Dx38 question on PEQ's

[Coytee]

I've been piddling around with my Dx a little bit and realized I've got some PEQ's that aren't used.

Some of them are in the "main input 1" or "main input 2" section and some are in the "output 1,2,3,4" sections...

If I wanted to add a PEQ to try to suck down a room resonance at say, 90 hz and -3 or -6 db...(no clue on Q but it doesn't matter for my question)

and here's the thrust of my question

Can I treat any and all of my PEQ slots the same? (I realize I won't put a 90 hz peq in my tweeter section though)

can I put any peq tweeks in ANY peq slots and it won't matter or do some 'belong' in the main in section and others belong in the output sections??

My suspicion is it won't matter. I want to verify that so that if it DOES matter, I don't create a problem and not realize it.

I'm just a Dx (and peq for that matter) dummy!!

[Edgar]

Which configuration (from chapter 6 of the DX38 manual) are you using?

Probably all of the PEQ filters are the same. (If they're still using the ones that I programmed for EV many, many years ago, then they are definitely all the same.) But whether you can use a particular PEQ filter in a particular place depends upon the signal routing.

Or perhaps I'm misinterpreting your question. Are you asking whether the EQ to tame a room resonance should go before or after the crossover? The answer to that question is that it should go before the crossover in each channel, with identical parameters in both. This is because a PEQ filter also affects phase, and using different settings in each channel, or in just the lowpass or highpass section of each channel, will cause some weird stereo effects due to the different phase shifts in each path.

Also, don't try to notch-out the room resonance exactly by using a very narrow filter. You'll never hit the frequency exactly, and in fact the frequency will change slightly with temperature, humidity, barometric pressure, and other factors. And when a peak and a notch don't align exactly, you'll hear some very unpleasant artifacts.

Greg

[Coytee]

Which configuration (from chapter 6 of the DX38 manual) are you using?

If I interpret you correctly, I'm doing stereo 2-way, 2 in, 4 out. (is that what you were asking?) If memory serves me, on my 4 outputs, channels 1/3 are LF outputs and channels 2/4 are HF.

But whether you can use a particular PEQ filter in a particular place depends upon the signal routing.

In other words, if I'm PEQ'ing at 90 hz (just an example) I'd not put it in the tweeter output line (channels 2/4 that carries 500hz on up). I could put it in the channel carrying my bass signal OR (I presume) I could put it in the master input line. The HF outputs wouldn't be affected by that.

Are you asking whether the EQ to tame a room resonance should go before or after the crossover?

No, I was really just giving that an example in case someone asked "why are you PEQ'ing anything?" (although I'll say that it DID calm the resonating a bit but MAS gave me a lesson why I shouldn't do that in lieu of room treatments for that same problem)

The reality is, I'm kind of retentive on some things. When I enter my PEQ's for example, I put the PEQ using the lowest frequency first followed by the next & so on. Couple years ago when I got the settings from Roy as a spreadsheet (instead of the RACE file), I received them with PEQ's at 5,200K, then 310K, then 6800K and so fourth...(just making that up to draw picture). They were all mixed up in their order.

So, I'm in part, trying to reorganize the PEQ's so they are in this ascending order and I hit my wall of confusion... "can I put this PEQ here or not" (here defined as main in verses putting it in channels one and three or two and four...If I have any left over PEQ's, I'd rather they be open on each channel and not at the main input channel (unless I'm backwards and it makes more sense to do it otherwise)

ok... I've been outed as a retentive....I feel better now. anyone else want to join me...

[:S]

[Edgar]

If I interpret you correctly, I'm doing stereo 2-way, 2 in, 4 out.  (is that what you were asking?)

Yes, that's what I was asking. Thank you.

In other words, if I'm PEQ'ing at 90 hz (just an example) I'd not put it in the tweeter output line (channels 2/4 that carries 500hz on up).  I could put it in the channel carrying my bass signal OR (I presume) I could put it in the master input line.  The HF outputs wouldn't be affected by that.

That's correct. Since you're not sending any significant 90 Hz signal to the HF outputs, adding EQ at 90 Hz only in that channel wouldn't do anything.

The reality is, I'm kind of retentive on some things.  When I enter my PEQ's for example, I put the PEQ using the lowest frequency first followed by the next & so on.

Ah, now I understand. Sometimes the order of the filters matters, but only sometimes. Basically you never want to have a cut at some frequency followed by a boost at that same frequency. The reason is that the noise floor stays essentially constant, so if you cut the signal level you also reduce the dynamic range. If you then boost the signal level you also boost the noise floor.

It's probably not much of a problem with the DX38 filters, though. Again, assuming that they are the filters that I programmed when I worked there, then they are capable of over 138 dB dynamic range. So boost and cut to your heart's content, in any order you want.

