Is a convolution module possible?

[10 Miles High]

Tats developed for Red Bull Academy this thing:

https://www.electronicbeats.net/the-fee ... ike-never/


Granular, but he says convolution also.


10 Mh

[wavejockey]

Hi guys,


Anyway.. the biggest design choice is now to go for:

A. full sized version with proper sized display, proper large pots with knobs on them. 500 EUR

B. uVersion, with mini display, naked pots. 350 EUR?

A; full size please

take a look at knobs & OLED in Rossum Electro Morpheus & Assimil8tor

what about loading IRs ? (instead of creating your own)

[earx]

what about loading IRs ? (instead of creating your own)

Oh that's already possible. A basic function of any convolution response. Just plug in USB/SD and scroll through them using the rotary encoder..

[NS4W]

What parameters would be actively tweaked? If its a matter of setup and then just playing I dont mind a small interface. Modules made for wiggling are always nice if they are on the roomy side

[earx]

The Rossum modules look ace btw! Thanks for the tip!

[earx]

What parameters would be actively tweaked? If its a matter of setup and then just playing I dont mind a small interface. Modules made for wiggling are always nice if they are on the roomy side

Right now we have the following knobs:

* predelay 0..2000 ms

* envelope attack

* envelope release

* dry

* wet

* inputgain .. this is now digital, but analog would be much better (right?). The idea is to accept both doepfer and line levels.

* stereo width (mid-side)

The following digital controls:

4 buttonleds: Record, Store, Audition (play response only), and one more for which have no idea yet.

1 rotary encoder to select responses from USB/SD/internal. Push it to scroll faster.

And the following CV's:

- HP cutoff

- LP cutoff

- Stereo width

I chose these because in a block convolver it takes time to recompute the response if you apply envelopes, predelay, etc. Filter and mid-side settings are immediate.

[Hooft]

FULL SIZE please.

It's all about being able to play/perform with.

Related: E-RM Grain Convolver, in depth:

[video][/video]

[cackland]

Why not have a single control for dry / wet? Also, I would consider having cv over this parameter.

Also, what kind of HP are you thinking for each version? And would the micro version have the same control parameters but with the 9mm song huei potentiometers?

[earx]

Dry/Wet on CV sounds like a plan actually! Low pass is probably not useful anyway on a reverb. High pass much more so!

Check out our latest news for a pic of the unit including all the controls. We went for naked Alps pots now. They feel good, but well.. mini is mini. https://www.tastychips.nl/?p=3526

I'm really glad most people here prefer full size!

Thanks for the granular convolver. I actually heard that it was partly inspired by our GR-1. A great thing to be noticed by the Korg veteran designer in this way!

[cackland]

There is a lot of extra room on that control board.... I would definitely urge to go full size pots, with an improved layout. If this is the 'mini' version, I would be concerned about the regular sized version.

Assuming there will be a secondary component PCB, or will all the components be on the bottom layer?

[earx]

There is a lot of extra room on that control board.... I would definitely urge to go full size pots, with an improved layout. If this is the 'mini' version, I would be concerned about the regular sized version.

Assuming there will be a secondary component PCB, or will all the components be on the bottom layer?

Oh, for sure.. Yeah. This prototype is a single PCB, but already working on a sandwich version. Most definitely will place proper knobs on there.

[jorg]

I'm not interested in digital implementations, but am now quite intrigued about the possibility of an analog approach to convolution filtering.

Well, I suppose you could call all of our analog filters and delays "convolution" but they don't do the sort of general-purpose thing the OP is talking about. With digital convolution, you can model real spaces with uncanny realism. It is pretty costly in CPU, but even little ARM controllers these days are fairly amazing, and might be able to do some limited convolutions. The most important piece is a good FFT.

[wsy]

There's really several ways to go about this problem, and the best way is different depending on the inputs.

If one or the other inputs is basically an impulse (or a few impulses, at various times) then
the brute force implementation is often faster. Yeah, FFT is O(N log N) but if one input is
only four impulses or something small like that, then log-1 N is sixteen and that isn't a big
window at all. And that's brute force.

