AutoEqualized Fun

[Dreadhead]

Today I finally made my fake ear (block of modeling foam with a hole) for my measurement microphone and hooked it up to my digital equalizer. It's bloody cool, you put the heaphones on the ear and run the auto eq (automatic algorithm using pink noise and a RTA) and you've got pitch flat response

I know for a fact it would drive many nuts but to be honest I'm having a great time. The only boring thing is that it makes d5000s sound like sa5000s or whatever sound like whatever (as long as both are equalized).

Anyway I look forward to trying this out with other phones too. It'll be interesting to see how much correction is required. For the d5000 it's a lot down low and for the SA5000 it's more even (as I would expect). Both have quite a lot of correction up high too.

[grawk]

How does it correct for the flaws in the microphone?

[Dreadhead]

Well it's a fairly cheap measurement mic but it's calibrated to +/- 1 dB from 20 Hz to 20 kHz and the auto eq is accurate to about that range too. It all depends on how patient you want to be to let it do its thing but in the place where the headphones have narrow banded effects the alorithm has to take the best fit but can not get exactly flat.

Perfect it is not, but it is better than uncorrected. In the end throughout most of the fequency range 1dB is below audible differences in levels (or at least that's my understanding).

[grawk]

I'd have to play with it I guess. My experience is that it's more complicated than just frequency response, but I'd love to see what it can do sometime.

[Dreadhead]

I'd have to play with it I guess. My experience is that it's more complicated than just frequency response, but I'd love to see what it can do sometime.

Certainly is is more complicated. I'm not saying the phones sound exactly alike. The control and pace are still whatever they were before. For example I still think the SA5000 has better control than the D5000 but to be honest the difference is MUCH smaller; it's like taking the difference between different pairs of headphones and making it the difference between two SS amps for easy to drive phones.

One thing that is truly interesting is how the character of the pink noise changes as the auto-equalization goes (you can always hear the output of the other ear). It really shows you how much you're changing the character.

Oh and I will warn you that all of the phones I have done this for now sound bright at first. The SA5000s have increased highs and even the D5000s have a couple dB of correction upwards. At first it sounds odd but it grows on you fast when everything sounds so right. At first I was like "what the hell" and after a well done classical recording all was well again (the drums still had great impact and).

The DEQ I have is the Behringer Ultracurve Pro DEQ2496 (used entirely digitally in this case) and the measurement mic is the Behringer ECM8000:

AUDIO TECHNOLOGY - EQUALIZERS & ACCESSORIES - ULTRACURVE PRO DEQ2496 : Ultra High-Precision Digital 24-Bit/96 kHz EQ/RTA Mastering Processor

MICROPHONES / HEADPHONES / WIRELESS SYSTEMS - CONDENSER MICROPHONES - MEASUREMENT MICROPHONE ECM8000 : Omnidirectional Measurement Condenser Microphone

I am still a little confused on the ECM8000 because its an omnidirectional mic so I thought it did not have the proximity effect and one box it comes in said it does not have it and then the inner box around the carrying case says it does. I included the supposed effects of this in one of the equalization I did (by modifying the target curve) and in the end the phones seem very very bass heavy which made me conclude it did not have a proximity effect. I'm going to email Behringer and make 100% certain. If I'm wrong I'll switch to doing that too.

[darkless]

Same DEQ and mic as I have.

[Pars]

Used to have those... worked pretty well once you figured out how to calibrate it in auto-eq (speakers->room correction).

[darkless]

Too bad its ability to do auto-EQ sucks major balls compared to using Room EQ Wizard and then plot in the final corrections by hand on the DEQ2496.

I know I've mentioned the software recently in the speaker forum too, but seriously... it's just that good. Just for the record, though, I did my measurements using a RadioShack SPL meter (analog version), because the DEQ2496+ECM8000 combo didn't allow me to send any measurement data back to my soundcard.

[Dreadhead]

Too bad its ability to do auto-EQ sucks major balls compared to using Room EQ Wizard and then plot in the final corrections by hand on the DEQ2496.

