Cytomic "The Drop" Resonant Filter

[meloco_go]

I'd buy THAT for a dollar!

How's about just go for gold and get a whole DAW going!!?

The DAW is crazy, but creating something like virtual console with basic EQ/dynamics and full interaction between channels (or at least analog modelled summing amp) would be great (I do have and like VCC but IMO they stopped one step too early). I guess it is almost impossible to code as VST/AU so maybe standalone app with ReWire ins/outs.
Oh, too much dreaming...=)

Anyway, EQ and another comp make me want to hybernate for whatever time it would take Andy to finish them!

[Vectorman]

Andy, puhleez an EQ first!

Hehe. The synth will be down the line a bit, I want to write an EQ and another compressor before that.

I gather that with the Cytomic name on it, the EQ will involve circuit modeling and will be anything but a bog-standard digital EQ. Sounds interesting.

I can't help but wonder if you're thinking of putting off a new synth because, based on your work on The Drop, you know how CPU-demanding it's going to be to create a full synth architecture with that level of modeling detail and you're not sure if even current i7 CPU's are quite up to the task yet. Is that a factor at all, or is it purely a matter of what you feel most inspired to work on next?

[Crackbaby]

As i've said to a couple of developers here before ........ start ... working ... on ... a ... cloning ... machine ... ! That daw could be a reality, tomorrow! With several included modeled synths, effects and a giant console!


Just sayin'!!!

[metrosonic]

That's actually the first comment I've read from a Live9 beta tester. Any other opinions so far - the first tester seats were taken very quickly.

Apologies for not getting back to you on this, I've not checked into KVR for 24 hr's or so.

I can't give too much away ref the Live 9 beta as Ableton have asked the following of their beta testers hence the relative silence outside of the beta forum (and I think it's a very sensible thing to do):

"- Since Live 9 is still in beta and is subject to change, please do not post any reviews or make other definitive, public statements about it.

- Please avoid speaking in-depth about the Live 9 beta on the general Ableton forum. We understand that everyone has questions, but we are trying very hard to avoid spreading misinformation.

- Please carefully consider what you are saying publicly about having access to the beta. We understand that you're excited - and we are too - but on the outside, posting about your access to the beta can look like bragging, which annoys other users who don't yet have access

So, in general, please don't go volunteering information beyond what has been publicly stated on ableton.com. If in doubt, feel free to say that you're unsure whether a certain feature will change."

I only made reference to EQ 8 as it's a Cytomic design. But overall Id have to say I'm very impressed with the beta. It's still early days but Live 9 is shaping up to be a very solid release.

JM
------------
https://soundcloud.com/leftside-wobble

[Crackbaby]

Website down, trial expired
Can i download the non-trial now or will it be trial during all beta?

[David Else]

Regarding new projects, i have to say i don't think the world needs another digital eq unless it is some kind of revelation!

I have tried them all and nothing touches nebula for eq. Its like all the other developers are just avoiding the elephant in the room, the volterra kernel technology just sounds much better. The strange thing is i have NEVER heard anyone argue about this. Why would anyone who cares about sound choose an inferior sounding eq?!

It's a choice between:

1. 'normal digital' eq (ie all sounds the same but with different curves and stick on a saturation algorithm).
2. convolution based eq. Sounds a bit better but loses flexibility.
3. volterra kernel/nebula. Loses flexibility but blows others away for sound.

The only reason i can see anyone would bother to develop another eq would be if they could extend the nebula technology and somehow make it more flexible. I don't think there are any patents on volterra kernels. A nebula technology based EQ with full control over 10 bands at once, now THAT would be a new eq worth making

[realmarco]

problem is with convolution is that u need to have a test tone to be emitted and a mic to records the room's reverberation and then its just a matter of blending the pure tone with the recorded tone with DSP-Math plugin.

However with Dynamic processors(compressors,limiters etc)and EQs need to sample each little knob twist,

lets say a hardware compressor with 3 knobs each with static knob settings(click 1,2,3)

sample knob 1 settting 1 : impulse recorded
twist knob 1 setting 2 :impulse recorded
knob 1 setting 3 : impulse recorded

3X3X3 times of Impulse files
thats with only a old "classic" harware units. and we havent tried Over-driving the unit

when u come across modern compressors and EQ that have normal knobs, u can then understand how complicated and time-consuming it gets. and How much space of HD it takes.

Focusrite put out an external hardware unit called "liquid mix"
http://www.soundonsound.com/sos/jul06/a ... uidmix.htm
Odds are Focusrite did not take their reputation lightly so the product did what it said on the box

However with Plugin devs, its improbable they have the rescources(money) to do something like that.

[andy-cytomic]

Website down, trial expired
Can i download the non-trial now or will it be trial during all beta?

The web page is back up and running. My web hosting service is having a lot of trouble at the moment in the aftermath of the New York storms, sorry about the outages.

