Hi all
I got a stupid question. Why you can not capture reverb pedals with the QC? I mean the technology is there for convolution reverb, with a frequency sweep and all. Is this due to to the modulation or pitch of certain pedals? As a first step to capturing reverb, wouldn’t it be cool to get a test frequency within the QC to capture room reverbs (convolution)?
Cheers
The QC cannot capture any time based effects. No current capturing technology will capture time based effects etc.
Yeah i got that but WHY?
Assuming they would also need to develop a convolution reverb device and software for the device to create an impulse response for it. Probably more involved than they would want.
But could be cool!
A lot of time-based effects captures are HUGE files, digitally, so they don’t make sense to use as captures but rather to just reproduce in software versions. A solid reverb IR would be like 10x the filesize of a cab IR, for example.
Digital reverbs are massive, complicated algorithms that change at every point along the decay of the reverb in terms of EQ, volume, density, etc. There’d be no way to accurately “capture” this with a static setting of the target reverb. They’re not just a filter that the sound passes through like a drive or amp. They’d be more like a filter feeding into a filter feeding into a filter … feeding into a filter, each with variables adjusted for each iteration.
I’m sure someone will eventually make it happen, but I suspect it would take active participation on the user’s behalf, following instructions (now, turn the decay all the way up. Now, turn the mix all the way up, etc.)
Interesting idea, though!
Digital reverbs have been recreated with IR’s for many years now actually. Plugins such as this have existed since the mid 2000’s - https://www.sweetwater.com/store/detail/Altiverb8--audio-ease-altiverb-8-convolution-reverb-plug-in-download
Yes the convolution reverbs exist for a long time now. I am not speaking about adjusting the decay or any parameter. I talk about a static IRs of a Room. I get that is too complicated with modulation and such.
The QC is locked at 21ms (or so) for IRs = 1024 samples @ 48khz.
21 ms is way too short for a reverb IR (even a simple room reflection).
…sadly
Current approaches to machine learning cannot capture things that change over time without providing some input that can be added to the training set that correlates with the change. This is how NDSP can capture the controls on an amplifier - by including the knob positions and their effect on the amp output in the training set. But doing that for reverbs or modulation hasn’t been covered yet.
To note a few things as I’ve made some reverb IR’s.
The process isn’t complicated. It’s pretty much the same as capturing an amp IR, just longer. The software to do it isn’t specialized either, you can actually use what’s included in Logic.
The file isn’t that large. 1.9mb is still rather small when we have gb’s available. You could fit the QC with 6,000 high quality reverb IR’s and still have over half the memory left.
This has already been done in a small unit. The strymon MX. To be fair the QC is running 2000Mhz and the MX is running 800Mhz tricore. But overall the MX can still process parallel 10 second IR’s ro stack two for a total of 20 seconds of IR.
Mathmatically the difference: the QC’s 21ms IR length is 0.21% of what the BsMX is capable of on one of its lanes.
So yes, should the hardware be able to do it? Yep.
Can the programmers program it? Yeah probably.
Can you do more than reverb? No not going to happen, but you can get some delay-ish sort of effects sometimes.
Will it eat DSP like crazy? probably.
Like many other things. Is it possible? Yes. Has it been done? Yes. Is it a priority? No.
And I don’t mean this as a slight but just as a fact. NDSP is more n the business of selling things that go CHUGGACHUGGACHUGGA and not Ack ack ack ack ack ack ck ck k k k . . . k. … …k …k. …