MTurboReverb preset making action, let's make the ultimate reverb! ;)

[caulixtla]

I have made three demos of MTurboReverb. As a demonstration of its flexibility, all three demos use a single algorithm which I have developed for MTurboReverb. In the first demo, we hear this algorithm being used to simulate a large recital hall (we first play the piano dry, then add MTurboReverb to the piano sound):

https://soundcloud.com/caulixtla/mturbo ... piano-hall

By adjusting a few parameters in MTurboReverb’s “Easy” screen, we can hear the same algorithm being used to simulate a small chamber (we first play the drums dry, then add MTurboReverb to the drum beat):

https://soundcloud.com/caulixtla/mturbo ... um-chamber

And, we can even have the same algorithm to make long ambient sustained sounds:

https://soundcloud.com/caulixtla/mturbo ... g-ambience

I have already submitted this algorithm to Melda Productions for consideration to be one of the many algorithms which MTurboReverb will come with.

[caulixtla]

Here is another drum chamber using this same algorithm:

https://soundcloud.com/caulixtla/mturbo ... ber-demo-2

Without changing the algorithm at all, I was able to reduce metallic resonances by increasing the modulation. While this would give a noticeable warble with pitched sounds, the modulation is not very audible with these drum sounds.

The attached image on the Soundcloud page are all of the “Easy” parameters used in the reverb for this demo.

[Chandlerhimself]

Caulixtla - Sounds really good. I'm interested in trying it.

[Batty]

So can this Reverb simulate the clarity of a bricasti or valhalla or maybe a good Lexicon variant?
It's the clarity I am searching for, so that a mix doesn't get muddy while using several instances of MTurboReverb.

[MeldaProduction]

Sure it can, I quite doubt there's something which it cannot match, the problem is to do the programming . I quickly discarded basically all reverbs out there except for valhalla, but I'm already past that too. Now I'm on bricasti, the last man standing . I must say, to my ears some of the designs I have here already exceed bricasti by far when it comes to realisticity, but I still want to make a few of sounds similar to it, they are good after all. We'll see where it leads. I'm already trying the bricasti hall for the 3rd time from scratch, because I always end with something "better to my ears", so I ditch the bricasti simulation and just do it my way . I just hope one day I won't wake up and say "these sounds suck" .

Edit: For the record, I usually diverge from simulating bricasti exactly because it seems quite muddy to me...

[Batty]

Sounds like you're on the way, Vojtech!
I do care for room verbs and do like vintage verbing - if it advances a song, but I mostly care for clarity.
In most reverbs I heard muddiness after the initial sound ends, which in turn decreases clarity in the song, multiplied with the many instances used in different tracks.

Valhalla is the thing I can currently compare with on my computer, having the demos and since them I disuse Mreverb (sorry :p hehe), but instead grab for MConvolution all the time.
From my expierience (recherching on the Internet, not actual expierience) most people and the professional engineers say that bricasti is the one on top to reach for. And not only the software implementation, but the original hardware. They are talking about something like a more "alive" sound and more "musicality" that helps to enliven a song more easily.

Also Bricasti and Lexicon are called 'Industry standard'...but that always encompasses only a few of the algorythms in there.
Most Verbs stuff the song tight after only 2 instances of normal use. The better ones can be used in 4 to around 10 with caution.
Most are hard to keep subtle while having an impact at all. But from what I have heard in this thread...MTurboReverb sounds verrrrry promising.

Will it possibly arrive in version 10.8 or more likely later?
Will there be a listening function, where one can check on the added Hall soloed out from the original for A/B-ing?

[Tagirijus]

Will it possibly arrive in version 10.8 or more likely later?

I guess Vojtech wrote that it might be published mid or end of january 2017 ... but I assume that this is even just a guess from Vojtech himself.

Will there be a listening function, where one can check on the added Hall soloed out from the original for A/B-ing?

You can just turn the dry / wet knob to 100% to listen to the reverb signal "soloed". I cannot imagine anything else which could make a sense regarding "soloing the reverb wet signal"

[caulixtla]

In most reverbs I heard muddiness after the initial sound ends, which in turn decreases clarity in the song, multiplied with the many instances used in different tracks.

To me, a “clear” sounding reverb is one which does not add any color or character to a sound. While such a reverb can sound plain, it can be used to enhance the sound of a mix without the reverb affecting the character of a mix the way a “colored” reverb does.

To get in to the technical details of reverb design, it means that ideally, the reverb responds to all frequencies equally. An impulse response of the reverb should, on a spectrogram, have the same decay for all frequencies across the spectrum — its impulse response should sound like decaying white noise.

To make an algorithmic reverb with this design, I would have relatively few early reflections, since the goal of this reverb is to enhance the sound without imposing its own character on it. For the decay, I do not think a FDN will work: A FDN adds a lot of color and character to the sound of a reverb. That’s good if trying to, say, emulate the sound of a 1960s reverb chamber (something someone requested about 10 pages back), but not so good if trying to get that Lexicon “you can’t really hear the reverb” sound. The “smooth” algorithms in Valhalla Vintage Verb have this feature too — the reverb, while rich, does not impart too much of its own sound (which, to my ears, makes these algorithms in VVV sound a bit more plain than some of the algorithms I have made in MTurboReverb, but I see the appeal for this kind of sound).

The Lexicon designs tend to not use comb filters nor FDNs. They use allpass filters, modulators, and delay lines. The MTurboReverb parlance for this is “a” (allpass), “v” (modulating delay line), “fam” (modulating allpass filter), “d” (fixed delay line), and something called “stfb” to tie it all together in a Lexicon-like structure. Before sending the signal to the main “stfb” reverb tank, I would either use one of MTurboReverb’s diffuser early reflection modules, or have a number of allpass filters with low diffusion (allpass filter feedback) in series (this delays the onset of reverb, but has a nice transparent diffuse sound to it).