If most of the values are exactly zero, then the "brute force" implementation, operating
on a run-length encoded table, is O(N).

Or if all of the values in one of the tables is constant (or slowly varying) then again, it's O(N) [[ run double buffering and crossfadel]].

And now ya got me thinking about this....

[wavejockey]

* dry

* wet

* inputgain .. this is now digital, but analog would be much better (right?). The idea is to accept both doepfer and line levels.

The following digital controls:

4 buttonleds: Record, Store, Audition (play response only), and one more for which have no idea yet.

1 rotary encoder to select responses from USB/SD/internal. Push it to scroll faster.

And the following CV's:

- HP cutoff

- Stereo width

.

if going analog for the input gain means bringing in artifacts and/or noise i would not vouch for it
(if one would like analog gain, that they'll put in a valve before then)
knob 4: shift?
a cv input that can be assigned arbitrary? (like on my ensoniq DP/4)

[Paranormal Patroler]

+1 on the smaller version. If IR and convolution is your thing at the studio there are full fledged options outside the modular. If you're a performing artist that needs convolution (like I am) then a smaller version would be more appealing for a live oriented case.

Just my 2 cents.

[wavejockey]

a) what other hardware versions ?

[Paranormal Patroler]

a) what other hardware versions ?

I didn't claim hardware, but I'll humor you none the less. Logidy EPSi is one. Mungo module C0 is another. Software options (as per my studio reference) are abound.

[mcnys]

i would like a smaller version tbh, and here is why: a convolution reverb is more of a set it and leave it kind of device, not very performable module so smaller footprint makes more sense for me (also cheaper price hehe). Other than that i'm hyped for whatever version it will be in the end.

*puts 50 euro in to the piggy bank*

[Paranormal Patroler]

By the way, IR are really nice for percussion purposes. So a smaller module would also make more sense if want to buy a couple to enhance the sounds of your percussion.

[jorg]

...If one or the other inputs is basically an impulse (or a few impulses, at various times) then the brute force implementation is often faster...

Yes but that's not a very general case. OK as a percussive or plucked sound source, but not practical as an effect.

[earx]

By the way, IR are really nice for percussion purposes. So a smaller module would also make more sense if want to buy a couple to enhance the sounds of your percussion.

Ok, noted. But you can already use 2 separate ins and outs on this module. It would probably be possible to have 4 ins and outs as well..

[Buyakasoundman]

By the way, IR are really nice for percussion purposes. So a smaller module would also make more sense if want to buy a couple to enhance the sounds of your percussion.

Ok, noted. But you can already use 2 separate ins and outs on this module. It would probably be possible to have 4 ins and outs as well..

YES YES this!!! 4 ins and outs please

[wavejockey]

now we are on DP/4, DPSv55/77 & Lexicon territory

if (enough) routing possibilities are available, this is a no-brainer

[earx]

Current revision:



As you can see we went for full sized. Otherwise the user interface would suffer.

This baby can now:

- Record and store responses to USB stick
- "Scan" the envelope over the response, moving it around from early reflections up to the tail. using the pot directly under the display
- And last but not least it can crossfade between two responses in real time. This can also be done using CV.

Another improvement:
It can accept stereo jack input directly from your laptop or phone, or use two separate mono jacks.

Now looking at options to pitchbend the input and output, to keep the response consonant with the input.

The main issue here is the eurorack form factor. I'd actually want to add a crossfader, and a pitch knob.. but where to leave it all?

[brandonlogic]

Current revision:



As you can see we went for full sized. Otherwise the user interface would suffer.

This baby can now:

- Record and store responses to USB stick
- "Scan" the envelope over the response, moving it around from early reflections up to the tail. using the pot directly under the display
- And last but not least it can crossfade between two responses in real time. This can also be done using CV.

Another improvement:
It can accept stereo jack input directly from your laptop or phone, or use two separate mono jacks.

Now looking at options to pitchbend the input and output, to keep the response consonant with the input.

The main issue here is the eurorack form factor. I'd actually want to add a crossfader, and a pitch knob.. but where to leave it all?

It looks like theres space for a new more knobs on the left side?