Reverb

[Aleksey Vaneev]

I've released CRTIV Reverb today, it uses the aforementioned Hadamard FDNs. Except that due to true stereo processing the total number of FDNs per channel is 32. Takes 3.5% i7700k CPU . I know it's not absolutely perfect in early reflections part, some discrete elements can be heard on transients, but currently it is as good as I could get it.

http://www.voxengo.com/product/crtivreverb/

[xoxos]

what a treat! though it will be a while before i do any more music.

Hay, doesn't Stevie Nicks live in Arizona? I bet that Crystal Visions Road is named after a lyrics line in her Dreams song.

my sister spent a weekend on her ranch in the 80's. it has f**k all to do with dsp.

[karrikuh]

I'm using 4x4 Hadamard matrix FDNs, with absolutely all delay line lengths at random in the range 5 to 100ms. I run 16 of them in parallel with the same input signal. So there is a total of 16*4 delay lines of random lengths. The number of FDNs can be reduced I think, but even at that number the CPU load is quite low. Used random delay length range can be reduced e.g. to 30ms, but this will produce a "plate reverb" character.

Thanks for the valuable info.

[nonnaci]

Anyone have experience with modulating the feedback matrix of FDNs?

https://www.audiolabs-erlangen.de/conte ... ht2015.pdf
seems give a general case of time-varying rotation transforms by modulating the exponential of the eigenvalues. Are there plugins and/or sound samples of anyone doing something similar?

[Christian Schüler]

Yin and yang, echo density and mode density.
Try to improve one, and the other one goes south...

Aside from it sounding metallic, it has horrible, horrible wolf notes and tends to sound twice as loud on one channel compared to the other. None of this behavior is random, either, which might actually be charming if it was.

That description sounds very much like it's lacking mode density. So you need to improve that by increasing the order of the system (increase the delay lengths). Which means ... echo density is going to suffer unless you also add more scattering connections.

[Richard_Synapse]

Anyone have experience with modulating the feedback matrix of FDNs?

https://www.audiolabs-erlangen.de/conte ... ht2015.pdf
seems give a general case of time-varying rotation transforms by modulating the exponential of the eigenvalues. Are there plugins and/or sound samples of anyone doing something similar?

I've tried rotation transforms some time ago but didn't like it. That doesn't really mean much though, my experiments were not based on the above paper & it's always possible that there's some sweet spot, specific matrix, some trick or whatever to make this sound good.

Richard

[syntonica]

Omnidirectional sound radiation is a special case, there's not many things that naturally do this.

Being in the center of a sphere would do it. It sounds like an uninteresting case, of course, depending upon your physical shape and damping qualities.

[Christian Schüler]

Anyone have experience with modulating the feedback matrix of FDNs?

https://www.audiolabs-erlangen.de/conte ... ht2015.pdf
seems give a general case of time-varying rotation transforms by modulating the exponential of the eigenvalues. Are there plugins and/or sound samples of anyone doing something similar?

You can do this quite literally in AriesVerb. That's what the angular controls are for. I have a preset (I think it is calls "Formant filter" in the filters category) that has a very short delays, and by changing the rotation angle of the feedback matrix you can move the relative locations of the resonant peaks, thereby changing the "formant" from AH over OH to IY. So its basically simulating the reverberant structure of the laryngeal cavity

[nonnaci]

Anyone have experience with modulating the feedback matrix of FDNs?

https://www.audiolabs-erlangen.de/conte ... ht2015.pdf
seems give a general case of time-varying rotation transforms by modulating the exponential of the eigenvalues. Are there plugins and/or sound samples of anyone doing something similar?

You can do this quite literally in AriesVerb. That's what the angular controls are for. I have a preset (I think it is calls "Formant filter" in the filters category) that has a very short delays, and by changing the rotation angle of the feedback matrix you can move the relative locations of the resonant peaks, thereby changing the "formant" from AH over OH to IY. So its basically simulating the reverberant structure of the laryngeal cavity

Did the choice of orthogonal matrices produce any interesting effects? e.g. choice of Givens rotations, householder reflector planes. The paper aimed at asynchronous randomized movement of poles which in a way smears the modal density and thus smoothing out resonances but I'm sure there could be ways of moving the resonances in a guided fashion.

[Christian Schüler]

As of now, my rotation angles are static, not modulated. I didn't try that yet, but it would be trivial to add. However my guess is that it would result sonically in a similar effect as the "inverter modulation" that is already implemented. The inverter in AriesVerb is a one-pole frequency dependent phase shift that sits with the other filters in the feedback loop, and you can modulate its corner frequency. This lets you shift the resonant peaks by up to 1/2 the peak-to-peak distance. This creates a modulation that feels very differently from delay line modulation because the effect does not scale with frequency. Like the difference between PM and FM. Something *very* similar should happen when dynamically rotating the feedback matrix.