Frequency Chart - is it ok?

[glennconti]

From this thread:

http://www.diyaudio.com/forums/digital-line-level/140451-behringer-deq2496-firmware-update.html

it looks like V2.5 firmware came out in mid 2009. I have contacted the seller for the version and date code. No response yet. From the photo in the ebay ad it looks like the date code on the mic is 09??. I found the V2.5 firmware online and the procedure to upgrade to it. I don't have a MIDI board on my PC but I'm sure my local music store can do the upgrade for me if I bring the unit in with the .syx file on a CD. I may get lucky and not even have to upgrade though.

You are right too about the CPU in the device. That is the difference between the V1 and V2 firmware. From Behringer's website, V1.4 (with the older CPU) is equivalent in functionality and bug fixes to V2.3 (with the new CPU).

Thanks for the heads up. [Y]

[Quiet_Hollow]

Haven't used the Antimode personally. I do own and use the AS-EQ1. I wouldn't give the manufacturers too much credit. They're apporach is far from cloak and dagger.

1. They are a DSP, nothing more. Just a chipset with some memory and an I/O board.

2. First and foremost function is to flatten frequency response without causing more issue present than at the initial condition.

3. In the case of the AS-EQ1, it also matches up the impulse response of the subwoofer signal (samples both the sub and the mains) across the passband to facilitate good crossover, within certain constraint.

Beyond that..."there is no spoon" so to speak. The bulk of what we're actually paying for is convenience in the form of speed and repeatability.

Their marketing department (if they employ one) has to try and sell that to the layman ad hoc, because the only person that will buy it on technical merit alone are those that have already put in the time doing things manually.

2 years of measuring and adjusting my subwoofer by iterative process, and I was ready for something that could automate the whole ordeal. I certainly don't mind rolling up my sleeves and getting dirty with the numbers and gear (there's good learning in that), but when I found out I could spend more time listening to various states of tweak than actually tweaking...I bit on it.

[glennconti]

Hi Artto, I am trying to piece together what you did by comparing Subwoofer response test 1 to Subwoofer response test 8.

I have read the DEQ2496 manual and am wondering about your equalizer setup. From your frequency response chart it looks like you have 1/12 octave resolution if I'm not mistaken. How are you able to acheive that? Or maybe the question is how did you make that RTA printout. It looks like RTA for the DEQ2496 is only 1/6 octave resolution.

And what is FEQ? You show that on your printout as a DEQ2496 EQ setting and there doesn't seem to be a menu for that on the Behringer.

It looks like you are using a combination of both GEQ and PEQ to achieve your flat response.

Is it correct that you agree to the following?

"If used as a room response correction device (as I
do), I would recommend avoiding the Auto EQ function completely as it's
rather OTT and will radically alter the sound balance of your carefully
selected amp and speakers. Which assuming you like them originally, is
not a great idea.



Instead I would suggest using the mic and RTA feature with the Pink
Noise generator and just try gently targeting obvious peaks with the
Parametric EQ function. This allows you up to pick up to 10 target
frequencies where you can boost or cut (though I'd always just cut and
not boost as this is less intrusive) and vary the width of the bandwidth
and frequency of each. This is also the best way to deal with bass
nodes, although in this case I'd use this tool identify the frequencies
and add a few db of attenuation at a time and listen for improvements"

Thanks in advance for your reply. []

[glennconti]

The DEQ 2496 is very handy device... The RTA function is very quick to setup...

Yeah but what you failed to realize is that the DEQ 2496 only has 1/6 octave RTA function resolution. This is not accurate enough to do proper subwoofer integration which needs 1/12 octave resolution. []

Therefore it looks like I'm going to have to get more hardware to interface a mic to my laptop. That is I am going to have to convert my laptop to an RTA. There is well documented software, TrueRTA. And I think REW will work also. Any comments would be appreciated.

[Quiet_Hollow]

Yeah but what you failed to realize is that the DEQ 2496 only has 1/6 octave RTA function resolution. This is not accurate enough to do proper subwoofer integration which needs 1/12 octave resolution.

Although the higher the filter resolution, the merrier it would seem...there is a point of diminishing return and it lies at the 31 band EQ. It's not a matter of electronic capability, it's a matter of physiology...based on the biological construction of our ears and the nature of perception. It boils down to the difference between what it takes to make a microphone happy vs. what it takes to actually be heard. At the lower frequencies, the ear is less physically discriminate...the usual number of filters available at those lower frequencies reflect that.

See:

http://www.sfu.ca/sonic-studio/handbook/Critical_Band.html

http://www.feilding.net/sfuad/musi3012-01/demos/audio/100_critical_bands.htm

and

http://en.wikipedia.org/wiki/Critical_band

[Quiet_Hollow]

Yeah but what you failed to realize is that the DEQ 2496 only has 1/6 octave RTA function resolution. This is not accurate enough to do proper subwoofer integration which needs 1/12 octave resolution.

