Smart EQ 3 is out !

VST, AU, etc. plug-in Virtual Effects discussion
KVRist
209 posts since 30 Jan, 2015

Post Tue May 18, 2021 6:33 am

Ploki wrote:
Tue May 18, 2021 4:26 am
martinjuenke wrote:
Tue May 18, 2021 4:15 am

I already have Gullfoss as well and see both compliment each other nicely (Sonible on tracks and Gullfoss on Mixbuss). To overdo something is at the end the responsibility of the user, not a flaw of the tool (at least in my opinion). However, tastes differ obviously…
It is, but smartly tuned parameters make better specific tools.
You can do basically ANYTHING with melda. But anything also means a lot of garbage, because they’re so open (still love ‘em tho)

I had the same quip with gullfoss. Stop fckin boosting, cut and clean and let me do the boosting. (Thats why i like soothe2)
It would have to be renamed "smart cut and clean" :D

KVRAF
1840 posts since 11 Aug, 2012 from omfr morf form romf frmo

Post Tue May 18, 2021 11:36 pm

Ploki wrote:
Mon May 17, 2021 7:35 am
Question is how is a static curve derived from a small portion of an audio better than i.e gullfoss or teote or similar plugs.
And is boosting really a part of mundane mixing? I dont think i ever boost to “clean and tidy” something.
The applications are somewhat different. Gulfoss is something I'd use at the mastering stage where it needs to be very dynamic as there is a lot going on. TEOTE "balances" the spectrum and cleans up resonances.

Smart EQ 3 like I described is a mixing assistant, as in individual tracks, where it has a reference for several types of instruments (the ones with presets) or from a "small portion" can tell plenty what it sounds like. For example, an acoustic bass drum's or trumpet's timbre is going to vary only so much. The tracks interact with each other and it'll make room for some and prioritize others, like Wavesfactory Trackspacer.

As for how it EQs, cutting only is perfectly valid. However, cut and boost is really no different—with the exact same curve shape, the overall amplitude is simply higher. If I must hazard a guess, Smart EQ 3 is trying to maintain a certain gain level so the instances on other tracks can play nice with each other.

If it really bugs you, add a node with a Q of zero. Now drag it down until the entire curve is under 0 db, noting the node's dB. Then increase the instance's gain by that amount.

KVRAF
4123 posts since 17 Dec, 2009

Post Wed May 19, 2021 2:28 am

I tried demoing it again - jesus load time is atrocious.

I don't really use Teote for resonance notching - i don't think it does as good of a job at it as Soothe2, and it's CPU intensive.
And i wouldn't use gullfoss at mastering, it screws up the phase so much it annoys me, shit starts to sound smeared. Teote however, yes.

screenshot:
same violin mic at 3 different places in the same song
even if I apply "dynamic" it still retains the general curve. (The crazy 10dB boost from the last screenshot remains also on the low notes even if i set dynamic to 100).
So which of the 3 curves is best for this violin in particular?
Should I capture every part of the song separately?
(also boosting that 2Khz makes the violin nasal)

If the curve is based only on a 3-second clip it should be MUCH more broad and not as surgical, and i don't see any control to adjust resolution.

even if i lower the gain so it "doesn't" boost, the way overly aggressive curve is still there.

So i guess i shouldn't say "boosting" i should say not making so many aggressive adjacent interventions.

You'd have to have REALLY bad source for this to make things better or faster without intervening too much into the source sound in my opinion.
You do not have the required permissions to view the files attached to this post.
Image

KVRist
101 posts since 31 Jan, 2021

Post Fri May 21, 2021 9:11 pm

From my testing, it appears the universal preset profile in Smart EQ3 at 100% is close to -4.5db/oct noise, with some slight differences. Using a spectrum analyzer, I see mainly a slight 2db bump from 500hz to 3khz and a slight 2db dip from 3khz to 15khz. (While I generally like the profile on the master track, it would be great if it had other references to choose from.) In contrast, Ozone is pretty much flat with -4.5/oct noise, with a little Hp and Lp on the ends. Teote is also fairly flat depending upon the slop selected.

Unfortunately, as of April, Sonible is not no longer offering their simplified auto EQ plugin Balancer as a free download. Although lacking almost all the customization features, besides the level of mix, there were a couple of things I like more than the full version--less Cpu intensive and the ability to easily choose Warm, Neutral, and Bright spectrums. It appears that Sonible has updated their profiles in Smart EQ 3 as the profiles differ than Balancer.

