The RV-5 Analysis Thread

Well...


It is a start.
This is a custom algorithm I started working on btw. It uses 8 early reflection stages, 6 input diffusion delays, and a Hadamard feedback delay matrix of 4x4. The delay buffers are read using a Cubic interpolation. Linear interpolation sounded okay until I applied modulation, then it all went to shit.

Now... Here are some single sample dirac impulse plots of the RV-5:
1757146113272.jpeg


These are with the time knob at positions - 7 (minimum), 8, 9, 10, 11, 12, 1, 2, 3, 4, and 5 o'clock. You can clearly see that the time decay has no bearing on the initial set of impulses used to get the reverb going.

If we zoom in a little:
1757146205190.png


You can see the single sample impulse at the top. You can also see the initial response of the algorithm. Now how do you turn a single sample 'click' into a smeared rollercoaster shape like that? You use all pass filters.
 
Okay.... phone the horses and put them on hold!! Here's a thing...

I was obsessing over trying to dial in my all pass filter to give me the same kind of result that I see at the very first tap on the RV-5. Which is this:

1757152152328.png


But in my algorithm... everything I kept doing sounded alright. I'm totally able to get one or more all-pass filters to perform some smearing. but I wasn't getting that kind of shape. I was getting stuff more like this:
1757152254106.png


Or this:
1757152283506.png


Or even this:
1757152311147.png



And try as I did, I couldn't get my all-pass to match the RV-5. So I started thinking.... what if it isn't just the RV-5.

So I took the RV-5 out of the path, and just did UFXII -> Reamp Box -> UFXII. I got this:
1757152550124.jpeg


Then I tried UFXII -> UFXII, and I got this:
1757152568560.jpeg


In short... I am pretty sure that the anti-aliasing filter of my converters in the RME UFXII is causing this.

If I change my sample-rate from 48k to 88.2k, I get a much different pattern:
1757152724064.png


And for shits and giggles, 192khz sample-rate:
1757152770316.png


So nailing the pre-smearing characteristics of the RV-5 is going to be quite tricky. Because I currently cannot isolate it well enough from the rest of my system. This is where a VST plugin of the Boss RV-5 would come in handy!!

Oh, and not only do we have these anti-aliasing filters in the audio interface to contend with, but it is entirely possible we have some in the pedal as well. Throw noise-floor and ground or mains hum noise concerns in there too, and it means any attempt to capture and digitally reproduce the same all-pass timings and gain factors, is pretty much impossible.

.... maybe.
Fun times!
 
Very interesting and cool thread!

Have you ever checked out the Boss RV-70 rackmount unit? I believe it's equivalent / parallel to the RV-3 era tech, and no modulation, but it has a great MIDI implementation that lets you adjust a lot of parameters not normally accessible, which might be revealing. They can sometimes be found for pretty cheap. Sounds cool too (or maybe I just miss my RV-3 lol)
 
Back
Top