[Funkybot's Evil Twin]

The only reason i can see anyone would bother to develop another eq would be if they could extend the nebula technology and somehow make it more flexible.

"We don't need another filter. Filter technology is essentially the same from plug-in to plug-in with just a saturation built in, or different curves, etc." People said crap like that all the time.

Then The Drop beta came out. The bar has been raised considerably (DIVA's filters were also a huge improvement over existing models IMO). The Drop is every bit as good as analog hardware when it comes to the sound, and much improved compared to hardware when it comes to workflow, additional features, the lack of maintenance, costs, and instance counts.

Why do you think Andy couldn't do the same thing with an EQ?

There's always room for the technology to improve.

[andy-cytomic]

Regarding new projects, i have to say i don't think the world needs another digital eq unless it is some kind of revelation!

I have tried them all and nothing touches nebula for eq. Its like all the other developers are just avoiding the elephant in the room, the volterra kernel technology just sounds much better. The strange thing is i have NEVER heard anyone argue about this. Why would anyone who cares about sound choose an inferior sounding eq?!

It's a choice between:

1. 'normal digital' eq (ie all sounds the same but with different curves and stick on a saturation algorithm).
2. convolution based eq. Sounds a bit better but loses flexibility.
3. volterra kernel/nebula. Loses flexibility but blows others away for sound.

The only reason i can see anyone would bother to develop another eq would be if they could extend the nebula technology and somehow make it more flexible. I don't think there are any patents on volterra kernels. A nebula technology based EQ with full control over 10 bands at once, now THAT would be a new eq worth making

Now I have nothing against the Nebula plugin, or dynamic convolution, and you are obviously happy with the results, so please keep using it and continue being happy.

I am assuming you mean you've tried lots of EQs, but perhaps you can let me know if you have tried this one?

http://www.soundonsound.com/sos/jul10/a ... assive.htm

Frederick Norén has very good ears, and a very good listening setup, and although I've not tried the UAD plugin myself, I believe what is written in his review. One part of the review also to note:

"...a big part of it is emulating the characteristics of the inductors, which not only can saturate, but also display hysteresis. Given its complex behaviour and non-linearity, hysteresis cannot be simulated with dynamic convolution."

Now Frederick is right here, but perhaps this statement is a little confusing, since it's not just the non-linear inductors that are the problem, but also how they are placed in the circuit as well that can cause even more problems if a full non-linear analog model is not done to emulate the circuit.

Almost every good sounding analog circuit have one thing in common, multiple nearly instantaneous feedback loops with non-linearities and some sort of time dependence in them like hysteresis, capacitors, inductors, and lots of the time combinations of all these things.

In practical terms this means that it is very difficult if not impossible to characterise and reproduce the full sound of one of these types of analog circuits via black box methods, unless the circuit is in its very linear area of behaviour. This is why methods using dynamic convolution kernels are better at subtle things like almost linear EQ and small amounts of compression, but can't handle things like screaming self oscillating filters with drive and compressors pushed hard. Especially of note is that large transients momentarily push circuits into their non-linear region for small amounts of time, even if you are otherwise operating in an almost linear region.

But don't take my word for it, please go ahead and ask the Nebula guys themselves how they feel the volterra kernel black box method stacks up to a full non-linear analog model like The Glue, or the UAD Massive Passive EQ. In limited situations I can see the usefulness of the volterra kernel or other dynamic kernel approaches, but I am more interested in reproducing the full range of behaviour of analog circuits. I love pushing them hard since they ooze tone, and even if you don't push things hard the transients in your material will.

So although you have listed some of the options your list is not complete. When I do an EQ it will be a full non-linear circuit model, just like all my other plugins.

[K-Slash]

Andy,

You're the man !

I can't wait for your other plugins.

You're one of the few that I trust to be able to do a faithful analog model of high-end tube mastering gear.

If you give a go an an all-tube device like the Fairman TMEQ and Fairman TMC or Manley Vari-Mu like compressor, I'll be the first to drop the cash !

I really like your "no-compromise/I'll do the best I can/math can get it done/stop wishing ultra-light CPU usage as you now it's gonna f**k the sound" approach.

For example, I take the White 2A and Black 76 from IK. They did a great job at modeling the overall compressor behaviour (at least to my ears), but failed at modeling all the non-linearities that add the trademark character of the units, even if at the time they were released, they had the modeling experience to do it fully.

I mean, what the point of having for example an ALL-buttons mode, that is famous for getting gritty and adding nice saturation/distortion/clipping and such, if you don't model the actual output distortion ?

Hope that makes sense, almost 5 AM here and I'm tired. Time to go sleep and dream about the wonder virtual gear coming next !

PS : I wish I could win Euromillion so I could fund all your plugin projects .

