HoRNet SW34EQ analog modeled equalizer

[Frantz]

Is it important if it doesn't?

Llast time I checked this plugin (granted, it was the first maintenance fix or so), I didn't see drastic aliasing issues. Just did so with the recent demo (which is actually still 1.1.0), and also no issues from 44kHz to 96kHz.

I haven't demoed this yet but became interested based on the positive feedback here. My laymen's understanding is that oversampling would be required when running at 44.1 to ensure aliasing does not occur since this is implementing analog modeled filters. Let's see what the dev has to say about it.

http://en.wikipedia.org/wiki/Oversampling

Oversampling can make it easier to realize analog anti-aliasing filters. Without oversampling, it is very difficult to implement filters with the sharp cutoff necessary to maximize use of the available bandwidth without exceeding the Nyquist limit. By increasing the bandwidth of the sampled signal, design constraints for the anti-aliasing filter may be relaxed.

[Compyfox]

Again, I didn't see any throwbacks on the harmony plot.

But the dev can answer that question better than we users could, indeed.

[EcHo2K]

The quick reply is no, SW34EQ doesn't have any kind of oversampling.

The long one is that I tried oversampling in the early days but the filters don't need it since they don't get distorted noticeably near nyquist (the only one that get distorted is the second peak filter, when set at 8khz).

Actually oversampling would make sense for the saturation part, but the original console is so clean that aliasing appear slightly only when you drastically overdrive the input and this is an extreme unintended use of the plugin (after all we have an EQ here not a saturator)

This is to explain that adding oversampling would just add a nearly noticeable improvement in accuracy ad the cost of double the cpu usage

Saverio
hornetplugins.com

[Compyfox]

Still a bit confused on the "Boost = Saturation" part.

I don't know why I remember this, but I had the impression that the saturation is only applied on boost. Apparently, the saturation is "global". Meaning... the channel strip's preamp is running alongside the EQ. And funny enough, you can turn off the EQ and only use the saturation

That is interesting (yet a bit boring, since a lot of EQs do that), but kind of cool at the same time. Especially if his EQ counts to the more "rare console/channel strip" EQ ports.

[Frantz]

Actually oversampling would make sense for the saturation part, but the original console is so clean that aliasing appear slightly only when you drastically overdrive the input and this is an extreme unintended use of the plugin (after all we have an EQ here not a saturator)

Thanks for the response. Interestingly you linked to a review on your website where the author likes to use the plugin firstly as a saturator. For people who like to use the plugin in this way, it could be helpful to offer oversampling as an option.

http://modernmixing.com/blog/2014/02/24 ... in-review/

The first one is by driving the input section until it lights up red. Yep that’s right this plug-in sounds amazing when you start digging in and and hammering that input section. I find it works best for percussion and acoustic instruments but I am willing to try it on anything because it just sounds so damn good.

[EcHo2K]

Ahaha you know what reading that people like to use that plugin overdriven made me think that probably alias is not that bad

No really, I think it all comes down to the proportion between odd and even harmonics that this eq produces, probably it's a "good sounding" mix despite of the small aliasing you can find on frequencies above 4K.

Anyway the best way to find out is to bass the saturation algorithm through some spectrum analyzer.
Regarding the boost, this EQ being an analog emulation add saturation at each eq filter so if you boost at 2K and then boost with the other peak at 4K you are actually working on a "saturated" signal and this all adds to the final sound. Usually i found that other emulation only adds harmonics ad the beginning or at the end of the processing.

Saverio
hornetplugins.com

[Compyfox]

But wouldn't your EQ do the same in this case?

I see it like this:
Input Gain -> Saturation Module -> EQ module -> Output gain.

Once the SAT module is engaged, of course the EQ responds to that. So SlickEQ is currently the most prominent EQ that adds harmonics on boost, but nothing on cut.

[Frantz]

Ahaha you know what reading that people like to use that plugin overdriven made me think that probably alias is not that bad

You know one of Tone2 Electra2's many filter models is called Alias. It actually produces some fantastic sounds by intentionally aliasing. So perhaps a little aliasing is a feature.

[EcHo2K]

But wouldn't your EQ do the same in this case?

I see it like this:
Input Gain -> Saturation Module -> EQ module -> Output gain.

Once the SAT module is engaged, of course the EQ responds to that. So SlickEQ is currently the most prominent EQ that adds harmonics on boost, but nothing on cut.

The internal structure of SW34EQ is like this:

input -> sat. -> hpf -> sat. -> low shelf -> sat. -> low peak -> sat. -> high peak -> sat. -> hi shelf

It's done like this because the original unit has an operational amplifier for every filter, and an op amp works like an AB amplifier having one transistor for the positive half wave and one for the negative. The difference in tolerance between these two transistor produces even harmonics.
Each of this saturation unit is calibrated so that the final harmonic content resembles the one of the emulated channel.

Having all these saturation stages requires quite a bit of optimization to not make the SW34EQ eat too much CPU, that's why adding oversample would really make it heavy...

This doesn't exclude that in the future i could add it as an optional feature for those who whant a "perfect sound"

[EcHo2K]

You know one of Tone2 Electra2's many filter models is called Alias. It actually produces some fantastic sounds by intentionally aliasing. So perhaps a little aliasing is a feature.

It seems somebody likes it so probably is a feature!

[Compyfox]

The internal structure of SW34EQ is like this:

input -> sat. -> hpf -> sat. -> low shelf -> sat. -> low peak -> sat. -> high peak -> sat. -> hi shelf

But wouldn't that mean, that if you turn off the "global" SAT mode (which is pretty much the preamp), that the SAT modes for the filters should respond and "add" to the global saturation?

At least this is how I'd understand it.

Then again, we're talking about an OP AMP. This can be either a hit-and-miss dirty one, or a fairly clean one (depending on the used chips). So this might not register at all.


But... if I analyze the plugin through VSTPluginAnalyzer, I really don't see the behavior as just mentioned. It's either global saturation on, or off.

Might still misunderstand the concept however.

[EcHo2K]

It's simple, if you turn off the saturation you turn off EVERY saturation module. What's cool about having a saturation stage for each filter is that each filter works on the saturated signal of the previous so gain staging it's actually pretty important here since adding 12/15 dB could saturate internally the EQ!

Saverio

[zeep]

What's cool about having a saturation stage for each filter is that each filter works on the saturated signal of the previous

Interesting. I wonder what other eq's have that "summing" characteristic.

This conversation got me very curious to try out this EQ.

[Compyfox]

It's simple, if you turn off the saturation you turn off EVERY saturation module.

Can I make a FR, where you can turn off the "global saturation" (read, the preamp), but keep the OP amps on the filters? If that would make any sense to you, that is.


Like with Slick EQ, an extra button on the GUI:
"Engage EQ SAT"

Or have three "noise" buttons:
Noise, SAT Pre, SAT EQ

For those purists that want the saturation for the EQ boost, but not the whole module. Or those that want the preamp saturation, but no EQ saturation (read: a clean EQ). Or just turn off both. Again, SlickEQ does something similar.


Granted, this would turn everything a bit over-technical and overrun the non-tech heads. But I'm sure somebody would have made that FR at some point after my comments. So let's get it checked right from the start.

The possibility would be there if we look at the flow chart-

[EcHo2K]

Oh sure i could, but this feature doesn't belong to this EQ, this is an emulation of a specific board so it shouldn't be too far away from the original unit...

Saverio