Coming Soon: The Relab 176 (Retro 175) Tube Compressor/ Limiter

[jens]

Strangely enough my current GoTo compressor has neither here nor over at Gearspace mentioned as an alternative yet. It's a VariMu hardware emulation as well and all the descriptions of the Relab 176 (huge sweet spot, makes everything sound better, dead-easy to dial in) apply to it too - plus it has a few nice tricks up its sleeves and can often be had for 50% off.

People love hype, not gear - which is why I won't say what it is.

[audiouser720]

Strangely enough my current GoTo compressor has neither here nor over at Gearspace mentioned as an alternative yet. It's a VariMu hardware emulation as well and all the descriptions of the Relab 176 (huge sweet spot, makes everything sound better, dead-easy to dial in) apply to it too - plus it has a few nice tricks up its sleeves and can often be had for 50% off.

People love hype, not gear - which is why I won't say what it is.

What is it?

[jens]

huge sweet spot, makes everything sound better, dead-easy to dial in

forgot to mention: quite versatile

[jens]

What is it?

It's an officially sanctioned emulation of a highend boutique stereo vari-mu compressor with an absolutely brilliant and gorgeous GUI.

That it's my GoTo doesn't mean I'm not also sometimes owning&using the UAD 176 by the way - that one is really nice too imo.

[Ploki]

The hype about this one isn’t whether it sounds “better” - that’s always subjective.

The hype is about authenticity, and it’s completely justified. UAD one is a joke in comparison.

So, is it better in every context? Likely not, no single unit - sw or hw - is, ever.

Is it the most versatile compressor? No, not in the slightest.

Is it behaving more like HW than any other compressor plugin? Absolutely.

I hope enough people recognise that, so developers stop wrapping soft clippers into fancy gui and call it emulation.

If you’re enthusiastic about audio, the least you can is appreciate the level of detail and meticulousness that went into coding this one.

I guess “simulation” is a better descriptor for the Relab one anyway.

I’m more cynical than not about new plugins which anyone who came across me here or over GS should already know.

This is really good work, and i appreciate good work when i see it

[El°HYM]

I hope enough people recognise that, so developers stop wrapping soft clippers into fancy gui and call it emulation.

This might not age too well on kvr.

[jens]

The hype about this one isn’t whether it sounds “better” - that’s always subjective.

The hype is about authenticity, and it’s completely justified.

And yet someone @Gearspace managed to correctly identifiy the cuts between the hardware and the plugin in the first AB-example Relab posted. Someone from Relab then said it's like the difference between two hardware devices, yet since they AB-ed with the actual hardware they emulated, I don't get this explanation. So no, it's not a perfect 100% emulation, it's a close approximation -
like a lot of other plugins - most of which use a lot less CPU though.


It's a theoretical possibility that the method Relab uses here isn't that much more accurate than other techniques, however a lot less practical for realtime-useage, which is why e.g. Andy from Cytomic isn't using it.

[Ploki]

The hype about this one isn’t whether it sounds “better” - that’s always subjective.

The hype is about authenticity, and it’s completely justified.

And yet someone @Gearspace managed to correctly identifiy the cuts between the hardware and the plugin in the first AB-example Relab posted. Someone from Relab then said it's like the difference between two hardware devices, yet since they AB-ed with the actual hardware they emulated, I don't get this explanation. So no, it's not a perfect 100% emulation, it's a close approximation -
like a lot of other plugins - most of which use a lot less CPU though.


It's a theoretical possibility that the method Relab uses here isn't that much more accurate than other techniques, however a lot less practical for realtime-useage, which is why e.g. Andy from Cytomic isn't using it.

Yeah that someone was me lol.

And I stand by what i said.

Also, as far as i know cytomic uses a really similar method (also u-he)

Edit: And i agree with Relab’s explanation. If you ever toyed with tube gear - this sort of difference can easily happen, this thing has been in development for more than a year. Tubes can be finicky

[audiouser720]

There will be a time when people will desire the sound of vintage 2010- plugins with wrapped soft clippers into fancy gui

EDIT: my secret weapon is already early 2000 vst. They sound more raw, more vintage edge, more like real hardware. The complex component modelled plugins have no soul. They too perfect, too sterile

[Ploki]

There will be a time when people will desire the sound of vintage 2010- plugins with wrapped soft clippers into fancy gui

I work with a producer that swears by Dverb since pt5, so yeah, probably.

Some of old plugins have a specific sound that’s been imprinted into countless records (i.e decapitator or waves L2) despite technical shortcomings.

