MeldaProduction

Alphacodex · Post by **Alphacodex** » Sat Mar 18, 2017 9:30 pm

Just a quick question. From watching the YT tutorials I learned, that with MMultiBandConvolution I can add an instance before and one after an effect to capture its impulse response.
However when I have a VSTi with no audio input, and want to capture its internal effect, it can't be done with MMultiBandConvolution, right?
I'd need something like Voxengo Deconvolver...?

Cheers,
Codex

jmg8 · Post by **jmg8** » Sat Mar 18, 2017 9:34 pm

There is a test tone in the advanced settings. Use that on the first instance (before) and use record on the second (after)

Thavma · Post by **Thavma** » Sat Mar 18, 2017 9:39 pm

Oh my God! What crazy thing! Melda are have Alien technology!

Sorry don't know how to help.
But could you send me that video with that capturing technique you say? Would like to see it.

Thanks

Alphacodex · Post by **Alphacodex** » Sat Mar 18, 2017 10:20 pm

Just a standard tutorial from Melda:

https://m.youtube.com/watch?v=cg8RdeB-dwI

Starting at 3:25 the process is briefly shown.

@jmg8
As I wrote I want to capture the internal effect of a VSTi (instrument), so there is no before and after. Also there is no external input.
All I can do is, render an internal click noise or a sine sweep first dry and then with the effect...

jmg8 · Post by **jmg8** » Sat Mar 18, 2017 10:32 pm

Which instrument? Some come with an FX version too.

Thavma · Post by **Thavma** » Sat Mar 18, 2017 11:12 pm

thanks! wow!

Chandlerhimself · Post by **Chandlerhimself** » Sun Mar 19, 2017 12:00 am

You can't capture the sound of an instrument because there is no input. Unless the vsti has an input it's impossible. Also unfortunately capturing modulation isn't really possible, although maybe you can capture multiple IRS and switch between them to fake it.

Alphacodex · Post by **Alphacodex** » Sun Mar 19, 2017 6:44 am

It's the ArtsAcoustic reverb built into Nexus. Someone in another thread told my it should work with Voxengo Deconvolver following those steps:

1) Create a testtone within nexus (e.g. a shot sample, or short white noise burst or a sine-sweep)
2) Render the test tone dry (switch of the internal effect)
3) Render the test tone with effect on
4) Enter both dry and wet wav-file into the deconvolver

I haven't tried it though, first wanted to check, whether I could do it with my melda tools first.

jmg8 · Post by **jmg8** » Sun Mar 19, 2017 8:26 am

You could record Meldas test tone and load that into Nexus.
Then use MConvolution after it to make the IR.

Alphacodex · Post by **Alphacodex** » Sun Mar 19, 2017 2:16 pm

Unfortunately I can't load anything into nexus, I can only work woth the material that is already in nexus. All basic waveshapes are there and I can program modulations, so a sinus-sweep is possible or a short burst of white noise, but no external sounds...

jmg8 · Post by **jmg8** » Sun Mar 19, 2017 2:24 pm

Well if I remember correctly. Deconvolution is not what you need. I does the opposite of make an IR. It decodes IRs.
If you make a short white noise burst and record it dry and then again wet.
Then phase invert the dry to remove it from the wet, leaving just the tail = IR.
However, I think the burst is supposed to be exactly the same both times so that it can phase cancel correctly and white noise is random!
Maybe there is a way you can create a burst that will be the same every time?

steve2KVR · Post by **steve2KVR** » Sun Mar 19, 2017 2:33 pm

...possibly...record say a one-second burst or so...then edit the file such that you have the length you need with zero crossing points at the start and end of the selection, save...I've used something like this for other "artistic" purposes...hth.../s~

Alphacodex · Post by **Alphacodex** » Sun Mar 19, 2017 4:27 pm

Hmmm... I could try that, as nexus is a sample player it could be that white noise is also a sample, thus always exactly the same.

If I could find a very short (only some samples) "click" sample and just cut that away, would that also work?

Cheers,
Codex

jmg8 · Post by **jmg8** » Sun Mar 19, 2017 4:42 pm

Not sure, I think it has to be white noise as it contains all frequencies at equal energy. If you use something else I think it would make the IR sound filtered. Same with sine sweep, I think it sweeps through every frequency with the same energy.
Try white noise, record it twice, then phase invert one of them, if it cancels to silence then you know it was a sampled white noise and happy days. If they are different, hmm, we will have to keep thinking.

Alphacodex · Post by **Alphacodex** » Sun Mar 19, 2017 5:12 pm

Dammit... no luck with the white noise... It seems to be generated somehow. I even tried MAutoAlign because I had the idea that maybe the sample startpoint is maybe randomized, but still no luck. Interesting, because that also means, that nexus is NOT just a rompler... or did they add a round robin for white noise???

So maybe I'll try a shot sample next, I read a gunshot also contains all frequencies... although there might be some filtering involved so the response wont be exact...