Illusion - new reverb from LiquidSonics

VST, AU, etc. plug-in Virtual Effects discussion
liquidsonics
KVRist
496 posts since 30 Jan, 2009 from UK

Post Mon Oct 29, 2018 6:10 am

ghettosynth wrote:
Mon Oct 29, 2018 5:42 am
liquidsonics wrote:
Mon Oct 29, 2018 5:32 am
ghettosynth wrote:
Mon Oct 29, 2018 5:05 am
This seems like it is similar to H-Reverb, no? Reading in between the marketing, which is thick with this tone for sure, they seem to be synthesizing Fusion-IRs and then using those IRs in the reverb, no?
Waves' tech in H-Reverb is not to be confused with Fusion-IR Synthesis which is fundamentally different in a number of key ways to do with the reverb generation and modulation, though I don't want to offer a blow-by-blow comparison of a competitor product of course as it'll just come across as me saying x y and z is better when it's best for people to decide if they prefer it or not themselves.
I'm sure, but, the differences were not my point so much as the similarity is. You are generating IRs on the fly and then using those IRs in a convolution reverb, no?

I'm making no judgement to sound quality here, only commenting on the similarity because it does impact usability. Moreover, the ability, of this reverb to be of sonic interest like H-verb, comes down to how well you can generate/simulate those spaces. My comment about H-Verb was just a first shot at categorizing this reverb in terms of where it fits in the reverb spectrum. Of course there are going to be differences, but, to be clear, one of the reasons that I like Reverberate 2 is that it loads Waves IRs and I use it much more as just an IR verb that is super flexible as I do a Fusion-IR verb.

I understand that any firm has a need to capitalize on their technologies as best as they can with respect to what kinds of products the market has shown a willingness to buy, but, these kinds of generative IR products are less interesting to me than either pure algorithmic reverbs or more traditional IR reverbs. I too would have preferred to see Reverberate 3 as opposed to this approach.
Off-line computation of impulse responses is going to be a massively broad topic in future, I don't think it fits in a general characterisation. You couldn't really put a full and comprehensive ray-traced environment, a static physical model of a plate with tens of millions of simulated resonances (which can't be done in real-time), and all sorts of other things including the inherently modulated Fusion-IR Synthesis algorithms of Illusion in one general bag, it's too much of a generalisation. All of those would sound nothing like a regular tapped delay/FDN algorithmic reverb or each other. I suspect over time attitudes around the superiority of algorithmic and the general negativity around convolution are going to evolve. Seventh Heaven proves Fusion-IR can reach incredible levels of accuracy and quality using the best hardware source material, and Illusion takes it in a really novel new direction which sounds great but comes with a new level of flexibility without leaning on the M7 to do it.

This tech is just not suited to fitting into Reverberate, if I'd squeezed it in people would complain bitterly about the added complexity and lose so much in the very focused way the reverb components can be controlled. Any future version of Reverberate would double down on what it does best, which is provide a great way to load general purpose and Fusion-IRs from a wide range of sampled sources and allow editing of them in powerful ways, but with a new focus on improvements to the interface. Of course that is a product I am interested in creating as well.
Matt
www.liquidsonics.com
Follow us on Facebook

ghettosynth
KVRAF
11421 posts since 13 Oct, 2009

Re: Illusion - new reverb from LiquidSonics

Post Mon Oct 29, 2018 6:56 am

liquidsonics wrote:
Mon Oct 29, 2018 6:10 am
Off-line computation of impulse responses is going to be a massively broad topic in future, I don't think it fits in a general characterisation.
Until we get to the point where parameter changes appear to be transparent to the generation and loading from the user's perspective, yes it does fit into a general characterization, and that was my point. Like H-Verb, you are generating and loading, and so, like H-Verb, one cannot smoothly automate the parameters that cause new IRs to be generated.
You couldn't really put a full and comprehensive ray-traced environment, a static physical model of a plate with tens of millions of simulated resonances (which can't be done in real-time)
Today.
I suspect over time attitudes around the superiority of algorithmic and the general negativity around convolution are going to evolve.
Tomorrow.