[realmarco]

interesting though the liquid-mix is meant to emulate the hardware on the market

not just become another compressor that sounds good.

Analog EQ and Compression does its thing with electricity, Digital Adjust Bit-depth.

a Drumloop in 24 bit going thru a hardware Analog compressor will still be in 24 bit.

a Digital one will degrade the dynamic range, on extreme setting you can turn a 24 bit audio file into a 12 bit audio file

am I wrong ?

[andy-cytomic]

interesting though the liquid-mix is meant to emulate the hardware on the market

not just become another compressor that sounds good.

I have never used a liquid mix in person. I did do their online blind test using headphones in an open plan office in central London some time again, and could pick the difference between the analog and digital with 85% success rate.

Many developers claim to emulate hardware, usually something along the lines of "every nuance captured" or "every detail has been meticulously modeled" or "replicate precisely the sound of" and many variations on that theme, that's marketing for you. Of course they are trying their best to achieve this goal, they just have varying levels of success.

Analog EQ and Compression does its thing with electricity, Digital Adjust Bit-depth.

a Drumloop in 24 bit going thru a hardware Analog compressor will still be in 24 bit.

a Digital one will degrade the dynamic range, on extreme setting you can turn a 24 bit audio file into a 12 bit audio file

am I wrong ?

I don't know where to begin here since these statements are pretty vague and I can't really tell what you mean, sorry about that.

If you record a signal with a decent pre and have a low signal to noise ratio then there is nothing wrong with then taking this full 24-bit audio file and doing further processing on it, including large amounts of dynamics. Both in the box processing and external analog units that process this signal are in an equivalent position, neither can magically create more resolution out of thin air. If you take a 24-bit file and then gain it up by 72 dB then you will have to throw away the top 12-bits and you will be left with a 12-bit signal, it doesn't matter if the gain is done in digital or with ultra clean and perfectly linear analog, the results will be the same.

[meloco_go]

But don't take my word for it, please go ahead and ask the Nebula guys themselves how they feel the volterra kernel black box method stacks up to a full non-linear analog model like The Glue, or the UAD Massive Passive EQ. In limited situations I can see the usefulness of the volterra kernel or other dynamic kernel approaches, but I am more interested in reproducing the full range of behaviour of analog circuits. I love pushing them hard since they ooze tone, and even if you don't push things hard the transients in your material will.

Andy, I don't have UAD, so can't compare but I have Nebula, both Glue and Drop and a bunch of plugin EQs. With that experience I can relate to what thermal wrote. There is something lacking with common plugin EQs, and Nebula just makes EQing more rewarding. So, I don't think Nebula can do something like Neve line level to mic input distortion trick, but even without that it's great.
But with that I have very high hopes for your EQ plugin, cause Nebula is PITA to use and if I stuck up too much instances the project becomes unresponsive because of latency etc.

So although you have listed some of the options your list is not complete. When I do an EQ it will be a full non-linear circuit model, just like all my other plugins.

I'm really looking forward for it.

I have a question though, do you plan to make exact 1:1 model of a particular circuit or do you plan to be more flexible? E.g. a model of, say 1073 with filter circuit from SSL?

[David Else]

So although you have listed some of the options your list is not complete. When I do an EQ it will be a full non-linear circuit model, just like all my other plugins.

Thanks for the reply Andy. I can't try UAD Massive Passive as i don't have the hardware, and you said you had not tried it either, so not a good point of reference

I am not a DSP programmer, so can't really join the discussion on the technical side. Nebula does indeed have major limitations, i don't know if they can be solved.

What i can say for sure is that i have been mastering commercially for 10 years and know what i am listening to. So far every single EQ plugin claims this and that about how great they sound, and **in my opinion in comparison to a nebula eq they don't**. Its all marketing, same flat and sterile sound.

I really hope you can be the first developer to truly better nebula for EQ's, it would be fantastic and i would be your first customer. My point is that it is going to take something **new** to do it. Not just more of the same. Nebula was a new technology, how about the next one?

I did get involved in a forum shootout between the glue/hardware/alexb nebula emulation, and heard your website demo. My conclusion was while the glue did handle the dynamics well, it didn't have the 'dimension' of the hardware. It was flat, the hardware was alive and 3d. It is this extra dimension that is missing in all emulations other than Nebula. Nebula could not actually do the dynamics properly in this example, it emulated the 'dimension' but f***ed up the actual dynamics part. We need both.

Also, to be honest, it's only a small percent of your customers are going to hear the 'dimension' difference, it is subtle, but it's also very important for those people who do hear it. I don't want to sound arrogant, it's taken me 10 years of very intense listening to pick up on this stuff. Its all very subjective, I can't tell someone what to hear, I can only say what i hear.

Nebula has some major problems for workflow and limitations in what it can do, i don't want to come over as a fanboy. I look forward to checking out the drop soon, and any future eq.