Understand, though, that it does matter where you put the filters, depending upon whether you're trying to EQ an individual driver or you're trying to do some room EQ. EQ for an individual driver belongs in the output channel (after the crossover) feeding that driver. Room EQ belongs in the main in of both channels with identical parameters, for the reasons that I outlined in my previous message.

I've been outed as a retentive....I feel better now.  anyone else want to join me...

Hey, I'm an engineer. Look up "engineer" in the dictionary and it says, "see Anal Retentive".

Greg

[Islander]

The reality is, I'm kind of retentive on some things. When I enter my PEQ's for example, I put the PEQ using the lowest frequency first followed by the next & so on. Couple years ago when I got the settings from Roy as a spreadsheet (instead of the RACE file), I received them with PEQ's at 5,200K, then 310K, then 6800K and so fourth...(just making that up to draw picture). They were all mixed up in their order.

Personally, I prefer the term "obsessive", since it sounds less rude and may be more accurate in certain instances. Besides, wouldn't you want the mechanic who services your brakes or any surgeon who works on you to be obsessed with getting it right? I know I certainly would.

When Roy sent me the Dx38 settings, I noticed that they were out of sequence, frequency-wise, and figured there must be a good reason. After some thought, I concluded that it was to prevent any interactions between PEQs that were fairly close together, like the 86 Hz and 90 Hz settings. Accordingly, I did it just the way Roy listed it.

[mikebse2a3]

Room EQ belongs in the main in of both channels with identical parameters, for the reasons that I outlined in my previous message.

Hey Edgar

In Richards example of a 90Hz PEQ for a room mode. Assuming a crossover setup in the 500Hz range between LF and HF sections please explain why this PEQ couldn't/shouldn't be placed in either the Main In or LF OUT channels as long as it's done identical in both channels(ie:Left and Right in a stereo setup) with the same parameters?

mike tn

[Edgar]

In Richards example of a 90Hz PEQ for a room mode. Assuming a crossover setup in the 500Hz range between LF and HF sections please explain why this PEQ couldn't/shouldn't be placed in either the Main In or LF OUT channels as long as it's done identical in both channels(ie:Left and Right in a stereo setup) with the same parameters?

Because I'm being anal retentive. In this case it won't matter much. But if the PEQ center frequency was reasonably close to the crossover frequency, it would. This is because many crossovers (such as Linkwitz-Riley) are designed so that the LF and HF sections operate as complimentary pairs. If you EQ only the LF or only the HF, that relationship is destroyed.

Greg

[mikebse2a3]

Because I'm being anal retentive. In this case it won't matter much. But if the PEQ center frequency was reasonably close to the crossover frequency, it would. This is because many crossovers (such as Linkwitz-Riley) are designed so that the LF and HF sections operate as complimentary pairs. If you EQ only the LF or only the HF, that relationship is destroyed.

Greg

Thanks Edgar

I didn't think it should matter in the above example but just wanted to make sure I wasn't misunderstanding something.

mike tn (part time anal retentive)[]

[mikebse2a3]

When Roy sent me the Dx38 settings, I noticed that they were out of sequence, frequency-wise, and figured there must be a good reason. After some thought, I concluded that it was to prevent any interactions between PEQs that were fairly close together, like the 86 Hz and 90 Hz settings. Accordingly, I did it just the way Roy listed it.

I really don't believe the sequence of the PEQ filters matters but it might or might not matter if they are placed in the main in or channel out section depending on what Roy's purpose was when doing the EQ.

For example something that might confuse someone is if you have a case of the crossover being set at 500Hz for the LF and HF sections and you notice a (- amplitude)PEQ is placed at 700Hz in the LF ouput channel then in that case the purpose of the PEQ could be to create a combined acoustical roll-off much steeper than the intial slope of say a 500Hz 24db linkwitz riley filter would do on its own. You would not want this PEQ placed in the Main Input channels because it has a very specific purpose for the LF section even though it is at a higher frequency than the crossover frequency which might cause some confusion for some but just because you have crossed at 500Hz electrically it's the acoustical response that is the end goal! You must be able to measure and confirm as Roy can that what you want to happen in the (amplitude and phase domains) is what does happen though!

mike tn

[Islander]

I just didn't want to second-guess Roy, who knows a very great deal more than I do about this stuff. I followed his directions and I'm very happy with the sound.

[mikebse2a3]

I just didn't want to second-guess Roy, who knows a very great deal more than I do about this stuff. I followed his directions and I'm very happy with the sound.

If in doubt wise choice Pat![][]

mike tn

[Coytee]

Because I'm being anal retentive.

Tell em bro!!