Please excuse my curt response in advance. I just got done typing a 45 minute detailed response to your message, only to have this POS[8o|] billboard software vaporize it! [:@] If the moderators release my message ( would be a F****ing miracle) , you'll see what I originally posted.

For now..google "critical band" and read up on how it applies to determining filter points.

[glennconti]

Please excuse my curt response in advance. I just got done typing a 45 minute detailed response to your message, only to have this POS billboard software vaporize it!

I am truely sorry. I feel the loss as much as you do. I have done some research on "critical band". Apparantly it has to do with the proximity of nerve/hair combinations in the ear which cause two closely pitched sounds to not appear to sound as loud because they are only stimulating one nerve. Alternitavely it allows two more widely pitched sounds to sound louder as they are stimulating two separate nerves. The concept of a band-pass filter was implied whereas if the frequencies are both in the same band-pass they do not sound as loud as two separate frequencies both hitting distict band-pass filters (band-pass filter being a nerve/hair combination).

I dont really understand the following chart but it appears to support a greater sensitivity to frequencies below 200Hz.

[glennconti]

Ask the seller for the date code and version of firmware on the DEQ2496. The firmware should be V2.5. The date code should be within the last few years.

I have heard back from the seller. The version of firmware is V2.5 and the date code is 0912. So the unit is 2-1/2 years old and has the latest firmware.

[artto]

Sorry for the delay Glenn. I've been kind of busy.

I used multiple filter types.

The first order of business was to filter out the higher

subwoofer frequencies since the built-in crossover on the Empire Epik only goes

down to 40Hz @ 12dB/octave. The Epik has pretty good extended higher response

for a sub. It probably mates well with speakers that don't go down very low but

this creates some interference problems with Klipschorns which go pretty deep

on their own.

So the first thing I did was to use the Graphic EQ (GEQ) to

reduce all the frequencies above 50Hz at maximum cut of -15dB.

I then used the GEQ to generally smooth out the larger bumps

below 50Hz using cut only. Test Response 8 shows the center frequencies and the

amount of cut. The GEQ is fixed at 1/3 octave.

Then I used the Parametric EQ's (PEQ) shelving filter

capability to create a low pass filter at the maximum 12dB/octave. Here I

actually played with the frequency entry point that gave the best response on

the meter and by listening. It turned out to be around 85Hz.

Then I used the Parametric EQ (PEQ) to make further adjustments

by targeting the specific frequencies. The PEQ bandwidth is variable down to

1/10 octave. I've shown the bandwidth and amount of cut with the center

frequency used.

After that I still had some minor issues so I decided to try

and use the Feedback Destroyer as a "sharp shooter" EQ (FEQ) which is a cut

only filter but has very sharp bandwidth varying from 1/10 octave to 1/60

octave.

I know it seems kind of weird that the FEQ would make any

difference at that point but I tried to duplicate it with the PEQ and couldn't

so I went back and used the FEQ in addition to the other filters.

After all that I implemented the time delay to put the subs

in the same time domain at the Klipshorns. Looking back I probably should have

done that first.

I haven't tried the auto EQ yet. And I didn't have the Dayton calibrated mike at

the time (they were out of stock). Behringer recommends not using the Auto EQ

on frequencies below 100Hz anyway.

When I acquired the second DEQ2496 I used it for the center

channel and implemented the time delay first. I then used the GEQ to cut all

frequencies below 50Hz -15dB.

Hopefully after I get this CD project knocked out I'll be

able to try the Auto EQ on the Belle center speaker. In the meantime I'm

probably going to get a third DEQ2496 for the mains and try the Auto EQ on

there too.

I always try to use cut and avoid boost. The only frequencies

I might like to boost are below 20Hz and the DE2496 doesn't allow center

frequencies below that anyway. And since most microphones are already rolling

off down there the frequencies below 20Hz are in reality, probably already "boosted"

due the cut using the filters at higher frequencies.

The problem with boost is its really easy to drive excessive

power to the speakers especially at extreme frequencies leading to the

increased possibility of failure due to input power or excursion. Since these

subs have built-in power amplifiers with individual gain controls I simply had

to increase the over all gain at the subs to bring the level back up to where

it was before EQ.

BTW, the correct term is "mode" not "node". They are not interchangeable

and mean different things.

[glennconti]

Sorry for the delay Glenn. I've been kind of busy.

Not at all. I'm am thankful for any input you can give me.

I used multiple filter types.

Yes I am going to use GEQ and PEQ to start off with.

So the first thing I did was to use the Graphic EQ (GEQ) to
reduce all the frequencies above 50Hz at maximum cut of -15dB.

My set up is Preamp -> DEQ2496 -> Subwoofer (w/ 75Hz lowpass and 70Hz highpass) -> Amplifiier -> Main Speakers. So I need to EQ everything.

