Sonalksis Essentials Love

[rydan]

I have to be completely honest though... I'm blown away that you don't like the sound quality.

Sorry, I think this is a misunderstanding. I do like the sound quality! If I didn't already have the Cambridge, I would probably have bought the 517 instead. As I said before, they are pretty similar in character and features, the top end in the 517 is slightly better than Cambridge i think, but to my ears it sounds just a veeery slight bit more grainy than the UAD 1073 (which oversamples, and eats 1/3 of a card by the way, so yes, oversamling is quite expensive CPU-wise...). However, since I have the Cambridge, the 1073, the UAD PultecPro and also Elemental Audio Firium (and since I'm not making any serious money from my music right now), I can't really justify spending money on another eq, my EQ:ing needs are pretty much filled. Hope that cleared things upp a bit...

Regards

/Patrik

[DaveSonalksis]

I agree that the fidelity of the sonalksis eq needs to be upgraded, i plan to buy it and the compressor in the future, but the fidelity is what kept it from being a must have to me when i demo'd it.The demo struck me as something i'd buy later but if more fidelity was there i would have bought it then instead of going elsewhere.

Hmm... This sounds like something I need to investigate... can anyone give me a way to measure "the fidelity of the eq"? If I can find a way to measure what's wrong, I'll get it fixed!
Fidelity .... to what, precisely, are we seeking here? Textbook analogue curves?

[DaveSonalksis]

voxengo have oversampling in some of their eqs. algorithmix have it, universal audio (pultec eq, precision eq, neve eq), refined audiometrics, massenburg.

I stick with eliosound's comment on oversampling...

as far as advanced transform technique:
http://www.eliosound.com/amlt
imho this is closer to analogue units (more symmetrical) than the sonalksis eq at 44k, i compared them.

This is VEEEEERY interesting... Hmm... certainly, up against the other "decramped" EQs on that list we stand up very well... Like I said, the Orfanidis stuff is textbook compared to what the Sonalksis guys invented.
I'm not 100% sure about their measurement of the shelves. This is some very meaty business here I'm going to investigate this!

THANKS for linking me!

Dave.

[DaveSonalksis]

voxengo have oversampling in some of their eqs. algorithmix have it, universal audio (pultec eq, precision eq, neve eq), refined audiometrics, massenburg.

I stick with eliosound's comment on oversampling...

as far as advanced transform technique:
http://www.eliosound.com/amlt
imho this is closer to analogue units (more symmetrical) than the sonalksis eq at 44k, i compared them.

This is VEEEEERY interesting... Hmm... certainly, up against the other "decramped" EQs on that list we stand up very well... Like I said, the Orfanidis stuff is textbook compared to what the Sonalksis guys invented.
I'm not 100% sure about their measurement of the shelves. This is some very meaty business here I'm going to investigate this!

THANKS for linking me!

Dave.

Yes, ok. I've figured out how it works, and I can confirm that it is a BRILLIANT technique... and it WILL sound as good as the Analogue 'original'. In fact, it's as close to perfect as a technique can get, and as close as the developer wants it to be. AirEQ is the real thing people
I'm curious as to whether the guy from eliosound ever worked at Digidesign...
Btw, his write-up for d2 is wrong. If he inspects a little more closely, he'll find that d2 is Analogue like curves... Well, so the guy that made it showed me
But that was many many years ago, and there's more cpu around now, and Forte and AirEQ are closer to Analogue than d2.

This technique appears to be almost identical to that which I used on the Forte Suite plugins.
No. I'm not going to reveal details.

Wow. Fabrice Eliosound, you are a legend. I agree, this technique is a challenge. It took me three months with a very fast computer for Forte!

The harsh limitation of this technique is that it's an ENORMOUS pain in the ass to do... and there's a process in the step which is incredibly complicated, and very very weird... that took months to solve... but then, I never had Mathematica back then...

Alas, it's a real pain to add new filter types... ENORMOUSLY painful, (but certainly not impossible).

In a plugin, this technique lacks the elegance of the orfanidis technique... or the crazy stuff in 517 (which still scares me)... although the process of implementing it is as graceful and elegant a mathematical process as one could hope for.

