32C Channel Plugin

[madmaxmiller]

What if you drive it a little harder. So put in a slightly stronger signal…

But that's just one component of the original console. I guess no one thinks that it alone can get 32C sound. That's why we have the Mixbus 32C.

Personally, I rely on Harrison's statement. It's just a channel plugin. Other candies can be found in the Mixbus 32C DAW. But definitely quality stuff.

I'll do some more testing for sure. However, driving it harder I expect making it even more unlikely to cancel out between the two. Hotter signals also, but I wanted a "real" world scenario for the first run, hence -6dB. Also, if you want to proof/disproof something you need to limit the scope first. The guy in the video was fiddling with the compressor also and whatnot, while the question was: is the 32C EQ different from any stock EQ? So I put EQ vs EQ as clean as possible.

Next thing I'll test: "null" or cancel out as much as possible one setting, then changing the neighbouring freq/gain knobs on 32C and trying to match this in ReaEQ's respective dials without touching the former one. This is when the push-pull of an analogue circuitry comes into play, changing frequencies, Qs and gains of the whole ever so slightly when touching one of the dials. If the 32C EQ does this (and ReaEQ doesn't) it must be impossible to null (or get it to the same low level as with one dial only) and hard to match even if I use all available dials on ReaEQ.

After that I might use the actual built-in EQ of a Mixbus32C instance and do the same tests.
Finally I could put Mixbus32C versus AVA 32C Channel Strip to check what Ben said - if there's a gain/dithering problem with the plugin. These two should Null perfectly with all settings.

Again: All tools are freely available and everyone is invited to try it out themselves. We all know how easy it is to get the "desired" result in an experiment while subconsciously messing up method and apparatus to comply. That's why peer reviews exist, otherwise scientists could just make a YT video and declare this truth.

Cheers, MMM

P.S.: If someone owns the real hardware strip and checks it vs the 32C software -  that would be interesting, too - it wouldn't null I guess as there are too many other components like ADC/DAC in the path, but how close can you get?

[Sojuzstudio]

MMM I appreciate your effort.

[madmaxmiller]

Ok peeps, day 2 of testing - this time with the goal of nulling by hook or crook


Same arrangement as yesterday but different aim: I do a setting in 32C and then I try to match it as good as possible in ReaEQ. I'm confident that it must be possible, given enough time and as many fully parametric EQ bands as needed, to null out any other EQ

Experiment, driving a bit harder this time:

- stick with around -6db input - we don't want under any circumstances have any "drive" here!
- using only bell characteristics in the most "linear" part of the Fletcher-Munson curve
- set 32C upper mid to 2.44 kHz (neutral position) and +10.7 dB gain
- starting from the same values in ReaEQ tweak for greatest cancellation
- then add bands and see if you can tweak the result further
- then change one dial (frequency or band gain) on the 32C and try to follow with ReaEQ
- I will use 1 band ReaEQ, 4 bands ReaEQ (the default) and 6 bands ReaEQ

Results:

1-band on ReaEQ
- it took me a while to fiddle with freq/gain/Q but at 2484Hz / +10.3dB / 2.36 oct I reached a rest noise in a band of 800Hz-8000Hz with a peak of -63 dB at 2500Hz

- not bad but still well audible
   

4-band on ReaEQ
- I ended up with:
- Band 1: 1994.5 / 0.1 / 1.61
- Band 2: 2226.2 / 0.1 / 0.58
- Band 3: 2484.0 / 10.3 / 2.36
- Band 4: 3407.6 / -0.2 / 0.61

- and got:
- some very low dithering between 250Hz and 2000Hz
- a little "bell" from 2500Hz to 8000 Hz with a peak of around -66dB at 4000 hz
- this was very good however still audible because the human hearing is most sensitive at 4kHz haha
- also, it took four bands of EQ with 3 parameters each to get there which took a considerable amount of time to dial in
   

6-band ReaEQ:
- here we go beyond the number of bands we usually have in a channel EQ, let's see how that works out
-this resulted as follows:
- band 1-4: unchanged
- band 5: 5019.6 / -0.1 / 0.32 this contributed most
- band 6: 7037.7 / -0.3 / 0.13 this took just 1 db at 4000Hz away
- overall: dithering below -68 dB from 400 Hz to 6300 Hz with a peak of -68 dB at 3150 Hz and a tad lower at 4000Hz, hiss still audible at normal monitor settings
   

In all setting I was checking if i can improve the result by varying the 32C input gain in increments/decrements of 0.1 dB - in all cases I ended up having it back at 0.0 as the best result

Final test for today:
- I changed the gain on the Harrison to +4.7 dB (-6dB from the previous)
- I left all 6 bands of the ReaEQ engaged and changed only band 3 which is the "matching" band to the 32C

Result: playing only with band 3 (2469.3 / 5.2 / 2.58) I achieved a reasonable result of 2000Hz - 8000Hz "leftovers" with a peak at -67 dB at 3150Hz. This all while the other 5 bands still did their part.
   