I'm not ignorant of the technology and I don't have an "attitude" about convolution reverbs. My point holds, it applies to this reverb, it applies to H-Verb and it applies to UVI plate. While there are benefits to that approach, there is, currently, a cost as well and these kinds of reverbs are, point blank, not capable at the moment of replacing algorithmic reverbs where one wants to modulate the very parameters that you are using to generate IRs. That use case is what lumps them into the same category.
This tech is just not suited to fitting into Reverberate
I don't think that anyone is arguing that it is.
Any future version of Reverberate would double down on what it does best, which is provide a great way to load general purpose and Fusion-IRs from a wide range of sampled sources and allow editing of them in powerful ways, but with a new focus on improvements to the interface. Of course that is a product I am interested in creating as well.
Awesome, I can't speak for anyone else, but that's what I wanted to hear. I think that's all that people are saying. You have to create the products that work for your business goals, but, for some of us, those aren't the products that we're looking to buy. I want the best and most flexible IR reverb. But I want to load spaces that are interesting to me, not figure out what parameters I need to generate those spaces. It would be great to be able to create Fusion IRs, either from existing IRs, if possible, or from a sampling process. I think that's a product that those of us looking for Reverberate 3 would pay money for.

I'm sure that I'll demo this, but I'm less excited than I would be for a new Reverberate and that is because of the reasons expressed here.

liquidsonics
KVRist
496 posts since 30 Jan, 2009 from UK

Re: Illusion - new reverb from LiquidSonics

Post Mon Oct 29, 2018 7:35 am

ghettosynth wrote:
Mon Oct 29, 2018 6:56 am
liquidsonics wrote:
Mon Oct 29, 2018 6:10 am
Off-line computation of impulse responses is going to be a massively broad topic in future, I don't think it fits in a general characterisation.
Until we get to the point where parameter changes appear to be transparent to the generation and loading from the user's perspective, yes it does fit into a general characterization, and that was my point. Like H-Verb, you are generating and loading, and so, like H-Verb, one cannot smoothly automate the parameters that cause new IRs to be generated.
I have plenty of software reverbs that exhibit audible artefacts when some parameters are changed so one wouldn't as a result choose to automate those parts, and even on a hardware M7 you can edit some parameters and it causes the audio to drop out while something inside rejigs itself. In a v1 preset the modulation control for instance, or a v2 preset also the density does it. I wouldn't say its delay time change is super smooth on automation either like a basic feedback delay loop reverb would be because clearly it's doing something way more complex than just increasing and decreasing a loop gain somewhere. I don't see this as a problem with the hardware and haven't ever read any complaints about that, so I am just making the case that even the best hardware algorithmic reverbs aren't always transparent to automations so it feels like a harsh judgement criteria.

In cases like mine (Illusion, Seventh Heaven) it would be feasible to run a parallel reverb and smoothly transition to every update switching off the old reverb thread when it's finished a transition if I didn't mind the memory and CPU penalty of doing so. I don't do it as in general I don't think it's worth the cost of hitting the CPU for double the work during transitions, but maybe some would disagree. There's no really fundamental limitation of this sort of approach as far as I see it and it's something I'll be considering in future designs.
Matt
www.liquidsonics.com
Follow us on Facebook

User avatar
solidtrax
KVRian
578 posts since 6 Nov, 2011 from The Netherlands

Re: Illusion - new reverb from LiquidSonics

Post Mon Oct 29, 2018 7:50 am

ghettosynth wrote:
Mon Oct 29, 2018 5:05 am
This seems like it is similar to H-Reverb, no? Reading in between the marketing, which is thick with this tone for sure, they seem to be synthesizing Fusion-IRs and then using those IRs in the reverb, no?
As soon as you start to compare the sound of Illusion and H-Reverb, it's miles apart in my experience. I just finished testing Illusion for the bigger part of last hour and I think it sounds much better than H-Reverb. I would have purchased it, if it wasn't for the fact that I have 20 reverbs in my plugin arsenal already. I will continue to test it some more though, who knows.
Dawn of Dreams - 100 lush and shimmering presets for Reveal Sound Spire
Drumspiration - 80 electronic drum presets for Reveal Sound Spire
Website | Soundcloud | Twitter