I want to do a new EQ plugins now!!!!!
AAAAGH

Ugh. I have to sort out the fader cap on FreeG, and adjust some things by 1 pixel, and make installers.
BLEH.

I wish i was making EQ plugins.

AirEQ is cool. Loads of people seem to really really love that it doesn't have a graph.
Well. I hear visually. When I go to EQ something, I can SEE the EQ shape that I want to apply to it.
So, mine is gonna have a graph. And 5 bands. And LOADS of filter types.
Maybe I could model them on SSLs and Neves... OOOH! THIS IS EXCITING!!!!

[beej]

Dave - you're a geek.

Welcome!

[bmanic]

I also think it could be worth to improve the 315 in this terms, too. I always found that it works extremly well as one expect from a compressor, but it seems to flatten things a bit in the meaning of loosing depth or spatiality (not dynamics of course), like a slightly milky glass or so, hard to explain... a problem with most native digital compressors (those available on mac at least), not so much with the bombfactory on protools or real analog ones. Maybe it's just my incompetence in using compressors, but maybe others feel it same. )

Well, I don't agree with that at all. The SV-315 compressor is, IMHO, perfect. And comparing it to bombfactory is like.. .. .. weird.

Bombfactory = early 90's plugin technology. Quite 'accurate' transfer curve. Accurate 'sound'? Nope, not even close.

Cheers!
bManic

[bmanic]

I agree that the fidelity of the sonalksis eq needs to be upgraded, i plan to buy it and the compressor in the future, but the fidelity is what kept it from being a must have to me when i demo'd it.The demo struck me as something i'd buy later but if more fidelity was there i would have bought it then instead of going elsewhere.

Hmm... This sounds like something I need to investigate... can anyone give me a way to measure "the fidelity of the eq"? If I can find a way to measure what's wrong, I'll get it fixed!
Fidelity .... to what, precisely, are we seeking here? Textbook analogue curves?

No, fidelity, as in how the transients survive.

There was a rather good thread about EQ fidelity here a while ago. I'll dig it up for you.

Also, like people have already said, Voxengo, Massenburg, Algorithmix, PSP etc. all use oversampling and it DOES sound better. Sure, I'm all to aware of the problems with upsampling/downsampling (That's why I bought Voxengo R8brain Pro) and it is a compromise, but when done right, it really does improve the sound a lot. Besides, you can always make it a toggle option as a button on the GUI and leave the default to off (like voxengo, which even has an "auto" option where oversampling is turned on only during rendering).

I'll try to post some specific Sonalksis EQ related problems and compare to some other free and prototype plugins that I have.

Last but not least: Sometimes it's good to leave the "programmer" in you at home and use the "enginer" aspect a bit more. At least while listening.

I'll prolly be shot for saying that..

Cheers!
bManic

PS. don't take the critique the wrong way. I only put an effort in constructive (hopfully!) critique to companies that I believe in! You guys rock!

[disturb]

Just wanna chime in and say that i'd love an env follower added to whatever is planned for tbk's update ...

[bmanic]

Here is the "How do you judge an EQ?" thread. There is a lot to read trough and a lot of debate but there are several audio examples too and awesome explanations from Christian Budde (developer of Electri-Q).

The issue with transients is discussed from page 6, ->

Cheers!
bManic

[DaveSonalksis]

No, fidelity, as in how the transients survive.

Ah, this will be a matter of phase response!
This is REALLY subtle stuff! You have cool hearing!

And good monitoring! What do you use?
I'm currently running Absolute2s, which are ok for most things...
I'm getting some Genelec 1030s, which I've been lusting after for years - back at Focusrite we always used Genelecs for testing - although that was mostly because they never exploded

Can some of you guys link me some tunes? I KNOW the mixes will be sick, and I'd LOVE to listen!
Email address is dave at sonalksis dot com, unsurprisingly

There was a rather good thread about EQ fidelity here a while ago. I'll dig it up for you.

That would be great! I'd like that

Also, like people have already said, Voxengo, Massenburg, Algorithmix, PSP etc. all use oversampling and it DOES sound better. Sure, I'm all to aware of the problems with upsampling/downsampling (That's why I bought Voxengo R8brain Pro) and it is a compromise, but when done right, it really does improve the sound a lot. Besides, you can always make it a toggle option as a button on the GUI and leave the default to off (like voxengo, which even has an "auto" option where oversampling is turned on only during rendering).