I then used the GEQ to generally smooth out the larger bumps
below 50Hz using cut only.

I'll do the same except instead of 50Hz i'll use 80Hz.

Then I used the Parametric EQ (PEQ) to make further adjustments
by targeting the specific frequencies.

Me too.

I decided to try
and use the Feedback Destroyer as a "sharp shooter" EQ (FEQ)

I'll see how things sound and look on the RTA before I try this. I want to keep all thing as simple as possible at first.

I implemented the time delay to put the subs
in the same time domain at the Klipshorns.

I'll wait. I want to keep all thing as simple as possible at first. I may wait until I have access to RTA "waterfall" plots on my laptop.

I haven't tried the auto EQ yet.

After I get my settings saved I may experiment with this.

I always try to use cut and avoid boost.

I will follow this advice.

I am in the process of setting up my laptop as an 1/12 octave RTA. I will use the Behringer ECM8000 Microphone that comes with my DEQ2496, a USB mic preamp, and the REW software. This is how I plan to get the "waterfall" plots. It is my understanding that by adjusting time delay (I may be wrong about this - I will have to experiment) I can attempt to eliminate ringing (make the waterfall appear more 2D). I think I may have a problem with my system distingushing the pitch of distict bass notes ("one note bass"). Edit: On more investigation, the "one-note bass" problem appears to be able to be solved through the use of frequency cuts and/or bass traps. I just need to make sure I have this problem.

Again thank you for your feedback. Any comments are sincerely appreciated.

[artto]

1. From your frequency response chart it looks like you have 1/12 octave resolution if I'm not mistaken. How are you able to acheive that?

2. Or maybe the question is how did you make that RTA printout. It looks like RTA for the DEQ2496 is only 1/6 octave resolution.

3. And what is FEQ? You show that on your printout as a DEQ2496 EQ setting and there doesn't seem to be a menu for that on the Behringer.

Is it correct that you agree to the following?

4. "If used as a room response correction device (as I
do), I would recommend avoiding the Auto EQ function completely as it's
rather OTT and will radically alter the sound balance of your carefully
selected amp and speakers.

I just realized I didn't quite answer all of your questions.

1. The frequencies I used for test signals are based on the actual musical notes and scales. 12 notes per octave, hence the 12 measurement frequencies. The measured frequencies have nothing to do with the octave resolution of the DEQ2496 which is variable in bandwidth and center frequency.

2. I made the chart in Word since I was running many tests, more than just Test 1 through Test 8 (a,b, c, etc.) so wouldn't have to keep rewriting the same basic things over and over.

3. Actually, on the DEQ2496 faceplate the corresponding key (button) says FBD (FeedBack Destroyer) not FEQ ~ my mistake. The FBD can be used as an equalizer and it has much sharper bandwidth (down to 1/60 octave). You'll notice that the bandwidth EQ I used with FBD doesn't go any sharper than what I used with the parametric EQ. I thought I could achieve the same results using PEQ but I couldn't. FBD seems to react differently than the regular EQ filters. The smoothness of response at that point was really quite good without using FBD but as usual I wanted to see just how good I could get it.

4. I really don't like the term "room correction device" because its not really doing any room correction at all. What its doing is modifying the frequency response at the listening/measuring location relative to the interaction of the room size and proportions with the speaker and its location. There are in fact some things like frequency cancellation that these devices cannot remove. Only having a room with dimensions much larger than the longest wavelengths, and to some extent bass traps, can do that.There are some instances, such as the Musikverein Concert Hall in Vienna where these dimensions are used to the hall's acoustical advantage to reinforce the lowest frequencies, but these are still relatively large spaces compared to residential rooms. Also, "avoiding the Auto EQ function completely as it's
rather OTT and will radically alter the sound balance of your carefully
selected amp and speakers" ~ so what. The sound of these has already been radically altered by the room itself, that's the whole reason for acoustical treatment of the room in the first place, to make it as neutral as possible so that not only do the speakers "disappear", but also the room.

A little more clarification on the time delay. In most situations, depending on where you locate your main speakers verses the subwoofer, it's probalby the main speakers that will need some delay. Indeed, thats why SVS for instance now includes a time delay on the main speaker pass-through on their subs. My situation is a little different. The Khorns are corner speakers. And the sound from the woofer travels about 40 inches before it reaches the corner. My subs are not located in the corners. I may try them there eventually but two of my concerns are reflections from the mid/treble horns off the tops of the sub cabinets, and two, the opposing subwoofer drivers facing each other causing cancellations between the subs (two subs on each side lengthwise in front of the Khorns. Placed sideways the cancellations would probably not be a problem but then the subs are blocking the mouth of the Khorns which is why I decided to place them along the front wall in 1/4 space fashion). The result is that the subwoofers are about 9.5 feet closer to the listening position than the Khorn woofers. They were producing a "prominence" effect in much the same way that a center speaker without delay does (which why in the old days the center speaker volume was usually reduced by -3 to -6dB). Getting the subs in the same time domain as the Khorns first would probably have reduced some of this prominence effect from a listening point of view. In the end I guess it doesn't really matter since I EQ'd the subs by themselves without the mains running. But as I go through this a little further taking measurements with everything running at once I may find that I need to make some more minor EQ adjustments.

