Eiosis AirEQ v1.3

[defjamm]

this behavior was described on the second page and visualized by christian budde with his graphs.

BUT you can't match the air eq with the waves q eqs the higher it gets in the frequency, because they have asymmetrical filters(44k), air eq has symmetrical filters.

[verstaerker]

doesn't work in FL - the UI works but soundwise nothing happens

[4damind]

There is no problem or bug with AirEQ, the way you tune it is simply different from other EQs

Aah, ok! Is no problem with this behaviour. But its interesting in comparison with other EQs. May be that AirEQ is more analog then most other "analog" EQs.
Nobody will equalize while thinking to mathematical curves, the results are important. And AirEQ gives very nice results!

[bduffy]

Interesting. Again, I guess it all comes down to the idea that AirEQ is supposed to be a different type of EQ! Well, whatever it does, it tends to sound like what I expect to hear from an EQ, and really suits my sound.

[Eiosis]

Aah, ok! Is no problem with this behaviour. But its interesting in comparison with other EQs. May be that AirEQ is more analog then most other "analog" EQs.
Nobodoy will equalize while thinking to mathematical curves, the results are important. And AirEQ gives very nice results!

When I thinked about the AirEQ characteristics, there was many choices to make, and I know that choosing such behaviour for the Q factor would seem exagerated for some people. I must admit that it may be surprising, but my personal experience showed that it is a more musical behaviour.

Interesting. Again, I guess it all comes down to the idea that AirEQ is supposed to be a different type of EQ! Well, whatever it does, it tends to sound like what I expect to hear from an EQ, and really suits my sound.

It seems that some people really dig AirEQ, so it is rather normal that some other do not like it at all.. But it would be a shame not to test AirEQ in a real mixing situation, and not to try to familiarize with AirEQ's workflow, in order to decide if such behaviour is suited to your work or not.

Fabrice,
Eliosound

[kilroy]

Fabrice...thanks for the well expressed post. This sets alot of things straight as far as where you are coming from in terms of design concept and execution in this EQ. Also, I would like to apologise for causing you the unnecessary trouble to repeat yourself in some of these explanations. I should have taken the time to read the full body of this thread, as that alone may have preempted some of my questions. My bad.

So, I concede that this EQ has a familiarity curve which, for some of us, may require adopting a different approach to carving and shaping than what we might be accustomed to...and I don't want to leave the impression that this is a bad thing.

Couple of things, though...

For some work, you cannot get narrow enough Q widths with this EQ for clean, critical shaping. This would definately lack the necessary finesse for some restoration/mastering/re-mastering tasks. I can tighten up the Q by assigning a series of serial filters sharing the same Q and frequency and incrementing their gains, but this is awkward, and not a good work around, especially considering the way this EQ structures gain at narrow Q. I am assuming the minimum limit on the Q value is a reflection of this, or no?

There is a kind of dilemma here : generally speaking, you need more (real absolute gain) precision and lower gains for wide Q factors; less (real absolute) gain precision and higher gains for narrow Q factors. If you need high precision for narrow Q, that can be annoying.

Well, at least annoying for me, apparently.

Perhaps auto scale the gain distribution of the control, then... so that it reflects a nice log distribution no matter what the Q width?

[kilroy]

You do know that analogue EQs behave like this?

Sure do, but...

I'm not sure if it's that radical...

Me either...hence the head scratching.

[Eiosis]

Fabrice...thanks for the well expressed post. This sets alot of things straight as far as where you are coming from in terms of design concept and execution in this EQ. Also, I would like to apologise for causing you the unnecessary trouble to repeat yourself in some of these explanations. I should have taken the time to read the full body of this thread, as that alone may have preempted some of my questions. My bad.

Don't blame yourself and don't bother about that, I love to repeat myself on such subjects !! It is not a problem when you are passionate to discuss the things again and again, like many people do on KVR.
The question may not had been sufficiently explained, and I am sure that some people would find this additional explanation very useful and interesting. You could also read the response to "What is the musical difference between sweeping with a visual interface or a knob ?" in the AirEQ FAQ.

So, I concede that this EQ has a familiarity curve which, for some of us, may require adopting a different approach to carving and shaping than what we might be accustomed to...and I don't want to leave the impression that this is a bad thing.

If you feel comfortable with your actual plugin EQs, and that you reach fastly and every time your goals when equalizing, so you could not need AirEQ. Personally, I was never really as happy with other plugin EQs, that I was with analog EQs. Sometimes I could never reach the sound I wanted, and now with AirEQ, I just have to bypass the one or two EQs I had on a track, insert AirEQ, and I get a greater result very fastly.