Yeah. I think that oversampling is appropriate for things like mastering eqs, and 'special' eqs, but your general mix eq probably doesnt want to oversample [or you'd say goodbye to all your cpu pretty quick].

It's totally feasible we could do some oversampling EQ at some point- I did a little bit of research into minimum phase antialias filters a while ago. Even longer ago, I'd finished an algorithm to generate arbitrarily perfect minimum phase antialias filters... and then the laptop got nicked :'(

I wonder if, perhaps, the times when you want that oversampling EQ is the same time that you want other particular properties to your EQ (curve shapes and such)...?
Or perhaps it would just be generally useful to have on any EQ...?

Nothing stops me adding it as a feature... it'll take a fair while to do properly though!

Perhaps you'd rather I did other things... like adding the SV315 sidechain?
Or perhaps you'd prefer no oversampling and an 'AirEQ' style mode for SV315?

I'll try to post some specific Sonalksis EQ related problems and compare to some other free and prototype plugins that I have.

That would be cool.

This might sound a little odd (but i'll explain it), and it's just a request, which you can totally ignore, but, if it's not any trouble, could you either post to, or copy to the Sonalksis forums?
The reason is that some of the other guys aren't KvR registered (and won't be swayed), so they can't read the forum here! :-/

Last but not least: Sometimes it's good to leave the "programmer" in you at home and use the "enginer" aspect a bit more. At least while listening.

I'll prolly be shot for saying that..

PS. don't take the critique the wrong way. I only put an effort in constructive (hopfully!) critique to companies that I believe in! You guys rock!

Hehehehehehehe. Thanks!
I appreciate it. I'm glad of the support! A lot of companies are actively scared of KvR, so it's nice to find friends here!

And I +DO+ use the engineer in me while listening... and while trying to come up with new product ideas!
Top-end of EQ is a really fiddly thing though- without the right monitors, you'll struggle to hear it.

The difference between AirEQ and the Sonalksis top-end is really tiny, and one is not necessarily 'better' than the other... [although I will say here and now that I think that the Sonalksis top-end is nicer than the textbook bilinear transform or Orfanidis stuff].
AirEQ is blatantly competition for us... but it doesn't have a graph, and it has only one set of EQ shapes... and as I've said, the sonic difference is a small small thing.
So, I'm not advocating AirEQ over 517, except in the specific cases where fidelity is crucial and you're happy with the AirEQ shapes. And I don't think anyone will disagree that both products are world-class?

I intend to answer AirEQ with my own EQ based on the same technique [since I was doing it years and years ago, it's not exactly cheating, is it? I just got reminded of it]... but with loads more features, and a graph... which can be switched off... and the 517 curve shapes... and perhaps a 'vintage' set..

[Len]

Dave

Just to say welcome here, and as with all forums there are some real dickheads here that I am sure you will ignore. Bmanic is a great help and I'm glad we have him around.

But please - don't spend so much time here thinking you have to respond to all posts - go get coding instead!!!

One small comment on the Sonalksis plugs (I bought the 315 and 517) - they are still my go-to plugs for comp and EQ but it has to be said that since they came out there has been a lot of good competition, and the prices should by now come down as they are no longer worth that much more than other plugs. IMHO anyway.

Welcome again!!!

Len
p.s. - all songs on my myspace site were mixed and mastered purely with Sonalksis 315 and 517. Voxengo Elephant 2 for limiting during mastering.

[Eiosis]

Yes, ok. I've figured out how it works, and I can confirm that it is a BRILLIANT technique... and it WILL sound as good as the Analogue 'original'. In fact, it's as close to perfect as a technique can get, and as close as the developer wants it to be. AirEQ is the real thing people
I'm curious as to whether the guy from eliosound ever worked at Digidesign...

lol Thanks Dave, you are really a nice competitor.. Welcome here, it's nice to see some plugins developers discussing about sound, technology and to read users' whishes..
And no, none of us at Eliosound ever worked at or for Digidesign !

Btw, his write-up for d2 is wrong. If he inspects a little more closely, he'll find that d2 is Analogue like curves... Well, so the guy that made it showed me
But that was many many years ago, and there's more cpu around now, and Forte and AirEQ are closer to Analogue than d2.
This technique appears to be almost identical to that which I used on the Forte Suite plugins.
No. I'm not going to reveal details.

