How good is Altiverb next to UAD-1 reverbs?

[Hemmick Reef]

I noticed that Altiverb is due out soon for PC.

It has IR's of classic hardware reverbs as well as real spaces - is this the plug that truly covers everything to a very high standard?
Do the reverbs still muddy the sound the more you process, like cheaper reverbs: an effect more than add space?

Thanks

 

[Dan Duskin]

Alitiverb can't do modulation...

 

[Hemmick Reef]

Is that bad?

Can it be added with other plugins?

Cheers

 

[Ashermusic]

Alitiverb can't do modulation...

AudioEase has done a beautiful job sampling great European Halls, Churches, Cthedrals, etc for Altiverb. They sound great.

Plate 140 sounds great but it is a one trick pony.

The other UA verbs are nothing special IMHO.

 

[s_sibs]

Yeah,

The only UA verb I like it the Plate 140. It's great for that 50s, 60s vibe. I never really got into RealVerb or DreamVerb.

I've never heard the AudioEase plug but I did download a demo for WizooVerb W2. For $249 it is a great plugin for the price. It has it's own IRs and you can use your own .wav IRs. I'm not sure how long this verb has been out but I eally like it.

 

[tiger vomitt]

ive used Altiverb a lot over the years. i own all the UA reverbs.

I use Altiverb more. the plates are meatier. the small rooms IR's ive got are amazing. like the bathroom one, inside of cars, and all the other weird ones (djembe, reverb mic, zube tube, etc)

the plate 140 sounds better than AV maybe one out of every 5 or 6 times. sometimes it really is better for the situation. but usually i prefer the sound of altiverb. it's a little thicker somehow. i find i dont need to EQ it so much (if at all), as opposed to the plate 140 plugin, which usually as soon as i turn it on im reaching for its EQ controls. the UA plug doesnt sit in a mix the way that wendy carlos plate IR does

im not trying to dis the UA stuff, i love it all. even nigel sometimes
but yeah, altiverb is pretty great too. i havent used any other IR processor reverbs so i dont know if it's an AV thing or something common to all reverbs of that type.

HTH!

 

[Tony Ostinato]

Convolution Verbs all use the same math process and pretty much all sound alike, they are very good for one reverb setting all the way thru the song, they arent very good at changing decay times on the fly or any other really big changes in settings.

Other than their sampled reverb impulses theres not that much difference between altiverb and all the other convolution reverbs.

Theres so many convolution verbs out now, SIR is free even, that i don't think theres any reason for UA to do one, especially since theres no such thing as a vintage convolution reverb.

I dont really get onboard with comparing convolution to algo based or modeling based plugins, i dont think its a real comparison. they are 2 different tools for different purposes.

 

[tiger vomitt]

that's a very good point about automating parameter changes. it's true, AV is not really made to do that.

 

[Miser]

Hello

What resolution do those Altiverb Plate IR`s have(Wendy Carlos`and Chapman Recording`s)?Are they at 24bit/44.1kHz resolution?
I don`t understand or remember how this convolution process works at the sample level,but wouldn`t this fixed
sample resolution be a problem when we have audio track at 24bit/96kHz resolution?
In other words in cases when IR`s resolution is lower than actual audio track.I`ve heard that IR`s resolution
should even be higher than processed audio file.I`ve been told that true convolution is not possible on current computers,
it is too demanding(especially if it is performed realtime).The UAD-1 Plate 140 works perfectly at 32bit/96kHz resolution.


In my mind this convolution process is second generation processing as those IRs are audio files recorded through
various signal chains.The signal chain in this case would be: impulse or sine sweep -->DA converter-->EMT Plate-->AD converter.
Then follows the deconvolving process that gives the final IR.Then this IR is used at the sample level to present captured gear or space.
UAD-1 Plate is algorithm and as such it is first generation processing.This is just my reasoning so please correct me if I`m wrong.

What reverb times do these Altiverb IRs represent?They obviously are fixed so if there is no matching reverb time in the library
what can you do then?I guess there is no maxed reverb times among those Plate IRs.The Plate 140 has all the controls of the original and some more like said earlier by others.
Of course the main thing is how these IRs sound,but it is the only thing that I can`t comment since I haven`t heard them.
There`s no test samples of these Plate IRs on the Net.Some snare drum and other samples through
them would be nice.However,what I do know is that IRs from Plate 140 plug-in do not sound the same or as good as the original plug-in.
So I guess the convolution is lossy process after all.