Modulation will probably be needed to keep standing waves to a minimum, but I would keep such modulation subtle.

So, yes, this is an interesting challenge, and a very different one than the request we got a few pages back for a 1960s chamber reverb.

I hope to have time to implement this kind of transparent reverb.

[MeldaProduction]

Actually no reverb is flat , it essentially cannot be done, there's always some coloration. The only IR that doesn't color at all is an impulse . That's why some FDN designs actually degenerate to impulses . Also Lexicons don't seem flat to me at all, far from it, plus often horrible spatial image. But no wonder, the designs are decades old, so personally I consider them worth learning from, but not worth simulating really.

As for FDNs - as I explained, they can theoretically simulate any basic approach, certainly including allpasses, nested stuff, delays etc... But designing them like that would be a horror, plus they would waste CPU like nothing else . But generally FDNs are no worse than anything else when it comes to their potential. Of course, in MTurboReverb you can do anything with the feedback paths, which is not a classic part of FDN design (though here you can do that as well). Btw. since you like the flatness - then serial taps are probably the worst things imho, these are really parts of the ancient design and the reason was probably mainly the efficient implementation, not audio quality. Though, maybe they just experimented with it.

[chris.r]

@caulixtla
love this piano hall

[plexuss]

Sure it can, I quite doubt there's something which it cannot match, the problem is to do the programming . I quickly discarded basically all reverbs out there except for valhalla, but I'm already past that too. Now I'm on bricasti, the last man standing . I must say, to my ears some of the designs I have here already exceed bricasti by far when it comes to realisticity, but I still want to make a few of sounds similar to it, they are good after all. We'll see where it leads. I'm already trying the bricasti hall for the 3rd time from scratch, because I always end with something "better to my ears", so I ditch the bricasti simulation and just do it my way . I just hope one day I won't wake up and say "these sounds suck" .

Edit: For the record, I usually diverge from simulating bricasti exactly because it seems quite muddy to me...

So basically all other reverbs suck except yours. Got it.

[MeldaProduction]

Absolutely not, read more carefuly before replying! (btw. if that would be true, wouldn't that be awesome ? )

But when it comes to realisticity (which is not what you are after, you made that clear many times, and I have absolutely no problem with that!), I'm generally very unsatisfied with existing algo reverbs. They just don't sound realistic to me at all, not even far... I'm actually trying to understand, why... But no idea so far. It almost seems to me like Bricasti and Lexicons are duplicating their designs despite they are not realistic at all (I'm mainly talking about the delayed buildup of reflections, any ideas?).

[onerob]

I would like to see presets in this plugin that emulate, as closely as possible, some of the classic reverb hardware and software, as well as model realistic spaces. If both of those are present, it would be pretty amazing.

[plexuss]

Absolutely not, read more carefuly before replying! (btw. if that would be true, wouldn't that be awesome ? )

But when it comes to realisticity (which is not what you are after, you made that clear many times, and I have absolutely no problem with that!), I'm generally very unsatisfied with existing algo reverbs. They just don't sound realistic to me at all, not even far... I'm actually trying to understand, why... But no idea so far. It almost seems to me like Bricasti and Lexicons are duplicating their designs despite they are not realistic at all (I'm mainly talking about the delayed buildup of reflections, any ideas?).

Well you can't really make a realistic sounding reverb when confined to reproduction over 2 speakers. or even 7 speakers. I am not sure that the reverb manufactures are trying to make realistic sounding reverbs but rather reverbs that are "musical" or otherwise work well with various kinds of recorded sound.

In fact, many real reverbs are not musical at all. The real reverbs that sound good are often manipulated to sound good with music in various performance spaces etc. Then again, to reproduce that you'd have to reproduce the actual space. I just don't think you can create anything that resembles a real reverb, musical or not, using digital code and speakers. You may want to go into some performance spaces and explore and analysis what makes a real reverb sound the way it does (rather just from memory etc).

So far your reverb sounds good but it sounds like a reverb plugin. it has its own unique character that will work for many people. but even if you had some kind of alien technology in your reverb plugin, the limitation is going to be the reproduction of that reverb via speakers and so it will never sound real. I think what most of us look for are reverbs that sound good with the audio we are working with, not so much how real it sounds because you just can't get a real reverb sound without having a real reverberating space.

[MeldaProduction]

Well, I agree only partly. When I first clapped in a cathedral, I was like "wowwwww" and since then I'm clapping every time . And one probably cannot simulate this using 2 speakers indeed, ever. Sadly. I'd really like that . But I think convolution makes things sound pretty realistic, nearly as far as one can get. And so I take these real IRs as benchmark.
(btw. after spending some time with them, I often seemed to like my algo version more than the original IR, but I'm a little worried if I'm not biased now )

When it comes to "musicality", I have one problem with it - everyone considers something else musical. After all that's why there are so many music styles etc. But I hear your point. The problem is, I'm having a hard time find the "use" for many Bricasti presets. Now I'm trying the Sunset chamber, which is recommended by pretty much everyone, and it has the distinctive delayed echo buildup, and I just don't see a way to use this, trying just about every audio material... It makes the mix completely muddy and filled with "stuff" . Any hints?

Btw. the point here is that MTurboReverb doesn't really have a distinct character. It's programmable, and so it can sound in various way, and if you'd make a topology like in say Bricasti, it should be possible to make it sound pretty much the same. Or better. Or worse .