I don't think that I am wrong for the D2 measurement, anyway I could be wrong, so I will measure it again. I remenber that D2 is based on standard Orfanidis filters, and I beleive that Forte is based on what I call "advanced" Orfanidis filters, as later EQ algorithms by Focusrite.
I'm also afraid that we do not use the same technique that Focusrite uses, I think that I see what is the technique you are talking about, and I can say that it produces closer bells shapes than standard Orfanidis ones, but not as symetrical than AirEQ ones.

Wow. Fabrice Eliosound, you are a legend. I agree, this technique is a challenge. It took me three months with a very fast computer for Forte!
The harsh limitation of this technique is that it's an ENORMOUS pain in the ass to do... and there's a process in the step which is incredibly complicated, and very very weird... that took months to solve... but then, I never had Mathematica back then...
Alas, it's a real pain to add new filter types... ENORMOUSLY painful, (but certainly not impossible).

LOL. Thanks for the "legend" !!
In fact, beleive me, our technique is far more painfull that the one I think you talk about. But it is also much more powerfull, and we could virtually make any type of filter with it. But I must agree, the technique used by Focusrite is very painful, and AMLT also took us a huge amount of time to achieve.

The difference between AirEQ and the Sonalksis top-end is really tiny, and one is not necessarily 'better' than the other... [...].
AirEQ is blatantly competition for us... but it doesn't have a graph, and it has only one set of EQ shapes... and as I've said, the sonic difference is a small small thing.

The difference is rather tiny for some bell settings, more important for some critical and extreme bells settings. But the difference is much more important for the Low Pass filter, where I found only analog, oversampling and AirEQ to be suited to what I wait from a Low Pass, i.e. a very gentle cut with a regular and smooth shape.

I intend to answer AirEQ with my own EQ based on the same technique [since I was doing it years and years ago, it's not exactly cheating, is it? I just got reminded of it]... but with loads more features, and a graph... which can be switched off... and the 517 curve shapes... and perhaps a 'vintage' set..

Again, I think this is not the same technique, but if it's the same that the one used in Focusrite plugins, I must admit that it's very good, even if the shapes are not always perfectly matched with analog shapes.

Competition : at the end, it is the customer who is the winner, and that are very good news !


Fabrice,
Eliosound

[diverdee]

Well this has turned into a realy interesting & very civilised discussion.
I'm not really clued up on the technical aspects of all this, but it is always interesting to read posts by you guys who are, especially when there is so much obvious admiration & respect in the discussion.

edit: by the way I demo'd the sonalksis compressor a long time ago, i'm nowhere near an expert but it sounded good to my ears - an update with sidechain functionality would indeed be awesome & (or so I believe) a wise move on your part.

[DaveSonalksis]

Yes, ok. I've figured out how it works, and I can confirm that it is a BRILLIANT technique... and it WILL sound as good as the Analogue 'original'. In fact, it's as close to perfect as a technique can get, and as close as the developer wants it to be. AirEQ is the real thing people
I'm curious as to whether the guy from eliosound ever worked at Digidesign...

lol Thanks Dave, you are really a nice competitor.. Welcome here, it's nice to see some plugins developers discussing about sound, technology and to read users' whishes..
And no, none of us at Eliosound ever worked at or for Digidesign !

YAY! I didn't think you were here on KvR!! Hello!

Btw, his write-up for d2 is wrong. If he inspects a little more closely, he'll find that d2 is Analogue like curves... Well, so the guy that made it showed me
But that was many many years ago, and there's more cpu around now, and Forte and AirEQ are closer to Analogue than d2.
This technique appears to be almost identical to that which I used on the Forte Suite plugins.
No. I'm not going to reveal details.

I don't think that I am wrong for the D2 measurement, anyway I could be wrong, so I will measure it again. I remenber that D2 is based on standard Orfanidis filters, and I beleive that Forte is based on what I call "advanced" Orfanidis filters, as later EQ algorithms by Focusrite.