For some work, you cannot get narrow enough Q widths with this EQ for clean, critical shaping. This would definately lack the necessary finesse for some restoration/mastering/re-mastering tasks. I can tighten up the Q by assigning a series of serial filters sharing the same Q and frequency and incrementing their gains, but this is awkward, and not a good work around, especially considering the way this EQ structures gain at narrow Q. I am assuming the minimum limit on the Q value is a reflection of this, or no?

The funny thing here is that many people will find that the maximal Q on AirEQ is too sharp, especially for mastering purposes. Generally speaking, mastering jobs require wider Q factors, to get a gentle tone shaping. Your post prooves that, once again, there is no absolute rule and that equalization is often a matter of taste. You could also work on some material which really needs very sharp corrections, which is not the case for most well mixed songs. Remastering and restoring tasks probably requires sharper Q factors, and other filters that you do not need for a recent mix, like dehissing, declicking, etc..

Anyway, in a future version, some options in AirEQ will simplify the work if you want to tune several bands at the same time.

Also you could have a look to the Mastering presets, they could be very useful. The Mastering II preset is more intended to correction cuts, while the two other are more for tone shaping. My favorite is the Mastering I preset, very intuitive !

I'm not sure if it's that radical...

Me either...hence the head scratching.

Maybe not.. So let's say that AirEQ is a "radical" EQ !
And it is also possible that designing an analog electronic circuit with such behaviour is very difficult (or even impossible). To get such variations, you would maybe have to adjust several resistors at the same in the filter circuit, with different values and differents variation curves.. It is very difficult in practice, because you have to deal with all Q and gain values for the component calculation, to get the variable resistors you need, which would need to be custom built, etc..


Fabrice,
Eliosound

[Arksun]

Ok, so I finally downloaded the demo of this eq and tried it out on a full mix, comparing to other eq's.

Don't really give a flying crap about technical specs, whats on paper, exact db or Q figures, I tested using my ears first n foremost.

Gotta say, very impressive Fabrice, there really is some magic in that thing for a relatively low cpu hit. The most fascinating thing about the AirEQ is the Q amount and how it affects the entire range as you change the Q shape. I haven't heard any other software eq with a Q that behaves like this one (without the need to press a button to change its shape, see below).

Is that good thing?, well its certainly different, and it seems to sound great!, so I'd say yes.

The closest rival to this I think is still the Sonalksis, some people say its long in the tooth. I say the Sonalksis SV-517 is still one of the best. In terms of CPU usage the SV and the AIR are pretty close on my machine, both well optimised.

In fact you could say that changing the Q amount on Air is almost like a morphing of the bell shapes on the SV, going from Bell type 2, to 1, and then 3. It seems to follow a similar morphing of curvature, at least to my ears it does.

They both have their own character, different colours, which is nice. AirEQ defo provides a smoother top end boosting if you work at 44K, but I work at higher sample rates so the Sonalksis is closer, but with the Air still having the edge.

HarmoniEQ is still the airy smooth king and beats the AirEQ there, but HarmoniEQ isn't as tweakable and chews up much more cpu.

Overall verdict, a great eq. well done

Now, does anyone want to lend me some cash?, i'm broke. lol

oh!, one final point. on your website you have the option to buy with or without the Dongle.

But they're both the same price, wouldn't it then make sense just to buy it WITH the dongle, even if we do have one already, so that that dongle then acts as a spare.

Surely it makes more sense to have the dongle-less version a cheaper option (he hinted subtley to try and reduce the price)

Arksun

[Eiosis]

Gotta say, very impressive Fabrice, there really is some magic in that thing for a relatively low cpu hit. The most fascinating thing about the AirEQ is the Q amount and how it affects the entire range as you change the Q shape. I haven't heard any other software eq with a Q that behaves like this one.

Wow.. Thanks ! Some magic !
Anyway, thanks for the comments and comparison with other products, it is very interesting and I think it will be useful for some people, it could help to choose the right EQ.

I also love the Q factor behaviour, I must say that this is the only EQ which allows me to easily "tune a resonance". For example, on a snare or a tom drum, you often need to add a resonance on the note of the tom to give "more pitch" or to add impact or body to a snare drum (generally around 200-300 Hz). You need to tune the "resonance amount", i.e. the width of the bell curve, but without changing the overall loudness of the sound, in fact you need to only adjust "how the resonance sounds". I found AirEQ to be perfect for this kind of job, I do not need to constantly adjust the gain knob while tuning a bell resonance.

They both have their own character, different colours, which is nice. AirEQ defo provides a smoother top end boosting if you work at 44K, but I work at higher sample rates so the Sonalksis is closer, but with the Air still having the edge.

Yes, even if you work at 96K, AirEQ is a little closer to analog curves than the Sonalksis, but the difference begins to be very subtle at this sampling frequency. At 44 or 48K, the difference is important, especially for Shelves and Low Pass filters.