As the Altiverb is iLok/Pace related it is a big problem to me personally.My system corrupts everytime I install
something that has Pace drivers.Ok, they work on my old Intel P4,but not on main AMD 64 based system.I try to avoid
them all now.It seems that there is less and less native plug-ins from major companies that do not have iLok/Pace connection.
The UAD-1 card is the best copy protection in the world!!

In the end it is how it sound not how we get it.

 

[Tony Ostinato]

What resolution do those Altiverb Plate IR`s have(Wendy Carlos`and Chapman Recording`s)?Are they at 24bit/44.1kHz resolution?
I don`t understand or remember how this convolution process works at the sample level,but wouldn`t this fixed
sample resolution be a problem when we have audio track at 24bit/96kHz resolution?
In other words in cases when IR`s resolution is lower than actual audio track.I`ve heard that IR`s resolution
should even be higher than processed audio file.I`ve been told that true convolution is not possible on current computers,
it is too demanding(especially if it is performed realtime).The UAD-1 Plate 140 works perfectly at 32bit/96kHz resolution.

you can really think of convolution plugins and sampler vsti plugins along the same lines, and much like you can run kontakt at 96k with a 44k sample bank in it you can do the same with a convolution verb, which is basically using a sample of a reverb impulse to drive the effect.

to extend that analog you could think of plate140 as a synth, specifically a physics modeling synth but lets not get all complexicated hehehe, the effect is driven entirely by algo with no required \"sample\" in memory.


Convolution plugins are easier to code, and in the end the quality is in the sampled reverbs, not in the plugin. Compare SIR, Pristine Space, IR-1, Wizoos, and all the other convolution reverbs out there and you quickly realize why they encrypt their samples so nobody else can mix and match samples and plugins and realize that, using the same sampled impulses, convolution plugins are basically more the same than different.

Plate 140 is a completely different approach, so different that i dont think comparisons are valid. Its a component by component physics model of the actual unit, very hard to code and more about giving you the exact behavior of a specific unit than a whole bunch of different reverbs.


You just wouldnt use these 2 things for the same purpose, with the exception of running a 140 sample vs the uad-1 140, and i would say at a static setting throughout the song the point would probably be moot, it isnt gonna be that much of a difference.


[/quote]

 

[polygen]

as was said before, convoverbs, modelled \"mechanical verbs\" like the plate 140 and algorithmic verbs are really apples, oranges and bananas. Each has it's own advantages and disadvantages. So in a real-life production situation you'll want all these varieties in your arsenal.
Amongst the convoverbs altiverb is one of the better sounding because of it's ability to do semi-true stereo (it can process the left input with a stereo IR and the right input with a different stereo IR...provided you have \"true stereo IRs\" of this type you get better results than with other convoverbs). Also its IR library and it's hi-/lo-shelf algorithms are excellent. Another convoverb that is really cool for entirely different reasons is Prosoniqs Rayverb...it analyses an IR that you load and models the room that would have this IR as its response. You can then alter the features of that room in a rather realistic way. Its simply a killer IIRC the \"inverse ray-tracing\" technology used in that is also used in astronomy to analyse sunspots on the far side of the sun....talk about rocket science

cheers

 

[Miser]

Hello again

Physical modeling surely don`t sound the same as static samples.
Let´s say we have analog oscillator that drifts and is constantly running.
You can`t model that with a static or looped sample.Samples phase is
always the same when it is triggered,but analog oscillator cycles through.
Like said by many here and else where, convolution reverb is as good as the IR library.So sample resolution is important and quality of AD/DA conversion is part of it(when sampling analog \"world\").
IRs are great for room reverbs and can be more life like than algos.

 

[cAPSLOCK]

I agree with much of what is being said here, but take issue with a lot of it as well.

Convolution as a method to capture and use reverb has strengths and weaknesses.

I would say that the biggest problem with convolution verbs is they can lack a certain life of depth due to the static nature of an impulse response file. The best reverb hardware always uses a certain amount of chaos to achieve something that sounds natural, or special. Since IRs are static 'samples' of a reverb then you can lose that chaos in the result. They can sound flat or two dimensional compared to the hardware.

But well made impulse responses are often nearly indistiguishable from the 'real thing'.

It really depends on the nature of the source, the application and above all the quality of the the IR.

As far as making something sound like it is in a real space goes convolution blows all competition out of the water.

Artificial reverb methods and more receltly algorithms have been designed first to sound like REAL space. Secondarily (and very importantly) the sound of artificial reverb has evolved to a point whre it sets it's own standard even outside of what we would consider 'real'.

Convolution is better at real (halls, rooms, cathedrals, and so on) than at artificial (plates, Lexi 960 etc).