Ah well.. I probably shouldn't give too much away, but you might as well know
d2 isn't Orfanidis, or even Orfanidis related - it IS the same as you - I'm almost 100% sure! (I might have used a hexeditor to confirm my suspicions too... SORRY!!! I +HAD+ to know!! )
Forte is the based on the same idea as d2, but I refined it, and made it more accurate.

At the risk of being -too- cryptic, if I were to say that I refined the metric... you might know that we're eye to eye...

However, for the Saffire EQs, I use some pseudo advanced-orfanidis stuff.

I'm also afraid that we do not use the same technique that Focusrite uses, I think that I see what is the technique you are talking about, and I can say that it produces closer bells shapes than standard Orfanidis ones, but not as symetrical than AirEQ ones.

Nah, the d2 is just a very very very early implementation of what we do now...
Groundbreaking, I feel. Not enough people know about Paul Neyrinck. He's another legend.

Wow. Fabrice Eliosound, you are a legend. I agree, this technique is a challenge. It took me three months with a very fast computer for Forte!
The harsh limitation of this technique is that it's an ENORMOUS pain in the ass to do... and there's a process in the step which is incredibly complicated, and very very weird... that took months to solve... but then, I never had Mathematica back then...
Alas, it's a real pain to add new filter types... ENORMOUSLY painful, (but certainly not impossible).

LOL. Thanks for the "legend" !!
In fact, beleive me, our technique is far more painfull that the one I think you talk about. But it is also much more powerfull, and we could virtually make any type of filter with it. But I must agree, the technique used by Focusrite is very painful, and AMLT also took us a huge amount of time to achieve.

Then it's definitely the same one

The difference between AirEQ and the Sonalksis top-end is really tiny, and one is not necessarily 'better' than the other... [...].
AirEQ is blatantly competition for us... but it doesn't have a graph, and it has only one set of EQ shapes... and as I've said, the sonic difference is a small small thing.

The difference is rather tiny for some bell settings, more important for some critical and extreme bells settings. But the difference is much more important for the Low Pass filter, where I found only analog, oversampling and AirEQ to be suited to what I wait from a Low Pass, i.e. a very gentle cut with a regular and smooth shape.

Yes, this is true. I never bothered with analogue style LPFs... at the time it didn't seem like anyone wanted them... I had a look at them, and the technique REALLY struggles with them... I'm impressed that you got it to work... although I do have a sneaky suspicion as to how you did it...

Anyway, I can have no criticism of AirEQ whatsoever... but there's a limit to how much I can endorse it, obviously!!

I intend to answer AirEQ with my own EQ based on the same technique [since I was doing it years and years ago, it's not exactly cheating, is it? I just got reminded of it]... but with loads more features, and a graph... which can be switched off... and the 517 curve shapes... and perhaps a 'vintage' set..

Again, I think this is not the same technique, but if it's the same that the one used in Focusrite plugins, I must admit that it's very good, even if the shapes are not always perfetcly matched with analog shapes.

Competition : at the end, it is the customer which is the winner, and that are very good news !


Fabrice,
Eliosound

Let's keep making sick plugins

I mean... EQs are really good now (well, ours are!)... but I reckon we can take this to a whole new level... mmmm.....

I found a new paper on filter topologies this morning which might be veeeery interesting... it's 10days old...

Email me, and I'll send it over

Dave.

[DaveSonalksis]

Wow. Fabrice Eliosound, you are a legend. I agree, this technique is a challenge. It took me three months with a very fast computer for Forte!
The harsh limitation of this technique is that it's an ENORMOUS pain in the ass to do... and there's a process in the step which is incredibly complicated, and very very weird... that took months to solve... but then, I never had Mathematica back then...
Alas, it's a real pain to add new filter types... ENORMOUSLY painful, (but certainly not impossible).

LOL. Thanks for the "legend" !!
In fact, beleive me, our technique is far more painfull that the one I think you talk about. But it is also much more powerfull, and we could virtually make any type of filter with it. But I must agree, the technique used by Focusrite is very painful, and AMLT also took us a huge amount of time to achieve.

Then it's definitely the same one

I should just point out, btw, that Forte DOES NOT use the technique for it's HPF and LPF - I don't want to make false claims!
Also, it's ProTools ONLY, and will stay that way.
And it's VERY VERY expensive
And no-one has ever heard of it :'(