[audiouser720]

There will be a time when people will desire the sound of vintage 2010- plugins with wrapped soft clippers into fancy gui

I work with a producer that swears by Dverb since pt5, so yeah, probably.

Some of old plugins have a specific sound that’s been imprinted into countless records (i.e decapitator or waves L2) despite technical shortcomings.

Jokes aside, old Waves is technically better than newer ones

[kraster]

The hype about this one isn’t whether it sounds “better” - that’s always subjective.

The hype is about authenticity, and it’s completely justified.

And yet someone @Gearspace managed to correctly identifiy the cuts between the hardware and the plugin in the first AB-example Relab posted. Someone from Relab then said it's like the difference between two hardware devices, yet since they AB-ed with the actual hardware they emulated, I don't get this explanation. So no, it's not a perfect 100% emulation, it's a close approximation -
like a lot of other plugins - most of which use a lot less CPU though.


It's a theoretical possibility that the method Relab uses here isn't that much more accurate than other techniques, however a lot less practical for realtime-useage, which is why e.g. Andy from Cytomic isn't using it.

That kind of design can change sound from day to day. Thermal factors, voltage fluctuations on the power grid etc.

Also while the Relab version is based on their hardware unit they’ve idealised some components because why not?

For example, an ideal triode should be symmetrical but generally they are slightly asymmetrical. Relab’s tube model is like the 1 in a thousand tube that’s perfectly symmetrical.

[Ploki]

That kind of design can change sound from day to day. Thermal factors, voltage fluctuations on the power grid etc.

Also while the Relab version is based on their hardware unit they’ve idealised some components because why not?

For example, an ideal triode should be symmetrical but generally they are slightly asymmetrical. Relab’s tube model is like the 1 in a thousand tube that’s perfectly symmetrical.

For real, i caught the slight frequency difference in the very high range (something i observed with tube gear i either own or have used in the past) but the difference is within “what if i swap a tube”.

Then you peek the spectrograph and spot the non-thru-zero cuts

If the slight frequency difference was fixed (and it can be adjusted so easily) and cuts through zero crossing, there is absolutely nothing in the a/b file that can be used as an identifier. It’s uncanny. And the frequency profile difference is really in a narrow region between 10-14k and can easily be observed with tube swapping on analog gear, and is completely inconsequential for the accuracy of the model in my opinion. Really, which was HW and which was SW was a lucky guess and nothing more.

I was swapping tubes in my gemini like two weeks ago, and i had everything from NOS sylvania, old philips and siemens, old nonames, new JJs etc, and the differences were far more dramatic than what’s in this A/B.

[kraster]

That kind of design can change sound from day to day. Thermal factors, voltage fluctuations on the power grid etc.

Also while the Relab version is based on their hardware unit they’ve idealised some components because why not?

For example, an ideal triode should be symmetrical but generally they are slightly asymmetrical. Relab’s tube model is like the 1 in a thousand tube that’s perfectly symmetrical.

For real, i caught the slight frequency difference in the very high range (something i observed with tube gear i either own or have used in the past) but the difference is within “what if i swap a tube”.

Then you peek the spectrograph and spot the non-thru-zero cuts

If the slight frequency difference was fixed (and it can be adjusted so easily) and cuts through zero crossing, there is absolutely nothing in the a/b file that can be used as an identifier. It’s uncanny. And the frequency profile difference is really in a narrow region between 10-14k and can easily be observed with tube swapping on analog gear, and is completely inconsequential for the accuracy of the model in my opinion. Really, which was HW and which was SW was a lucky guess and nothing more.

I was swapping tubes in my gemini like two weeks ago, and i had everything from NOS sylvania, old philips and siemens, old nonames, new JJs etc, and the differences were far more dramatic than what’s in this A/B.

You were also the only one among at least a hundred that could reliably tell the difference.

So for anyone to claim that it was somehow "easy" has a very different definition of easy to me.

I have a Manley Vari-MU with a T-bar mod and that never sounds exactly the same from day to day.

If I ran a mix through it on different days and phase flipped them they wouldn't null if I phase flipped them. They'd be very similar but not the same.

[Funkybot's Evil Twin]

I was swapping tubes in my gemini like two weeks ago, and i had everything from NOS sylvania, old philips and siemens, old nonames, new JJs etc, and the differences were far more dramatic than what’s in this A/B.

Exactly this! Just this week I put Russian NOS 6V6s in my 5E3 clone. The day before I had done a capture of the amp for ToneX. I didn't change the settings and hadn't unplugged the capture setup, so I thought, let me do another capture of the amp with the new tubes to see if I can hear a difference.

There was definitely a bigger difference than I heard with the 176 A/B.