Your clean sine is already not clean as you see in 01 pic. Try another sine made with cleaner generator. Huhuh)

 

I made it with WaveLab (should be clean). The slope is because of the way SPAN works.

 

Try reaper JS tone generator (sine)
SPAN setting:
Block size 65536
Overlap 93.8 (max right)
Avg Time 50 (lowest)
Freq Lo 1.0
Freq Hi (22k or higher)
Range lo -180dB
Range Hi - any
Slope 9 (max)

Now put dmg equilibrium after sine generator and before span. Then same with proq2.
Then work with output gain or input gain. With proq2 distortion grows. With equilibirium I can't see any distortion. Even with +72 dB. But if proq2 inserted before equilibrium with its gain. Helluva distortion!

 

If you're caring about distortion about -170dB , you're doing gain staging, or something else very wrong.

It's a marketing/showoff thing that is, in practice, useless, unless you're producing music at around -300 - -500dB.

 

A post from FabienTDR

 

@mild pump milk
I did your test on some synths. Sine wave from serum, dune 2 and operator shows distortion at around -160db. I think according to today's standards fabfilter is doing great. It'll be good to see every dev going 64bit. But only if it's worth it.

Edit:
From above post - "you'd have to amplify the signal 31622 times to make it perceivable to healthy individuals."

Fabfilter knows what they are doing.

 

Last edited by a moderator: Nov 28, 2018

Has a bit less aliasing.

These settings don't help here, it looks even more weird.

But nonetheless, all I wanted to show is that with analogue emu pluggies the occurring noise-floor level is way above any artifact and that the outdated Oxford is as good as EQilibrium in this measurement. And I doubt, but don't know, that the Oxford is working with 64bit.

 

Didn't think I needed it ... then I seen the feature to see other EQ in project.. hmmmmm interesting

 

"When i insert proQ2 there is distortion about -170dB" Omg, that's sooooo bad. I wonder if fruit flies can hear that?

 

These span settings show fast (possible fast, but without screen fps latency, so smoother fft, but yeah timing is weird), precise analysis (high block size precise fft, mostly useful for lower freqs), increased slope of low level (I wouldnt used that if there was range -200dB or so). There is no faster,precise setting here. Only slope increased for to see distortions increased from 180dB limited range. In usual cases I use another slope settings. Or try to reduce block size, 64k is useful for lower frequency precision, not timing, so try 8 or 16k.
My settings are just maximum for all, but lost in timing precision and if we used more block size, we would gain up slope to see distortions, because they would dissapeared below 180dB threshold. So yeah, less slope for less block size.

And why 990 Hz at 44100? Why not 980 (better) or 1002 (still precise tgan 1k, but bit less than 980)? 1k is for 32k, 96k, 192k.
1260Hz, 1050Hz, 1225Hz sine for 44100Hz as well. But it is all about side distortion around signal, useful for saturation/clipping/etc tests

And i would not say it is aliasing. It is quantization errors limited by bit depth. Aliasing is another, but is also digital enemy

 

Last edited: Nov 29, 2018

Start studying different eq curves. Those plugins aren't all the same because, for one, some of them offer different eq curves, as well as other features. I like ReaEQ but the high shelf is atrocious.

 

Dan Worall is Back.

 

Applause!

 

Last edited: Nov 29, 2018

How do you know their plugins are still 32bit float internal processing?
Ex:
https://www.fabfilter.com/forum/1566/64-bit-internal-processing#comment0
https://www.fabfilter.com/forum/1655/just-released-fabfilter-pro-ds#comment0

 

1. It is written in your first link, - either that or another, if needed we will do that. Read further.
And internal processing is weird as well. So, it might be this: input 64, internal processing 64, output 32. Output maybe different.
2. Measure and you will see. -170dB distortions from clean digital eq is 32 bit float. Not 64 bit.
Same 32 bit float distortions you may see on Kotelnikov 1.5.0 and below. And you will not see them on 1.5.1 with 64 bit input/output. Same with TDR Nova 1 vs Nova 2. They not mentioned that in changelog, because FabienTDR forgot, due to a lot of changes and new Features. And it is true.

 

Last edited: Nov 29, 2018

Production Expert

Exclusive Video Review & Test Of FabFilter Pro-Q 3 EQ Plug-in - The Best EQ Just A Whole Lot Better

​

 

As said on gearslutz there are NO attack/release settings on dynamic eq (automatic) , and NO triggering dynamics by using external sidechaining (only for external multitracking analysis) . Hmm.. I would call this update Pro-Q 2.5, so they just introduced simple automatic dynamic eq and improved other features. Not so revolutional upgrade this time, but just a very good update.
Nova made 2.0 way better. A major update for both free and GE versions. Pro-Q3 seems to be weaker

 

Last edited: Nov 29, 2018

Agree. It's a nice update only and PRO-MB is still a valuable tool.
TDR Nova GE dynamic capabilities are better because users can handle a fine tuning. LIke ToneBoosters Equalizer 4.

 

For most people all they wanna know is: Fabfilter Q at last with DynEQ don't give a s*t about nothing else lol