Smart EQ 3 is out !

[fairlyclose]

Nah, that science quip doesnt hold up, and anyway I have not seen that variatiin in the few tests Ive done, so a non issue for me.

[Ploki]

Yeah lol, who the f needs a smart time saver tool that’s indecisive. I have my brains to slow me down.
I’m maintaining that a 3-sec snapshot is a poor implementation of “smart eqing”.
I don’t EQ a track manually if i listen 3 seconds of it.

Even when i do TEOTE or Gullfoss or Soothe i listen to the whole section to tweak it.

[Ploki]

Besides, if you have a half-decent source the amount of EQing it does is way too invasive. Way too many notch boosts over 1k. Not one source i tried smart EQing made it sound better.
When I EQ manually its much more sparsely and broad 90% of the time

[fairlyclose]

Besides, if you have a half-decent source the amount of EQing it does is way too invasive. Way too many notch boosts over 1k. Not one source i tried smart EQing made it sound better.
When I EQ manually its much more sparsely and broad 90% of the time

You can have 2 smart bands, adjust their bandwidth, and set their influence to as small as you like. You could use it as a smart eq below 1khz, one band dynamic to some extent, another not dynamic, and have normal eq bands above 1khz. Its pretty flexible.

[Ploki]

You cant decrease what Soothe calls “sharpness”
(What i assume is FFT resolution) to get broader strokes

[captainfuture79]

I had Smart EQ 2 already, but find Smart EQ 3 sounds better. I also had trackspacer, and the group feature is likely automatically getting trackspacer working on 6 tracks with a few clicks, instead of having to setup the sidechain routing. A huge timesaver if nothing else.

Hello, does smart eq3make trackspacer superfluous?

[Ploki]

No

[Rickskii]

I had Smart EQ 2 already, but find Smart EQ 3 sounds better. I also had trackspacer, and the group feature is likely automatically getting trackspacer working on 6 tracks with a few clicks, instead of having to setup the sidechain routing. A huge timesaver if nothing else.

Hello, does smart eq3make trackspacer superfluous?

Smart Comp has similar ability.

[sambaji]

In acoustic music, you can hear the artifacts which some of these dynamic real-time EQ adjusters add (with their fast attack and releases). Overall, I am a bigger fan of static EQ, with perhaps a little targeted subtle dynamic adjustments of resonances.

[chagzuki]

I made the perhaps bad decision to buy this just recently. I'm undecided on its usefulness, since the spectral mixing via groups is interesting and seems to cut quite precisely in ways that traditional mixing won't, but as others have said in this thread the method by which it arrives as a corrective EQ curve doesn't make any sense. I'm baffled as to why the developers didn't allow for an unlimited sampling time to establish an EQ curve. Very obviously, the small section it derives the curve from isn't representative of the track. Ridiculous. And the detailed curves that it produces are a curiosity; what are they compensating for? Resonances? Or is it actually suppressing the naturally dominant harmonic content of the audio it's analysing, and a result making things sound worse? Hard to tell. I was just playing around with it and put two instances of smart eq serially. The second one produced a curve that actually compensated for the adjustments the first had made(!). Wtf? So the second one decided the decisions of the first needed rebalancing.

What would actually be useful would be if this could operate without creating a compensatory eq curve and still perform in the group mode like a multi-headed version of trackspacer.

[DCrown]

How could this eq be so smart to know what I want to boost or cut? Does it know my taste or goal, does it know the whole song when I just use it on one track, what is the reference for smart eqs decisions or isn't it rather settled decisions of the dev, who delveloped and programmed the plugin, is the dev an experienced mix or master engineer,
the last question because some of smart eqs decisions are not suitable and satisfactory at all?
Imo AI is more or less like presets,
presets of an eq are rarely useful,
you can cut a snare at 400 Hz, but also
boost depending on the snare and mix, so...?!

[fairlyclose]

Smart eq 3 is a fine eq, dynamic settings work well. Man there are some f**king whingers here

[chagzuki]

Smart eq 3 is a fine eq, dynamic settings work well. Man there are some f**king whingers here

I was looking at the dynamic setting/behaviour and it seems to me that it produces deviations from the static curve, so it would still make way more sense to allow the user to sample a much larger section of a track. Or possibly blend curves together so that one could sample a few points and merge them.

[fairlyclose]

Smart eq 3 is a fine eq, dynamic settings work well. Man there are some f**king whingers here

I was looking at the dynamic setting/behaviour and it seems to me that it produces deviations from the static curve, so it would still make way more sense to allow the user to sample a much larger section of a track. Or possibly blend curves together so that one could sample a few points and merge them.

Nothing to stop you using multiple instances but sampling longer intervals will not do the magic you think as the learning will average out short time effects. Better to use multiple instances working on different sections.
But - in practice I never have to do that but of course that will depend on the material you work with

[chagzuki]

Nothing to stop you using multiple instances but sampling longer intervals will not do the magic you think as the learning will average out short time effects. Better to use multiple instances working on different sections.

It's not that it produces a different curve from one section, e.g. verse, chorus, to another, rather that within the same small section of a track it will produce different results based around the specific 6 or so seconds it samples, so according to that logic one would have to break an instrument track down into multiples of 6 second sections. So I would argue there really should be a way to sample larger sections or whole tracks. If it's correcting for unpleasant resonances or suchlike, then these ought to remain consistent over entire tracks.