Getting the subs and the center speaker time delayed in the same aproximate time domain as the Khorns has made a really big difference in overall system "coherence", improving clarity and definition, especially on recordings that were made with retaining the sound of the original space in mind.

[glennconti]

Artto, thank you for your reply. Everything is becoming clearer.

I received my DEQ2496 today. I am going to get a mic cable by mail Monday. So in the mean time I wanted to play. The DEQ2496 was a snap to cable up (Main In and Main out). I powered it up and restored it to the initial factory settings. I then played the Rives 1/3rd Octave Test Tone CD and checked the levels on my Radio Shack analog SPL meter. For each test tone/meter level, I adjusted manually the DEQ2496 GEQ 1/3rd Octave. The graph is now the equalized frequency response of my stereo/room combination. (I made sure to primarily cut not boost).

When I get my mic cable and USB Mic Preamp with Room Eq Wizard, I'll be able to improve the resolution using PEQ and FBD.

[glennconti]

This is my new frequency response chart as composed by Room EQ Wizard. Green is with no EQ. Purple is with both GEQ and PEQ applied. The region around 2KHz I over equalized so I fixed that and didn't resave the graph.

[glennconti]

Before Equalization:

[glennconti]

After GEQ and PEQ:

[glennconti]

This is the final waterfall chart. It shows minimal ringing and level bass frequency response. I got the author of Room EQ Wizard, John Mulcahy, to comment on the chart over at the hometheatershack REW Forum and he said it "looks pretty good []"

All in all, integrating my Sunfire sub to the Heresy 3's was a straight forward and fun project. And now I get to sit back and enjoy. Thanks to Quiet Hollow who first suggested me getting a Radio Shack analog SPL meter. Thanks to willland for getting me one (Radio Shack analog meter) even though it is discontinued. And thanks to artto for suggesting me getting the Behringer DEQ249 and helping me figure it out. There are lesser models than the DEQ2496 but it is perfect for me because in addition to the Parametric EQ it also has Graphic EQ which I chose to use for the higher frequencies. Thanks to John Mulcahy for writing REW and providing it for free (donations optional). And finally thanks to all the other posters that commented and helped me out.

I think the Heresy 3/Sunfire combo is better than Cornwalls alone and dare I say it, I think my setup will rival a pair of La Scala's alone with better WAF too.

[artto]

Glenn, congrats, the last (final) waterfall looks pretty good.

What role did REW play in this? How did you use it? Did you use REW to program the DEQ2496? Or did you just use it for analysis only and then made the adjustments on the DEQ2496 manually?

I recently got REW but haven't had the time to get very far into it and according to the REW documentation the DEQ2496 is not supported which led me to wonder if the documentation was not up to date or if the DEQ2496 was similar enough to the Behringer DCX2496 for REW to support it.

[glennconti]

I used REW to calculate the PEQ filters for me and then I manually programmed them into the DEQ2496. The REW software works for the DEQ2496 the only trick is you need to tell the software that you are using the FBQ2496 as an equalizer. It works great. You "Measure" the room then you select "EQ". Then "match response to target" then Look at the "EQ Filters" it has calculated. I only had 5 PEQ filters so it was a snap to program the DEQ2496.The latest software is 5.09 beta or 5.0 standard release. I am using the 5.0 standard release. There is a MIDI interface from REW to EQ device but I didn't trust it since I couldnt explicitly select DEQ2496 as an equalizer type. Programming manually was a snap though no need for the MIDI interface.

So programming the DEQ2496 was a 3 step process for me.

1. Play the 1/3rd octave test tones on the Rives TEST CD 2 and read them with my Radio Shack Analog meter. Adjust the GEQ settings on the DEQ2496. Then play the 31 test tones again while reading with the RS meter to make sure every thing had leveled out. This included the bass test tones.

2. Using REW and an ECM8000 mic play the sweep to measure the room/stereo combination. Let REW calculate the PEQ filters for me and manually program them into the DEQ2496 PEQ. The PEQ was only in the range of 20Hz to 100Hz. Tweak any GEQ settings that may be off from step 1.

3. Run a final sweep with the REW and ECM8000 mic to record the final results. Note I did not need to use the FBD on the DEQ2496. I am just running GEQ and PEQ.

And then the fun part - listen to my collection to see how good all the music sounds.