Interesting. But no, I haven't tested it. From the GUI, the plugin name and the description of the developers, it seems to be an emulation/interpretation of a Quantec, rather than Bricasti.

 

There are many reverbs with feedforward path. As well as dual-mono and true stereo reverbs. Reapers stock reverb is just that. That doesn't make it a Bricasti. But I have heard a lot of good things about these two though. Will give them a try. Thanks for reminding.

 

This is probably a misunderstanding on your side. As I have indicated, the description of Verbity (follow the links I shared above to the corresponding pages) indicates that it's inspired by Bricasti. Here is an excerpt:

Verbity is your basic Airwindows Bricasti-style reverb, which also means it doesn’t have pitch shifting: Casey [the inventor of the Bricasti reverb] doesn’t like it [pitch-shifting in reverbs], and this is an exploration into the things Casey’s talked about publically, to see if his wisdom leads to better reverbs

My understanding is that ReaVerb is exclusive to Reaper and is a convolution reverb, so I have no idea what on earth you are trying to imply by saying that "Reaper stock reverb" shares the topology with Bricasti, unless you are talking about an entirely different, obscurre reverb plugin in Reaper?

 

If it's not be said already then https://pasttofuturereverbs.gumroad.com/l/Gqfpl?layout=profile is probably the best bang for buck, although it's pretty much indistinguishable from https://samplicity.com/bricasti-m7-impulse-response-files/ with a bit of saturation this thing is killer

 

You can add modulation to IRs too.

So take Samplicity free and Convology XT free, use modulation and you are set.
https://wavearts.com/products/plugins/convology

 

Okay, chill the horses. Let's take this from the top. I might have expressed myself in a somewhat misleading way.

Verbity is a reverb with feedforward topology, i.e. the signal is split several times in a series circuit, divided, induced into delay lines, subsets merged, induced into delay lines, divided, etc. and combined until it is finally fed into a feedback loop. This serves to minimize the generation of unwanted resonances and stabilize the frequency response over time. Nothing new, nothing special, because this is used in so many reverbs today, for exactly the reasons mentioned.

Verbity and Verbity 2 are dual-mono reverbs. That's why I mentioned the Reaper stock reverb. Verbity 2 also has some basic channel crossfeed modulation, what makes it a 'quasi-true' stereo reverb, I guess? But I wasn't referring to ReaVerb (Convolution), but to ReaVerbate (Algorithmic), which is also a dual-mono reverb. So my question is, does that make ReaVerbate a Bricasti inspired plugin or a Bricasti style reverb? Especially considering that the M7 is not a simple dual-mono reverb?

So Chris has taken a few statements from Casey and built a reverb that he calls "Bricasti-style", whereas it (especially Verbity 1) has hardly anything to do with what an M7 does? Please don't get me wrong, I think the plugins sound interesting (I might even use them someday), but this communication is close to snake oil. Of course, you can't be blamed for that. It's a Chris thing. Not a Hazen thing.

 

Last edited: Oct 7, 2023

This. "bitshift pan", "mantissa preserving", "slew clipping", "32bit dither". Chris often uses misleading (and/or factually wrong) terms and phrases to make his stuff appear radically more interesting, innovative and ground breaking than it really is. It's part of the "eccentric Jesus of DSP" show he's been putting on for a decade to mislead people without programming/DSP knowledge.

 

The "modulation" in Convology XT unfortunately seems to be a simple chorus effect. I wish it had at least an offset / delay, so that the initial early reflections don't get affected (= blurred). To get a good reverb sound from IRs this is crucial imho: Modulation (if any) only on the tail, never on the early reflections (initial part of the impulse)! (Sorry for going full blown OCD here).

On the other hand the Convology XT devs have been very responsive in the past to user feedback and made lots of improvements over the cause of various minor updates.

 

Too bad.
I was expecting a "rolling" modulation.
I tested it years ago, but only for "straight" IRs.

So back to Valhalla

Nice reading on the subject (and how to create modulation on reverb insert full wet) :
https://vi-control.net/community/threads/adding-modulation-to-impulse-responses.78693/

 

BitShiftGain isn't doing anything new, but it is cleanly doing doubles or halves of your audio, as is the BitShiftPan on individual channels. That is as little snake oil as you can get, as it's not using look-up tables or dithering.

 

The point in bitshifting is that you can cheaply and losslessly multiply and divide integers by powers of 2 without actually dividing or multiplying (so long as the resulting integer fits the underlying type).

You can not shift bits on a float. They're virtually completely different from integers. Floats are stored as sign/exponent/mantissa, not their "actual" value.



Integers are straight forward:



But audio samples are passed to plugins as floats - in every plugin format and DAW.
We're limited to +, -, *, /; we do not have access to << and >> for audio samples.

Looking at the source of the pan/gain plugins confirms that he is indeed multiplying and not bitshifting.

Calling it bitshift-something but then multiplying won't make it magically shift bits. No compiler ever has or ever will create a code path where floats are re-cast as a union of sign/exp/mantissa and then only operate on the exponent to mimic bitshifts [it still wouldn't be bitshifting].

It doesn't matter what multiplication or division you apply to your sample, whether it happens to be a power of 2 or not, the resulting number will be perfectly lossless so long as the error incurred is smaller than machine epsilon (which it virtually always is).

If bitshiftpan sounds better to you, then it might be because it's using a different panlaw from your other plugins (or it's using no panlaw at all).

 

Last edited: Oct 7, 2023

Looking at the source, i just wonder why someone has to write two switch cases with 32 lines each for a gain plugin that does 6dB level steps, when a simple function is sufficient.
I have no experience in c++ but if I interpret this correctly these 64 lines of code could be replaced by 1 line with

Code:

gain = 2^(bitshiftGain);

Sry, for being so offtopic. I just found it interesting. And confusing.

 

Last edited: Oct 8, 2023

this thread made me think im behind on verbs but since im here, and you guys are here.. i have another question.

Q: Travis Scott - Loove from his latest album utopia, i keep hearing this saturated roomy tone especially in the second verse. To me it even adds some weight to the vocal. I was wondering.. is this a room/short verb or is this the actual room i'm hearing?

 

i found out in this tread to, havent tried it yet but im eager to try