SURPRISE: switching off bands 2,4,5,6 and having only 1 and 3 intact with these settings - it NULLED!
   


Conclusions:

Does this proof the claim "all EQs are the same"?

In my opinion: certainly not!
It took a considerable effort to match both EQs close enough, taking 6 bands of fully parametric EQ. Changing a single setting in the 32C EQ required some meticulous matching again in the other EQ - just look at the settings of band 3 alone at the ReaEQ:
- at +10.7dB: Band 3: 2484.0 / 10.3 / 2.36
- at + 4.7dB: Band 3: 2469.3 / 5.2 / 2.58
This can be seen as an indicator that the Harrison EQ indeed slightly changes center frequencies and width and steepness of EQ bands depending on gain for the respective band.
Again: this is only ONE single settings change in ONE band - it will vary more dramatically if you change more frequencies and gains in other bands.

Is it possible to match the Harrison EQ with stock EQ of other DAWs?

Certainly YES. It even occasionally nulls.
But the effort to do so is enormous as you would have to change multiple bands in all parameters to match a single setting - change one setting at the 32C and you have to change multiple settings at the other EQ to match again - you will need a Harrison EQ to do so (for comparison) as the correct settings can't be worked out in your brain alone.

What does that mean for practical mixing? Is Harrison "snake oil"?

Certainly NOT snake oil. While it is possible to achieve the same overall curve with another EQ it can't be genuinely done. You can only achieve equality for one certain setting at a time and then maybe replace the 32C in that track with your tweaked ReaEQ or such.

The claim of the user "bachstudies" was that Harrison's modelling is a lie and the 32C EQ is no different to any other stock EQ, namely the ReaEQ. My tests have shown that Harrisons 32C-EQ is distinct from the ReaEQ and shows the characteristics Harrison Consoles are advertising.

Again, peer reviews are welcome and encouraged, the tools are freely available and if someone finds a flaw in my method - please prove me wrong. That's how science works.

MMM

[Sojuzstudio]

Good analysis. 

I was thinking of buying it to get 32C Eq for mixbuses (because of limited three band standard eq). But at the same time, I get a little more.

Ben should at least offer you coffee. Great effort.

[madmaxmiller]

Ben should at least offer you coffee. Great effort.

If I ever come to Nashville I will knock on the door there

[Lese]

I’m not sure that it resembles the sound of the MB 32C channel strip 100%, especially the filters. Having said that, I used the demo in Studio One 5.3 with great results, but it crashes hard if I open the plugin window when the music plays. I get the beachball and no graphics, I have to force quit the program. This on 10.13.6 High Sierra, never had any problems with other plugins. Back to MB 32C for mixing I guess..

[madmaxmiller]

I’m not sure that it resembles the sound of the MB 32C channel strip 100%, especially the filters. Having said that, I used the demo in Studio One 5.3 with great results, but it crashes hard if I open the plugin window when the music plays. I get the beachball and no graphics, I have to force quit the program. This on 10.13.6 High Sierra, never had any problems with other plugins. Back to MB 32C for mixing I guess..

That will be my next test setup: MB32C vs AVA32C - maybe the difference you experience lies in using the summing engine in Mixbus32C (vs the engine in S1)?
MMM

[madmaxmiller]

Tonight I tried to compare the Mixbus 32C DAW channelstrip with the AVA channelstrip - and I found something weird. I did some error hunting but couldn't find anything. I stripped everything down to the single utility bus in Mixbus - to no avail.
I'm too tired tonight, will set up from scratch tomorrow and see if the anomaly persists.

MMM

[Lese]

The summing engine in MB32C definitely makes a difference. I did some more testing in S1 and it turns out that another plugin was to blame for the crash , nothing to do with the Harrison 32C channel! I’m really sorry… not to mention I lost the introductory price, my fault!

[madmaxmiller]

The summing engine in MB32C definitely makes a difference. I did some more testing in S1 and it turns out that another plugin was to blame for the crash , nothing to do with the Harrison 32C channel! I’m really sorry… not to mention I lost the introductory price, my fault!

Wait for the next sales wave if you can.
MMM