DEQ's working differently

One of the main reasons for my chairside digital EQ (DEQ) is so I can fine tune the settings used for the Panel and Sub DEQ's from the listening position, and then later load them into the those DEQ's at the front of the room  (which requires kneeling down on a foam pad, to protect my knees which have gotten sensitive).  So it's a "sandbox" or "staging area."

It has been working for awhile, after I got everything (really everything) set identical between the two DEQ's, but now there seems to be a bug.

When I first copied the new August 2024 EQ presets for the panels from the chairside DEQ into the front Panel DEQ, they were way wrong.  But that was because I hadn't yet cancelled out all the GEQ's that the left channel used to need, but which I removed from the sandbox version in mid August.

After removing the GEQ's from the front panel "2024" setting, it was nearly but not entirely identical.  I quadruple checked the PEQ's and double checked everything else (including delay and gain offset) and there were no differences.  And I didn't even have any new or old PEQ's in the area of difference.

There were some GEQ's of course, but I also rebooted both interfaces after completely removing the GEQ's and the spectrum differences persisted.  Because the differences are around the crossover point, a plausible explanation for these differences is that there's a delay time difference between the two DEQ's even though I've set the delay parameter identically.  Perhaps there's different delays in different vintages of the DEQ 2496.

Left, Chairside panel DEQ

Left, front panel DEQ

Possibly a difference in latency with optical vs AES/EBU.  That would not be because of the actual transmission velocity, but the optical receivers and transmitters may add delay.  If it's less than 10ms I may be able to adjust it out.  (Because I set panel DAC at 10ms delay to allow other things to have less delay.)  Or I can change the 10ms baseline if necessary.

Buut, the thing that seems wrongest is the AES/EBU, used for everything else, and what was used in the original delay calibration.  Did I somehow stumble on a different delay that works "better" just because the optical method added delay?

What I want to do is make the front panel DEQ work like the chairside does, because that's clearly better.

I keep thinking about the steep cuts at 128 and 144 Hz.  But both of those are in the sub DEQ, which doesn't change in the comparison above.  Only the panel DEQ changes, and it's lowest filter is a narrow cut at 252 Hz which isn't even very big anymore.

*****

I seem to have figured out the problem, but with my chairside DEQ and microphone stand put away, so I could only do handheld measurements.  While those are somewhat different from the stand ones, they clearly and unambiguously show the necessary difference, if not exactly matching the above "chairside" curve, which I can't do without my microphone stand.

The problem appears to be the selection of the "AUX" delay window.  The suckout at 125 Hz goes away when I change it to the "MAIN" delay window, even though both windows show the same 10 msec delay.

I don't remember the setting in the chairside unit.

That seems to fix it in one fell swoop.  I also tried varying the delays and it made differences but I couldn't quickly find the exactly required delay difference to produce the desired response, though it seemed about 0 ms was best.

Well, perhaps what the Aux Delay does is that it adds to the Main Delay, it doesn't replace it, allowing you to have twice as much total delay if needed.   That makes sense, and would pretty much explain these results.

I possibly switched it to Aux when I was copying the value to the chairside unit in early August along with all the other Panel DEQ settings, and I wanted to be sure I had the settings in both Main and Aux set the same on both units, so I switched it to Aux just to see what the Aux setting was.

***** September 2

After my party on Sunday, I put my testing gear and chairside DEQ back in place on Monday and tried to tune a boost for the depression around 180 Hz.

Strangely, in sweeping the depression seemed lower, so I went for 144 Hz at first, then 133 Hz.  I remembered that I had boosted at 133 Hz in the right channel so it made sense.

But then, when I was testing the result on the Analyzer app, it looked funny.  To be sure what was going on, I turned off the boost I had just added.  Then there was a deep depression at the 125 Hz and just above.

That was the same issue that I had just resolved a few days earlier, but now it was happening with my chairside DEQ rather than the one in front.

I checked the delay in the chairside DEQ, and it was set to Main (which fixed the issue in the front, and then I "updated" the chairside unit to that as wel, but without testing it).

I tried setting the delay page back to Aux on the chairside unit...and the problem went away!

So now it seems that to get the correct delay, I must set the front DEQ delay to Main and the chairside DEQ delay to Aux to get the correct delay that doesn't cause crossover frequency cancellation.

This is despite both delays on both units being set to the exact same 10.0 ms.

Perhaps coincidentally, the chairside unit is running from optical digital and the front from AES.  So it seems if you want just the one delay (and not the two delays combined), use the MAIN when running AES, and AUX when running optical.  Otherwise it seems to add the two delays together.  Weird, but I can cope with it.

It also shows you can't trust anything without measuring it first, there may be some weird gotcha somewhere.

Speaking of which, I remembered that I'd last measured the impressive smooth right channel response before I'd changed the front DEQ to Main.  It might not look so good in Main (which fixed the crossover suckout in the left channel) since it was "tuned" with Aux.

I'd better check it now.

Sure enough, in the right channel, it seems to work the other way, I get the crossover suckout with the delay at Main, and smoothness at the crossover (and a depression slightly higher, at 180 Hz) with the delay at Aux.

So it seems, at least using the front DEQ, the right channel wants Aux and the left channel wants Main.  I can't do that with only one DEQ.

Right Channel, front DEQ on Main Delay

Right Channel, Front DEQ on Aux Delay ("wrong")

But the plot thickened as I discovered that with the chairside DEQ, things were very different, I basically get smoothness either way, Main or Aux delay.  There is only a slight difference between the two, with a tiny depression at the crossover frequency (1-2dB) in the "wrong" condition (which makes the left channel unsmooth, which for chairside DEQ is Main) though it's also smoother higher up.

Right, Chairside DEQ on (correct) Aux (best on Left)

Right, Chairside DEQ on Main (worst on Left)

With the Chairside DEQ, it seems the best result is obtained on Aux Delay for both channels.  The worst about the Main Delay with the Chairside DEQ is that there's a rather large suckout at 180 Hz which simply does not exist with Aux.  Both are sorta down at the crossover frequency, with the Aux showing a bump up at 140 Hz making the down-ness seem greater, but only slightly greater in fact.  The difference at 180 Hz is far larger.

It's clear there is a difference between the two delays that has some vague similarities to what I see with the front DEQ, but it works out very differently.

So, are the PEQ's set differently?  Both EQ's have been rebooted since the last adjustment.

I think the PEQ's are the same but it is not an easy visual inspection because they are ordered differently.

Which leads to an interesting question...is the order of PEQ's important?  I've always assumed not.  I could take a quick look for an obvious difference first, but then copying them over, since after copying it is FAR easier to see if they are exactly the same also.  I'd been tempted to do that already.

In a quick inspection, I see no relevant differences.  The only difference in the vicinity of the crossover frequency is the 3/4 octave boost of +4dB at 133 Hz, and it is identical in both DEQ's.  However in the chairside unit, it is last, and in the front unit, it is first.

But next I tried a reboot of the front unit (since I wasn't going to be doing that copying tonight, it's far too late for that kind of detailed work with my eyes).  And the reboot fixed the front DEQ operating in its correct Main Delay mode:

Right, Front DEQ in correct Main delay, after reboot

Maybe I didn't reboot the front DEQ after the last change.

Oh my, so many things to keep track of.  

Maybe should run a full test at the beginning of every day.  Compare measured pink noise with previous results.

I still pity one time audio dealer Ike Eisenson whose home system was allegedly out-of-phase when my friend George visited, and George would never let it down.  I think I'm going to call this kind of problem where something wrong slipped among many previous changes Eisenson's Revenge.