Overall verdict, a great eq. well done
Now, does anyone want to lend me some cash?, i'm broke. lol
oh!, one final point. on your website you have the option to buy with or without the Dongle.
But they're both the same price, wouldn't it then make sense just to buy it WITH the dongle, even if we do have one already, so that that dongle then acts as a spare.

We sell the two options at the same price because we don't want to charge the customer with the addtional dongle price.. If you already have a dongle, you can get and use AirEQ immediately after your purchase, but you can't if you have to wait for the shipment, because we send dongles with licenses already loaded. So it makes sense to get AirEQ without dongle in this case, to be able to use it immediately.

Surely it makes more sense to have the dongle-less version a cheaper option (he hinted subtley to try and reduce the price)

Maybe a KVR group buy would be a good solution to get AirEQ with a reduced price.. If some people are interested here, obviously.


Fabrice,
Eliosound

[4damind]

I do not find something special with the Sonalksis SV517. Have tried this on many tracks and verified with other EQs I own. I bought GlissEQ from Voxengo and the Hydratone from Tritone Digital after the demotime with the SV 517
For me, this time after the outstanding good Neve-EQs from the UAD, the only EQ which has potential to be a new star on the sky is AirEQ! Some new fresh ideas, some new algorithms and a workflow which gives nice results in shorter time.

Hope the development does not stop there and some further plugins based on this algorithms will be released in the near future

[aMUSEd]

Hope the development does not stop there and some further plugins based on this algorithms will be released in the near future

I was thinking an Eliosound Ampsim would be good

[bduffy]

Hope the development does not stop there and some further plugins based on this algorithms will be released in the near future

I was thinking an Eliosound Ampsim would be good

If it uses the fraction of CPU like AirEQ...sure! My vote goes for the good ol' compressor. I'd like to see what he would do with that. Or an Eliosound delay...

[Arksun]

Fabrice,

Another question if I may.

On the install of the Demo version, it asked me what instruction set the CPU uses. SSE1, SSE2, or Generic.

I'm on AMD Dual Core and decided to just leave it as Generic. The cpu hits still pretty low.

Thing is, I think I'm right in saying the AMD X2 cpu's do indeed fully support the SSE2 instruction set. So, If I had installed the demo and selected SSE2 instead, should I expect to find an even lower cpu usage?

[Eiosis]

Hope the development does not stop there and some further plugins based on this algorithms will be released in the near future

For sure, the development will not stop here ! We are already working on future aspects of these algorithms, and our goal is always to give some high quality processings, with low cpu and zero delay latency. But developing and tuning such algorithms takes a huge amout of time. There was more than one year between the first ideas behind the AirEQ algorithms and the final and working version of the algorithms..

I was thinking an Eliosound Ampsim would be good

In this case I would have to improve my guitar skills !! Because I beleive that to make a great amp sim, you need to be a great guitarist, and have spent hours to play guitar, tune many real guitar amps, etc..
I think that it is very important to be a professionnal user of the product you conceive, and to have a very accurate point of view about all the aspects of this product : the sound, the way you tune it, the technology that you will use, and about the algorithms that will be the base of the processing and the product, why and how you will use this product in a real and professional situation, etc...

Of course, I have used some amp sims in my songs, I played guitar with it, but I think that I am not sufficiently experienced for the moment to design THE ampsim. And there is already many excellent products out there !

Maybe if I find a great guitarist/sound engineer/scientist/crazy ultra perfectionist..
aMUSEd, what would you miss in the actual existing amp simulators ?

If it uses the fraction of CPU like AirEQ...sure! My vote goes for the good ol' compressor. I'd like to see what he would do with that. Or an Eliosound delay...

Yes.. The problem for us is not only the sound quality, but also the CPU and the processing delay. Same thing for a compressor ! Anyway, there's a lot of work before the release of any of these products !

Fabrice,
Another question if I may.

Yes, of course you may ! And you even must !

On the install of the Demo version, it asked me what instruction set the CPU uses. SSE1, SSE2, or Generic.
I'm on AMD Dual Core and decided to just leave it as Generic. The cpu hits still pretty low.
Thing is, I think I'm right in saying the AMD X2 cpu's do indeed fully support the SSE2 instruction set. So, If I had installed the demo and selected SSE2 instead, should I expect to find an even lower cpu usage?

Yes, the CPU hit is improved if you select the SSE2 version in the installation process. For the moment, we do not have made precise measurements on different processors and for different versions of the plugin, so I can not say more about the CPU gain you will have.

We also have to make AirEQ SSE3 compatible and also, maybe a little later, SSE4 compatible. We must evaluate exactly what is the gain for each SSE version, because this is very dependant on how the compilator handles the code when compiling.


Fabrice,
Eliosound