ghettosynth
KVRAF
11421 posts since 13 Oct, 2009

Re: Illusion - new reverb from LiquidSonics

Post Mon Oct 29, 2018 8:02 am

liquidsonics wrote:
Mon Oct 29, 2018 7:35 am
ghettosynth wrote:
Mon Oct 29, 2018 6:56 am
liquidsonics wrote:
Mon Oct 29, 2018 6:10 am
Off-line computation of impulse responses is going to be a massively broad topic in future, I don't think it fits in a general characterisation.
Until we get to the point where parameter changes appear to be transparent to the generation and loading from the user's perspective, yes it does fit into a general characterization, and that was my point. Like H-Verb, you are generating and loading, and so, like H-Verb, one cannot smoothly automate the parameters that cause new IRs to be generated.
I have plenty of software reverbs that exhibit audible artefacts when some parameters are changed s
Yes, and there are plenty of reverbs where this is not the case and, in fact, is a conscious design choice. I'm not sure what point that you're trying to make here, mine was simple, changing the parameters yields a long delay and so, like HVerb, it cannot be changed quickly.

It's not only about modulation BTW, it's also related to usability while dialing in a preset. The long delay slows down workflow. Even if there are artifacts in an algorithmic reverb, the rapid change facilitates quickly dialing in what one is looking for.
I don't do it as in general I don't think it's worth the cost of hitting the CPU for double the work during transitions, but maybe some would disagree. There's no really fundamental limitation of this sort of approach as far as I see it and it's something I'll be considering in future designs.
Yes, I definitely disagree. It's absolutely worth the CPU for some users. I think that it would be best as a per/patch option. It would make the difference for me in these types of reverbs. I wish that UVI plate did this.
Last edited by ghettosynth on Mon Oct 29, 2018 8:17 am, edited 1 time in total.

ghettosynth
KVRAF
11421 posts since 13 Oct, 2009

Re: Illusion - new reverb from LiquidSonics

Post Mon Oct 29, 2018 8:13 am

solidtrax wrote:
Mon Oct 29, 2018 7:50 am
ghettosynth wrote:
Mon Oct 29, 2018 5:05 am
This seems like it is similar to H-Reverb, no? Reading in between the marketing, which is thick with this tone for sure, they seem to be synthesizing Fusion-IRs and then using those IRs in the reverb, no?
As soon as you start to compare the sound of Illusion and H-Reverb, it's miles apart in my experience. I just finished testing Illusion for the bigger part of last hour and I think it sounds much better than H-Reverb. I would have purchased it, if it wasn't for the fact that I have 20 reverbs in my plugin arsenal already. I will continue to test it some more though, who knows.
Sure, like I said, I'll demo it, just not today, because reverbs that work like HVerb, no matter how good they sound, are just less interesting to me in general.

Of course, it goes without saying that sounding better than HVerb isn't the standard here. I'll be comparing it to the best of my algorithmic verbs.

I think that in all of this discussion of sound that there's something missing in the discussion that I value a lot, and that's interesting features and how they map to sound. I like, for example, the smoothly modulatable moving microphones in Eventide's T-Verb. It doesn't matter if you can get that sound through other means, it has a usability factor that causes me to gravitate towards it when I want to move things about in a space.

Similarly UVI's plate feature that allows you to switch materials is also useful in the same way. Now, do I have experience with titanium plate reverbs? Of course not! That's not the point, the point is that the attention to that detail maps to something that is memorable and sonically different than the other choices. So you recall the last time, for example, that you used a titanium plate it had a certain sound and you build connections in your mind between the features of the plugin and the sound.

At any rate, I'm putting the cart before the horse here, I do need to try it out. My comparison to HVerb though was about the basic underlying technology and the limitations that it presents to usability.

robotmonkey
KVRAF
2113 posts since 20 Jun, 2012

Re: Illusion - new reverb from LiquidSonics

Post Mon Oct 29, 2018 8:18 am

So... is it better than a Valhalla reverb?
No signature here!

