I am so confused right now^^
AW Hypersonic VST2 plugin shows 64bit float res. on WLs menue item & bit-meter, so its audio output is 64bit float?

hmmm. Sure AW and TDR are x64bit but to my understanding the menu item "resolution" in WL refers to the internal audio bit-depth.
So TDRs internal audio bit-deph runs at 64bit float but i/o is reduced to 32bit float, is this right? That means Wavelabs "Resolution" menue-item just refers to i/o res.?

I have set Wavelab to process audio & plugins at 64-bit float and the the plugin information & bit-meter shows 64bit float on all 64bit (resolution) plugins no matter if VST2 or -3.
Since WL 10 i prefer working fully at 64bit float because it sounds more open and smooth, especially the spatial information are clearly more defined to my ears. The ultimate goal was to achieve a 64bit float transparent signal path from the beginning until the final dither to 24bit without any dithering/truncation involved before.

I guess i miss something!?

Edit: I just tried to check some of the other free TDR plugins and they are all showing 64bit float (VST2/3) in WL11s but all VST2 versions failed to load and WL crashed. The VST3 versions are fine and the bit meter shows 64bit too.
But the demo of DeEdger VST2 for works fine and shows 64bit float on the plugin-menue & output bit -meter. :/

My further question is: is it possible to have a 64bit float i/o (bit-transparent) version of Ultrasonic to avoid the internal dithering in WL at the plugin input?

What Wavelab is showing is not the internal bit depth of the plugin. It can’t know that and is only able to see what it spits out based on the 64bit input. Just because a plug-in is outputting 32bit, doesn’t mean it isn’t processing at something else internally. As far as I’m aware, there is no dithering from 64bit float to 32bit float. It doesn’t work the same as with fixed point bit depth.

okay i sould have said "conversion".
So the conversion from 64bit float to 32bit float is without any loss auf audio quality?

Again, if i set up a chain of 64bit float plugins where the bit meter is showing 64bit float on all plugins i get tonally different results if WL is set up to 64bit plugin processing vs 32bit processing in the preferences.

Of course there is a loss, something gets thrown away.
But this doesn't affect audio quality in a negative way, quite the opposite.

If the plugin does nothing, the process is bit transparent.
If the plugin does something, the higher internal precision will certainly never cause additional issues.

As mentioned above, this is a floating point truncation. With music signals, it practically dithers itself (because the truncation is largely chaotic, thus doesn't create much patterns.).


Ultrasonic uses old code and never supported VST3 or the 64bit VST2 extension in the first place. But the forthcoming update will of course.

Again, any "plugin bus" beyond 32bit is largely esoteric, there are no problems that higher precisions would solve. Feel free to try! 64bit I/O exist because it is becoming "free" due to super powerful CPUs and plenty memory. And Steinberg thought it makes a cool and lasting marketing move.

It looks cool in plugindoctor, but really is irrelevant and out of scale. Inside the plugin it could possibly matter, especially with recursive algorithms (calculations reusing previous, usually diminishing results), such as certain filters or naive approaches.

Juggling between computer music and mastering (mostly electronic music) myself, I’d love to find an optimal way to band-limit my softsynths too. For my 100% ITB synthesized / engineered music, 96kHz sounds often do much better.

And an updated Ultrasonic filter would be lovely in my mastering session too, obviously.

We're working on an update.

I know ! That was some clumsy yet honest enthusiasm from me here eh

Great news.

Thanks @ Ghost Cable for getting the ball rolling

Fabien,

Do spline curves offer anything to antialiasing?

The free Moss EQ can make sharp vertical cuts, looking like a sword on the spectrum, unfortunately stopping at 20kHz.

That might be a little too close for comfort based on what I have experienced. I like the TDR default because using a few instances of it in a single chain remains out of ear shot, all while doing repeated and effective cleanups (if needed).

How about CLAP?

This stuff is subject to plugin-host politics, it's them shaping the standards. We don't have the resources to play around with experimental plugin formats, and don't feel much joy doing such type of work anyway! (they all more or less do the same, and we plugin devs usually only focus on the canonical features anyway, i.e. what the formats have in common). Once a format becomes widely supported and worth the effort, we are happy to support it, be it 32bit I/O or 64bit I/O.