[djk]

Since the room peak is in the time domain, please explain to me how putting a dip in the amplitude response will make it sound better.

I've tried narrow band EQs for room problems, the sound is not acceptable (ignoring the sound quality of the EQ itself).

[Edgar]

Since the room peak is in the time domain ...

The room reflections take place in the time domain, but the resulting "peak" is in the frequency response.

... please explain to me how putting a dip in the amplitude response will make it sound better.

If the room response peak is minimum phase, then a PEQ of the proper frequency, cut, and width can cancel it exactly. (See this article.) A room response dip generally cannot be corrected by EQ, however.

Reading through some online articles this morning, I find that there is quite a bit of disagreement as to whether room response anomalies are minimum-phase or nonminimum-phase. (See this article, for example.) I expect that there is some of both in most cases.

I've tried narrow band EQs for room problems, the sound is not acceptable (ignoring the sound quality of the EQ itself).

It is far better to correct the room itself than to try to correct its problems with EQ. But used in moderation, PEQ can help.

Greg

On EDIT: this article explains it better;

"Unlike the entire loudspeakerroom response curve, which cannot be inverted accurately to provide a room equalisation curve because it is non-minimum phase, controlling room modes is more practical because they do exhibit minimum phase behaviour.

In addition, while the magnitude esponse of real rooms ie, rooms which do not have perfectly reflecting walls may vary a great deal in different locations in the room, the decay time of a resonance can be measured successfully almost anywhere in the room, and does not vary much with listener location. In addition, the filters you build as a result, which will be designed to control the reverberation time at specific frequencies, will work anywhere in the room too."

[djk]

My point exactly.

The speaker/room response in non-minum phase and cannot be corrected by tinkering with an EQ.

Trying to control RT60 with EQ doesn't work, it just sounds bad.

The best compromise I have found is to EQ it flat, see how much cut you used, the reduce the amount of cut a couple of dB, listen to it, reduce the amount of cut again (repeat as needed).

The system ends up sounding lifeless it the room is lively and you try to fix the whole problem with EQ.

Anything you can do to the room to reduce the amount of EQ needed will sound better overall.

[Don Richard]

That's correct. Since you're not sending any significant 90 Hz signal to the HF outputs, adding EQ at 90 Hz only in that channel wouldn't do anything.

Out of band EQ on xover outputs are used in some pro sound systems to correct in band phase, as you hinted in an earlier reply to the OP. Note that this application of PEQ cannot successfully be implimented by ear or with simple test equipment, something like SMAART or some other dual channel FFT would be required.

Using delays to time align loudspeakers at the xover freq is a well known technique, but in some cases this can shift phase at freqs different from the xover freq in a detrimental way. Out of band PEQ, done properly, can correct or improve overall phase response.

[Edgar]

Out of band EQ on xover outputs are used in some pro sound systems to correct in band phase, as you hinted in an earlier reply to the OP. Note that this application of PEQ cannot successfully be implimented by ear or with simple test equipment, something like SMAART or some other dual channel FFT would be required.

The DX38 also includes allpass filters that can be used for the same thing, though not quite as flexible because they are low-Q -- designed for phase compensation in multiway Linkwitz-Riley crossovers.

Greg

[DrWho]

My point exactly.

The speaker/room response in non-minum phase and cannot be corrected by tinkering with an EQ.

Trying to control RT60 with EQ doesn't work, it just sounds bad.

The best compromise I have found is to EQ it flat, see how much cut you used, the reduce the amount of cut a couple of dB, listen to it, reduce the amount of cut again (repeat as needed).

The system ends up sounding lifeless it the room is lively and you try to fix the whole problem with EQ.

Anything you can do to the room to reduce the amount of EQ needed will sound better overall.

I've had the same experiences you describe, but what do you say to those that claim that a modal peak at low enough frequency is effectively minimum phase? In other words, they're claiming that because the reflections are arriving within a single cycle, that the delay isn't of concern anymore...

It seems to me like this approach is effectively increasing the group delay at the problem frequency while keeping the amplitude the same - which I think might describe some of the incoherancy I've experienced from such an approach.

If that's the case, then it might be possible with some digital filters to push everything else back a bit to result in no group delay at the listening position....but of course all that goes to crap once you move your head (not to mention the likely distortions that might impart).

I've never heard EQ come close to achieving the same results achieved by treating the room. In fact, I make a career out of pulling EQ's out of PA systems (especially 1/3 octave crap tuned with an RTA).

[Coytee]

I make a career out of pulling EQ's out of PA systems (especially 1/3 octave crap tuned with an RTA).

I 'spose it's time for me to rid myself of my Audio Control C-101 as well as my dbx 14/10..... [:S]