I like using Smart EQ for generating EQ ideas (I only have the demo). However, I tend to prefer TDR Nova GE since I can use my own reference profiles, plus I already own it. In Ableton, I keep a few modified noise clips of various "colors" on a track to select from as references to feed into the TDR. I create these noise reference tracks using the free plugins Pink, to generate pink noise, the free sTilTv2 EQ to change the slope (e.g., to -4.5db/oct), and a spectrogram. Sometimes I'll use another EQ for further alterations (e.g., attenuated upper frequencies).

Back to Smart EQ. I definitely suggest demoing it. Like I said the universal profile is quite pleasant (not at 100% though). I would be tempted to get it. However, with how slow I am at producing my own songs, I can't justified the extra expense.

KVRAF
1840 posts since 11 Aug, 2012 from omfr morf form romf frmo

Post Fri May 21, 2021 11:39 pm

Ploki wrote:
Wed May 19, 2021 2:28 am
I tried demoing it again - jesus load time is atrocious.
What host do you use? It loads very quickly for me (REAPER). I run a licensed copy, it may be doing a demo time check.
even if i lower the gain so it "doesn't" boost, the way overly aggressive curve is still there.

So i guess i shouldn't say "boosting" i should say not making so many aggressive adjacent interventions.

You'd have to have REALLY bad source for this to make things better or faster without intervening too much into the source sound in my opinion.
You have Strength cranked to 100% in every screenshot, of course that is going to sound extreme. Use your ears, bring Strength down to where it improves the sound, and use the shaping tools if necessary. Personally I start it at 0 Strength and bring it up.

If there's something else you prefer (mixing is subjective after all), you can add your own nodes to address it. They tell you do this in the manual especially if you want to create your own presets.

As for different curves at different points in the song, it will optimize for whatever you learned it on. No different from a human engineer, but they would generally make choices that work for the rest of the track as well. So again, adjust Strength, add your own nodes, etc.. For something like a violin that could have wildly different timbres, a mixing engineer might split up, say the sustains from a col legno section. Though I wouldn't use this for orchestral music.

Like I said previously, this is a mixing assistant, it does not bring you over the finish line. You do not have to agree with its choices; you would not necessarily agree with a human mixing engineer either. So do not expect it to do completely and exactly what you would.

KVRAF
4123 posts since 17 Dec, 2009

Post Sat May 22, 2021 12:52 am

Logic

That's not the point. two of the screenshots were same articulation, similar dynamics and a third apart.
If i set strength to 50% it will be the same just less intense.
If i need to do so much baby sitting how much did it really do for me? Gullfoss or teote - set up and it works.
this takes so much manual correction the question is how much time does it really save?
I just fail to see how a global track EQ works if you analyse 3 seconds of it.
Image

KVRAF
1840 posts since 11 Aug, 2012 from omfr morf form romf frmo

Post Sat May 22, 2021 5:04 am

Gullfoss and Teote have different purposes, I don't know why apples keep getting compared with oranges. Compare it to Neutron 3's mixing assistant functionality if anything.

The time savings aren't just in the EQing, it's the mixing. So if the automatic EQ isn't to your taste on a track, don't use it and set your own EQ curve. It's just a few broad adjustments, as you said. When it dynamically adjusts your custom curve in relation to the other tracks it's managing, you don't have to make any manual adjustments or ride any automation or set up sidechain ducking.

KVRAF
4123 posts since 17 Dec, 2009

Post Sat May 22, 2021 5:41 am

Its not that different because they make similar curves in similar spots :)
Its just truly dynamic.
Havent tried the group thingies yet tho
Image

KVRist
331 posts since 15 Mar, 2006

Post Sat May 22, 2021 9:25 am

The upgrade is well worth it to me.
The features, sound, and interface just seem to click for me.
I'm liking it!

KVRAF
3892 posts since 26 Nov, 2015 from Way Downunder

Post Sat May 22, 2021 1:43 pm

The thing with a lot of these auto-EQ type plugins everytime you capture or apply a 'profile', it's different, sometimes wildly. It's hard to trust something that tells you to boost 3dB at 1kHz one minute and not boost the same frequency at all the very next minute. And it's all going on "under the hood" so you don't have any real control - once you get into manually correcting the auto-EQ decisions, you are defeating the purpose of using it in the first place. You fall back to using it as a second opinion, but not trusting it to EQ for you.

So I dunno... TEOTE is the only one that's actually wowed the heck out me. Still, I'm planning to demo this...