Afaik, CLAP is interesting for instruments, synths in particular, much less so for processors.

I just wanted to know if CLAP also used 64-Bit I/O, that's all.

Yes, most probably.

Any opinions on Pro-Q3's brickwall LPF for ultrasonic duty?

No experience with the TDR filter... will look forward to trying it when it's rereleased.

However with that said... The AirWindows Hypersonic is amazing. I prefer to only use one instance of it though. At the beginning of my final limiting chain.

EDIT:

AW UltraSonic causes pops between tracks when playing back in WaveLab

Looking forward to this

The Ultrasonic remake inspired us to experiment with a radically new UI approach, almost fully driven by the GPU (making it ready for the next 30 years of "super high res retina mega plus displays", and also greatly relaxing the CPU side).

We also decided to develop whole series of plugins following the new paradigm. Ultrasonic will be one of the first releases, with a half dozen more following in short intervals.

Any (rough) idea when ...?

Likely September, but take it with a grain of salt

That's exciting news. Are the 6 plugins various utility type or updates to the existing lineup? Or completely new tools?

These will be new products. In case of Ultrasonic we finally invested the time to make it a proper product (the first Ultrasonic was just a proof of concept, purely made for the original GS thread).
While working on it we realized the same UI concept could be reused to implement a series of "little" ideas in our books.

Most notably, a proper spectrum analyzer, several special use filter/utility plugins in line with Ultrasonic, such as a extended version of the relative HF limiter found in Limiter 6 and more. All very focussed concepts offering unique/exotic twists in a modern UI.

We also made progress on a special-FX device, but unrelated to this little series.

Great stuff Fabien! I’m excited for the evolution of your already great HF limiter.

The HF Limiter with more control would be fantastic I bet.

The relative mode is already very good, stays clean in Insane. More control over the ballistics of reduction and a sidechain eq would add a lot.

Looking forward to what you folks cook up in the labs.

Mental !

I just discovered this thread and I already want the end of summer ASAP.

I would be extatic if TDR could make an advanced transient shaper plugin for exemple.
I've got quite a lot of them, and I'm never fully satisfied by their sound and features.
Regarding advanced features, the idea behind Transgressor for exemple is great but there are many artifacts and strange things happening in the sound that I don't like (and it's not updated since ages). I never used SplitEQ or Crave transient EQ, but I own Physion and never use it because of the sound of it (I don't speak about the fx, but about the tonal/transient separation), so i'm not that confident with the above mentionned plugins.
Newfangled Punctuate is great (and the idea behind Elevate, with the transient shaper/clipper combo) but is too cumbersome and has too much latency.

Actually, the one that sounds the best (or better said, the most natural) is Flash, offered for free by Wavesfactory, but it's very basic feature wise.

I'm pretty sure you could make something great combining clipper, EQ, transient shaper and filter, with amazing quality and great ergonomy.
A kind of Limiter 6 for transients.

Split EQ and Stillwell Transient Monster do zero harm to a signal. They both sound great on all genres of music.
TDR dropped progressive plugs like Simulathe and you want rehashed plugs because you haven't demo'd whats already out there? Let them continue to be progressive.

Split can get a bit artefacty but i do like it, otherwise Spiff gets the most use.

Transient control is an interesting field, though. The specific series mentioned above is about things that have a spectrum (special purpose filter, specialized dynamic EQs, analysers). Transients are rather wideband, so transients control is not part of this series, at least not directly. Working on analyser related topics, we learned a great deal about tone vs noise separation, an important aspect in getting transient designers to work properly (because tonal vs noisy signal need very different handling). So we have ideas in this direction of course, but no worthy offer yet.

The problem really sits in the transient detection itself, it's very difficult to solve. About "perfect transparency" in splitter concepts: You can split a signal in any way you like, even really aggressively, and reconstruct it perfectly. It could be a monkey riding a crossfader, splitting the signal onto 2 channels, somehow. You can add them fine and end up with the original. The problems arise once you start processing each channel differently and try to recombine them. Then the quality of the detection starts to matter.