liquidsonics
KVRist
496 posts since 30 Jan, 2009 from UK

Re: Illusion - new reverb from LiquidSonics

Post Mon Oct 29, 2018 8:24 am

ghettosynth wrote:
Mon Oct 29, 2018 8:13 am
I like, for example, the smoothly modulatable moving microphones in Eventide's T-Verb. It doesn't matter if you can get that sound through other means, it has a usability factor that causes me to gravitate towards it when I want to move things about in a space.
Illusion has some great tools for playing with the reflectivity and proximity of the early reflections. The earlies have fusion modulation in them for subtle and quite natural randomisation of the reflections (unless you switch it off, and you can also drive it hard to cause chorusing if desired). These aren't standard point-delay reflections (which often cause accidental phasing when static or chorusing in modulation), they're very well decorrelated in a unique way that personally I think sounds great in both small and large room simulation scenarios. That's not going to satisfy your desire to move the sounds around in automation if you wanted to do that with interpolated delay lines, but I'm just pointing out a nice feature of the reflections in Illusion.
Matt
www.liquidsonics.com
Follow us on Facebook

Neon Breath
KVRAF
3098 posts since 15 Sep, 2010

Re: Illusion - new reverb from LiquidSonics

Post Mon Oct 29, 2018 1:33 pm

robotmonkey wrote:
Mon Oct 29, 2018 8:18 am
So... is it better than a Valhalla reverb?
:lol:

Forget it. Apples and oranges.
__________________________________________________
http://redfog-drone.bandcamp.com
http://soundcloud.com/sleeping-on-lotus-ashes
__________________________________________________

User avatar
bmrzycki
KVRAF
4088 posts since 11 Aug, 2006 from Austin, TX

Re: Illusion - new reverb from LiquidSonics

Post Mon Oct 29, 2018 1:45 pm

Product licensing is via iLok 2/3 (sold separately) or iLok computer based authorisation.
:(
Feel free to call me Brian.

simmo75
KVRian
1160 posts since 25 Mar, 2016 from Seattle

Re: Illusion - new reverb from LiquidSonics

Post Mon Oct 29, 2018 1:55 pm

There reverb king is back!
Wow, another beautiful reverb!

User avatar
plexuss
KVRAF
2573 posts since 8 Jul, 2009

Re: Illusion - new reverb from LiquidSonics

Post Mon Oct 29, 2018 2:06 pm

After spending the day with Illusion I decided to buy it. I think the algorythmic approach to Fusion-IR turned out well with this tool. It has a nice natural sound, fairly easy to tune in. I have Seventh Heaven Pro so I am used to the delay when making parameter changes. I need to spend more time learning the reverb so I can be effective with it but it has what I like to hear in a reverb.

I'd love to see a resizable GUI in both Illusion and SHP however. I struggle with the small size.

:phones:

User avatar
VariKusBrainZ
KVRAF
8738 posts since 16 Dec, 2002

Re: Illusion - new reverb from LiquidSonics

Post Mon Oct 29, 2018 2:08 pm

Neon Breath wrote:
Mon Oct 29, 2018 1:33 pm
robotmonkey wrote:
Mon Oct 29, 2018 8:18 am
So... is it better than a Valhalla reverb?
:lol:

Forget it. Apples and oranges.
And then divide those Apples and Oranges into quarters

simmo75
KVRian
1160 posts since 25 Mar, 2016 from Seattle

Re: Illusion - new reverb from LiquidSonics

Post Mon Oct 29, 2018 2:40 pm

plexuss wrote:
Mon Oct 29, 2018 2:06 pm
After spending the day with Illusion I decided to buy it. I think the algorythmic approach to Fusion-IR turned out well with this tool. It has a nice natural sound, fairly easy to tune in. I have Seventh Heaven Pro so I am used to the delay when making parameter changes. I need to spend more time learning the reverb so I can be effective with it but it has what I like to hear in a reverb.

I'd love to see a resizable GUI in both Illusion and SHP however. I struggle with the small size.

:phones:
I totally agree about the resizable GUI Pluxxus :)

User avatar
dangayle
KVRian
806 posts since 30 Mar, 2014

Re: Illusion - new reverb from LiquidSonics

Post Mon Oct 29, 2018 2:54 pm

bmrzycki wrote:
Mon Oct 29, 2018 1:45 pm
Product licensing is via iLok 2/3 (sold separately) or iLok computer based authorisation.
:(
At least it's got computer based auth instead of the dongle only like Seventh Heaven.

Return to “Effects”