KVRAF
4123 posts since 17 Dec, 2009

Post Sat May 22, 2021 2:04 pm

Exactly!
Image

KVRAF
1524 posts since 4 Jul, 2019

Post Sat May 22, 2021 3:14 pm

MogwaiBoy wrote:
Sat May 22, 2021 1:43 pm
The thing with a lot of these auto-EQ type plugins everytime you capture or apply a 'profile', it's different, sometimes wildly. It's hard to trust something that tells you to boost 3dB at 1kHz one minute and not boost the same frequency at all the very next minute. And it's all going on "under the hood" so you don't have any real control - once you get into manually correcting the auto-EQ decisions, you are defeating the purpose of using it in the first place. You fall back to using it as a second opinion, but not trusting it to EQ for you.

So I dunno... TEOTE is the only one that's actually wowed the heck out me. Still, I'm planning to demo this...
Do you mean the same input gives different results at different times?

KVRAF
3892 posts since 26 Nov, 2015 from Way Downunder

Post Sun May 23, 2021 4:22 pm

fairlyclose wrote:
Sat May 22, 2021 3:14 pm
MogwaiBoy wrote:
Sat May 22, 2021 1:43 pm
The thing with a lot of these auto-EQ type plugins everytime you capture or apply a 'profile', it's different, sometimes wildly. It's hard to trust something that tells you to boost 3dB at 1kHz one minute and not boost the same frequency at all the very next minute. And it's all going on "under the hood" so you don't have any real control - once you get into manually correcting the auto-EQ decisions, you are defeating the purpose of using it in the first place. You fall back to using it as a second opinion, but not trusting it to EQ for you.

So I dunno... TEOTE is the only one that's actually wowed the heck out me. Still, I'm planning to demo this...
Do you mean the same input gives different results at different times?
Yes - with basically every matching EQ out there, they never give you the same result twice - even with a precisely fed section (same input). And I always find that a bit unnerving to trust. An offline analysis for such tasks might be more trustworthy, but I don't know of any.

KVRAF
1524 posts since 4 Jul, 2019

Post Sun May 23, 2021 4:32 pm

MogwaiBoy wrote:
Sun May 23, 2021 4:22 pm
fairlyclose wrote:
Sat May 22, 2021 3:14 pm
MogwaiBoy wrote:
Sat May 22, 2021 1:43 pm
The thing with a lot of these auto-EQ type plugins everytime you capture or apply a 'profile', it's different, sometimes wildly. It's hard to trust something that tells you to boost 3dB at 1kHz one minute and not boost the same frequency at all the very next minute. And it's all going on "under the hood" so you don't have any real control - once you get into manually correcting the auto-EQ decisions, you are defeating the purpose of using it in the first place. You fall back to using it as a second opinion, but not trusting it to EQ for you.

So I dunno... TEOTE is the only one that's actually wowed the heck out me. Still, I'm planning to demo this...
Do you mean the same input gives different results at different times?
Yes - with basically every matching EQ out there, they never give you the same result twice - even with a precisely fed section (same input). And I always find that a bit unnerving to trust. An offline analysis for such tasks might be more trustworthy, but I don't know of any.
not been my experience with smartEQ but that is just with a couple of tests. In some ways tho it doesn't matter if the result is a good result each time ie challenges the notion of there being one answer to the EQ - which no mixing engineer would believe

KVRAF
3892 posts since 26 Nov, 2015 from Way Downunder

Post Sun May 23, 2021 5:35 pm

fairlyclose wrote:
Sun May 23, 2021 4:32 pm
MogwaiBoy wrote:
Sun May 23, 2021 4:22 pm
fairlyclose wrote:
Sat May 22, 2021 3:14 pm
MogwaiBoy wrote:
Sat May 22, 2021 1:43 pm
The thing with a lot of these auto-EQ type plugins everytime you capture or apply a 'profile', it's different, sometimes wildly. It's hard to trust something that tells you to boost 3dB at 1kHz one minute and not boost the same frequency at all the very next minute. And it's all going on "under the hood" so you don't have any real control - once you get into manually correcting the auto-EQ decisions, you are defeating the purpose of using it in the first place. You fall back to using it as a second opinion, but not trusting it to EQ for you.

So I dunno... TEOTE is the only one that's actually wowed the heck out me. Still, I'm planning to demo this...
Do you mean the same input gives different results at different times?
Yes - with basically every matching EQ out there, they never give you the same result twice - even with a precisely fed section (same input). And I always find that a bit unnerving to trust. An offline analysis for such tasks might be more trustworthy, but I don't know of any.
not been my experience with smartEQ but that is just with a couple of tests. In some ways tho it doesn't matter if the result is a good result each time ie challenges the notion of there being one answer to the EQ - which no mixing engineer would believe
Not trying to be confrontational but I disagree. If it uses a "scientific" method of analysis, then results should be 100% repeatable.

But there are always variables in nature, right? Yes, in nature :)

Return to “Effects”