It stretches endlessly to where it becomes really thin. But i don’t think Olivier is at a thin point. If Van Gogh would have lived now he wouldn’t be putting oil mixed with pigments in a stretched piece of canvas.
But this is an endless discussion worthy of a new thread full of capybara pictures.
I wouldn’t call my modules “works of art” insofar as their design is subject to the constraint that there’ll have to be at least 250 people to buy the result. I’m always reminding myself not to be too crazy when I work on them.
But I have sometimes toyed with the idea of making more thought-provoking modules “for art’s sake”, in the same vein as what Gieskes sometimes does - just didn’t have the time for that.
Gieskes is a good example of contemporary art taking the form of Eurorack modules. I really don’t see in which way the Laserloper would be in a different category from works of bending/repurposing widely accepted as contemporary art (say Cory Arcangel’s Super Mario Clouds).
I’m seeing a lot of parallels; both you and him are running what’s essentially a one-man operation while at first sight it looks like your business is bigger than that (you both have a very nice website, for example). Also, both of you collaborate with a lot of different people to put out your products.
What’s still different is that Peter has managed to build a portfolio that contains products that are clearly commercially viable as well as products that are primarily experimental and unlikely to ever result in a proper return on investment (e.g. Karloff, Julien, and Elementar). Of course this is easier to scale for digital downloadable items such as fonts, and less viable for physical products such as Eurorack modules.
Still, you might want to consider raising prices on your “commercial viable” products such as Braids to fund the experiments you’d really want to make real. Go for it!
I think the whole mutable instruments as a whole with it’s philosophy is work of art rather than the indivudual products.
I don’t think it’s not art because it has to sell.
But I believe most definitions of art require the creator to define his work as art so maybe it’s not.
Formally, you can define pretty much anything as art. More restrictively, I personally believe in the “art as the individual expression of emotion” definition.
More specifically, I’m trying to point out it is in fact possible to build a single catalog of products that contains those that are obviously commercially viable and those that are strictly experimental and therefore clearly not. It seems like this is what would provide Olivier with the most happyness.