KVR Audio

mansardian · Post by **mansardian** » Mon Sep 12, 2016 10:48 pm

hi poshook,

First, let me remind you that I didn't say Q-clone would be bad. I said I think Waves is overrated. That may be provocative to some I admit, so let me put in other words: I think many independent plugin devs are highly underrated.
It was you telling us that you weren't satisfied with Q-clone's results. I just made the suggestion to try out MMultibandconvolution. Must be one of the most powerful convo-plugins I've ever tried. Really. Give it a try. Not because it does some special magic. Its options are the special aspect about it.

Now something to think about: Do you believe Nebula would null the sampled hardware? (btw, I'm a fan of Nebulas concept!) Or to go even further: Do you think one hardware module completely nulls the module in the very next channel? If not, do you believe one is worse than the other? Hmm. Do you know the reason why the SMPTE time code used to live on the very last channel of a console? If you can answer that question you should know why the whole null-thing couldn't be more uninteresting.

I don't make plugins to null another plugin off. For such tasks I use the CUT-button of the regarding channel. (a lot easier and saves one track/insert)

To me it couldn't be less important if it nulls Nebula. That has never been my intention; I never tried to COPY Nebula. (That's pretty pointless, isn't it?) Instead, my intention was to create an algorithmic W295b that would sound familiar to me. (Yes, I know the REAL one, but as I said before, I cannot tug the studio console into my living room. That's why I chose Nebula as my reference which is pretty much of a salute to Nebula). And as long as you can't tell me which sound processor is here or there active in my blindtest I'd say the funny null game is not that important as you thought it would be.

PEACE my friend

camsr · Post by **camsr** » Tue Sep 13, 2016 12:59 am

Eq nonlinearities come in many different forms, null tests are pointless.

mansardian · Post by **mansardian** » Tue Sep 13, 2016 6:08 am

camsr wrote:Eq nonlinearities come in many different forms, null tests are pointless.

Absolutely .

Synthetic Wav · Post by **Synthetic Wav** » Tue Sep 13, 2016 7:23 am

thanks for the sound example it seems close, this sound example make it not easy to compare cause of the different hits, but maybe one is more sterile with harder transients and this would be reaktor probably,but really the sound example make it not easy to compare this so i m not really sure if it s real or in my head

it s usually one of the major difference between hardware/nebula and algo plugins , but the sound example really make it tricky to compare the transients for me

mansardian · Post by **mansardian** » Tue Sep 13, 2016 8:00 am

Synthetic Wav wrote:thanks for the sound example it seems close, this sound example make it not easy to compare cause of the different hits, but maybe one is more sterile with harder transients and this would be reaktor probably,but really the sound example make it not easy to compare this so i m not really sure if it s real or in my head
it s usually one of the major difference between hardware/nebula and algo plugins , but the sound example really make it tricky to compare the transients for me

Not bad, that's what I call a competent comment. You are right, the W295 has a special character: It softens the high transients a bit and this is something where Nebula shines. It captures that behavior.

That's actually what makes it sound so silky. In theory I could simulate that in Reaktor, too. However I'm not sure if I will kill the main advantage by doing so - CPU-efficiency. If it turns out to be pushing up the cpu-demand I think I better let Nebula do that job.
However, I don't see that necessarily as a weak part. It's more a matter of yin and yang, or Stan and Ollie. I like both of them.

Synthetic Wav · Post by **Synthetic Wav** » Tue Sep 13, 2016 10:15 am

yeah not a weak part, many will prefer it that way

zenvoxpop · Post by **zenvoxpop** » Thu Sep 15, 2016 11:54 pm

Btw whats the saturation stage like, is it pre or post eq?

mansardian · Post by **mansardian** » Sat Sep 17, 2016 11:37 am

zenvoxpop wrote:Btw whats the saturation stage like, is it pre or post eq?

According to the real device saturation is everywhere you drive the transformers hard. So if you feed the EQ with a way too hot signal the input stage will distort. the same is at the output stage. But yet I'm not sure if this EQ will benefit so much of audible saturation. It was designed to AVOID distortion. Don't take the W295b as a coloration tool.
That won't work.

For such tasks I want to recommend Sonimus' Burnley73. One of the best 1073s out there.

Cheers!

mansardian · Post by **mansardian** » Sat Sep 17, 2016 12:09 pm

I have to post something I just realized...

As I wrote before, my reference for constructing is Nebula's W295b. Just when I thought I was done I compared it with Nebula and got pretty confused: The curves didn't match anymore!!