Still.. a well sampled hardware box can sound amazingly good. And impulses can allow you to use high quality verbs on some track you couldn't have before freeing you up to use the box on something really important in the mix like the vocal.

To argue that convolution reverbs are 'one trick ponies' is entirely absurd.

If fact, I would argue that there is nothing as versitile in the reverb world as a good convolution plugin and maybe 50 really nice impulses. I can do almost anything I want reverb wise with about 10.

I haven't even bothered to sample the oft used Plate140 plugin believe it or not...

But if you guys are up to a test I will do a A/B for you with plate 140 impulses and the plugin so you can get a feel for exaclt how different (or not) they actually are.

Any takers?

cAPS

 

[Miser]

Like said earlier convolution can sound better than algos when it comes to room reverbs etc.
I think small rooms are what those IRs shine at(difficult for algos).Let´s not forget that convolution is
a great sound designing tool as well.You can get quite farout sounds by experimenting.What convolution
cannot do is non-linearities and those are things what make \"chaos\" and depth like hinted earlier by cAPSLOCK.
My first hand experience tells me that if not made properly and with high quality deconvolving,
IRs lack the depth and won`t tail as well as the original source(in case of
hardware or software reverbs).
In the mix those subtleties usually get masked so I guess it`s down to what works best in given situation.
NI uses sort of dynamic convolution on their cabinet models that is designed to be easy on CPU (Guitar Rig,B4II).

 

[xist2005]

I'd like to know from \"outboard gurus\" here, what do they think about the IRs of famous gear like Lexicon, etc. which you can find in the Altiverb libraries...
From the explanations I have read, they should not sound that good because of the nature of the original machines.
Unfortunately, in the native plugs world, we still miss a good emulation of those classic hardware reverbs. Now there's that IK new product which seems to be aimed at that, but I don't like IK so much, to tell the truth!
ProTools and Vst users have more chances, like the 2016 emulation, which unfortunately will never be ported to AU (word from developers).

 

[Dan Duskin]

I'd like to know from "outboard gurus" here, what do they think about the IRs of famous gear like Lexicon, etc.

Lexicon was THE reverb company to add modulation to the verb tail... this was (and is) the biggest part of the lexicon sound. If you impulse a Lexicon verb with modulation (most lexicon verb presets have modulation) it will sound more like a plugin than a lexicon.

 

[Tony Ostinato]

I'd like to hear an example where you slowly increase the decay time to 8 seconds and then decrease it to 1 and then back out to 8.


I never said they were one trick ponies, but theres certainly tricks they cant do. Since i wasnt in the cubase forums i dispensed with the 3 page disclaimer explainer for all the snooty kids who love to misquote and add words never said.


I think people like convolution verbs in much the same way they like samplers, they get that \"im stealing the sound\" thrill.

but much like how samplers cant slur correctly and are severely limited in expression so are convo plugs, in fact convo plugs are worse since none do dynamic convolution (yet).

 

[Miser]

The true dynamic convolution seems to be too demanding process to run on
current computers so there is shortcuts in use or this I believe is the case.I don`t actually know if it is\"watered down\" dynamic convolution that NI uses in
Guitar Rig.I was just told by NI that it is something like that.So Tony or somebody else please enlighten me on this subject.
I have to emphasize that sample resolution is critical when making IRs from
analog world.IR at 16bit/44.1KHz resolution is different than at 24bit/96kHz.The reverb tail is preserved better by higher resolution sample.
So why Altiverb Plate IRs sound different than Plate 140?It is a question of what was the signal chain (how true the converters were etc.)and what kind of Plate were sampled(and of course that sample resolution).Every EMT Plate sounds different and that is just how
analog world is(so lovingly random and chaotic).
:wink:

 

[Miser]

Since I love the UAD-1 Plate 140, I want to defend it`s place against the
convolution reverbs here`s a little quote:
\"As plates are pretty linear within a broad range, any transducer-based non-linearities, and other non-linearities, can be emulated using some of the standard processes that we've added on top of the reverb.\" Joe Bryan-UA

So it seems that there is a lot more going on under the hood.:wink:

Sorry if I´m repeating myself all the time.I find it hard to express myself
the way I intend to.

 

[cAPSLOCK]

I never said they were one trick ponies, but theres certainly tricks they cant do. Since i wasnt in the cubase forums i dispensed with the 3 page disclaimer explainer for all the snooty kids who love to misquote and add words never said.

It wasn't you I misunderstood Tony, it was Ashermusic I misunderstood. And he was calling the Plate 140 a 'one trick pony' not convolution. So I was wrong, but not the paranoid way you thought I was.

And I am not a snooty kid.

cAPS