I scratched my head and started to search for the mistake in my construction but couldn't find any. And then I found the only thing about Nebula that does.not.work. like the real hardware. In the tech specs of Siemens they say the W295b has a roll off from 15kHz upwards, as well as from 40Hz downwards. Alright. Now, if you want to have the full emulation of that EQ in your insert you need to load at least two instances of Nebula (Tim Petherick), or even three (Germanos EQ AITB).

Do you know what happens if you stack low pass filters? I guess you do. the roll off gets steeper and steeper. That has a major influence on the curves. And that is a real problem. From that point of view even the harmonics (call it color) are much too "present". It's like stacking three EQ modules just to process only one band with each of them.

No matter how great Nebula sounds, this is something where my EQ will be more authentic. The signal passes through the process only once.

That downside of Nebula is not an opinion of mine. I can see it on my analyzer. That doesn't mean Nebula cannot be used to wonderfully emulate hardware EQs. It means those hardware EQs must have a linear frequency response. Or you use only one instance per track.
Even the harmonics aren't authentic anymore if you need more than two.

focusrite · Post by **focusrite** » Sat Sep 17, 2016 12:45 pm

http://www.timpetherick.co.uk/dualtech-v2-sampling/

Alex b has clean presets to counteract this also

mansardian · Post by **mansardian** » Sat Sep 17, 2016 2:54 pm

focusrite wrote:http://www.timpetherick.co.uk/dualtech-v2-sampling/

Alex b has clean presets to counteract this also

Oh, okay, so you found a workaround for this...cool.

Is there a V2-sampled W295b, too? I couldn't figure that out. AlexB doesn't seem to have a Siemens- or perhaps he has but I couldn't decipher it as such.
Anyway, I think it's great you found a way around that issue

However, I still feel the need for a CPU-friendly, zero-latency solution

zenvoxpop · Post by **zenvoxpop** » Sat Sep 24, 2016 9:53 pm

So, how's the project coming along?

mansardian · Post by **mansardian** » Sat Sep 24, 2016 10:29 pm

zenvoxpop wrote:So, how's the project coming along?

Hi,
I had to deal with some problems regarding the midband. that was a tough piece of work. Looks like I solved the issue. Tomorrow (Sunday) I'll finish the final parameter work, then the last band (bass). Everytime I try it on a vocal recording I'm amazed how natural that EQ sounds. It can really bring dull recordings to life. Those guys at Siemens knew how to set curves

The next chapter will be to inject a little individual character into this thing. Means I'll add a preamp module that can be driven to taste. Not an original Siemens design but a custom gain stage which can be switched on and off to get 1. the original frequency response and clarity, as well as a very moderate CPU consumption when switched off and 2. a stronger individual sonic fingerprint and an abrasive effect on transients when switched on.
After all I want to deliver a professional grade product that shall become kind of a go-to tool if you need to pull the damping blanket off your tracks

camsr · Post by **camsr** » Sun Sep 25, 2016 2:53 am

will it work at 96khz?

Zaphod (giancarlo) · Post by **Zaphod (giancarlo)** » Sun Sep 25, 2016 8:46 am

mansardian wrote:I have to post something I just realized...

As I wrote before, my reference for constructing is Nebula's W295b. Just when I thought I was done I compared it with Nebula and got pretty confused: The curves didn't match anymore!!
I scratched my head and started to search for the mistake in my construction but couldn't find any. And then I found the only thing about Nebula that does.not.work. like the real hardware. In the tech specs of Siemens they say the W295b has a roll off from 15kHz upwards, as well as from 40Hz downwards. Alright. Now, if you want to have the full emulation of that EQ in your insert you need to load at least two instances of Nebula (Tim Petherick), or even three (Germanos EQ AITB).

Do you know what happens if you stack low pass filters? I guess you do. the roll off gets steeper and steeper. That has a major influence on the curves. And that is a real problem. From that point of view even the harmonics (call it color) are much too "present". It's like stacking three EQ modules just to process only one band with each of them. No matter how great Nebula sounds, this is something where my EQ will be more authentic. The signal passes through the process only once.

That downside of Nebula is not an opinion of mine. I can see it on my analyzer. That doesn't mean Nebula cannot be used to wonderfully emulate hardware EQs. It means those hardware EQs must have a linear frequency response. Or you use only one instance per track.
Even the harmonics aren't authentic anymore if you need more than two.

Acqua plugins fix it... Basically they apply harmonic distortion just once

Reaktor: W295b EQ-Preview