I know I've mentioned the software recently in the speaker forum too, but seriously... it's just that good. Just for the record, though, I did my measurements using a RadioShack SPL meter (analog version), because the DEQ2496+ECM8000 combo didn't allow me to send any measurement data back to my soundcard.

So you're saying an SPL meter with a stated frequency range cutting out above 100 hz (I don't remember the exact number it's probably even higher, I have the same one) is more accurate than the ECM8000/DEQ combo? I don't buy it.

That software looks neat though. My FA66 has a 48v phantom powered input that I could hook the ECM8000 to and do the same thing.

Cheers,

Chris

[Hopstretch]

A lot of people use the ECM8000 with REQ. There are calibration files for it knocking around on the Home Theater Shack site where the software is hosted.

[darkless]

I retract my earlier statement. I see that I made the mistake of not stating that I used the SPL meter with REQ solely for calibrating the lowend output of my SVS PB13-Ultra subwoofer and my speakers playing together. I did not make any corrections above 100 Hz.

I've just played a bit with the auto-EQ again with my Audio Note AN-E/L speakers, and this time the result was much better than the last time. However, I'm still not quite sure what values I should consider sane for the "delta" and "max span" settings used when letting the auto-EQ do its thing.

[Dreadhead]

I messed around with REQ through my FA66 and it's downright impressive. I stupidly saved over the auto-equalized version but for my SA5000 I started with the auto eq and now I have (after a lot of tweaking) pretty much flat response between 20 and 20kHz with one dip at 6.5 kHz that I just can't get rid off but it is very narrow now.

I'll rerun the DEQ based auto-eq and compare the two with my ears and see what happens but it looks like it seems to accentuate the highs more than DEQ (I used the ECM8000 calibration file).

[Dreadhead]

Man oh man I should reread these things before I hit submit I apologize for all the spelling and grammatical errors in the above posts.

My guess is that this is working so well in my system because my GS-X still has a lot of juice to spare for these phones and can deal with the increased outputs where needed. In the end most of the corrections are on the order of 3-5 dB with 3 big ones (10 db parametric) to correct for some resonances (funnily they come out as dips not peaks). Maybe this will lead me back to a beta22 but I doubt it.

Anyway I've been giving it a lot of thought and I think that I'm no longer in the market for the Qualias. The equalizing has fixed the differences between the Qualia and the SA5000s. I think I now have phones that may beat the Qualias (The bass is certainly better: just as controlled with more punch) and all that for under a grand (Headphones, DEQ, mic and foam). I am going to let my WTB over at HF fade away I think and just be happy with what I have. I'll get a chance to hear Qualias again in a couple weeks so I can do a comparison then but I don't really expect any surprises. Of course I'm probably entirely wrong

[Hopstretch]

Very interesting. I've never even considered trying to EQ headphones, as I've always thought of it as a way to deal room modes rather than alter driver responses. Could you post a picture of the setup in action, ie. phones on the "ear" showing mic position?

[Dreadhead]

[Dreadhead]

[Dreadhead]

Red=uncorrected

Green=hand corrected (different than yesterday???)

blue= Auto Equalized

[Hopstretch]

I guess that's what they're talking about when they talk about a mid-bass hump!

[Dreadhead]

New Hand EQ job (blue=uncorrected, gold=hand). Most of the difference between the auto-eq and the hand eq is the fact that the calibration file for the ECM8000. From Now on I will modify the the target curve for the auto-eq to match the calibration curve from REQ. In the end it sounds nothing short of amazing in my opinion. So liquid

[aerius]

In the end most of the corrections are on the order of 3-5 dB with 3 big ones (10 db parametric) to correct for some resonances (funnily they come out as dips not peaks). Maybe this will lead me back to a beta22 but I doubt it.

EQ doesn't correct for resonances unless we're talking about a crossover where the response is rolled off well before the driver's resonance point(s). Also, how do you know it's a resonance unless you've run a cumulative spectral decay plot on the headphone?

[Dreadhead]

EQ doesn't correct for resonances unless we're talking about a crossover where the response is rolled off well before the driver's resonance point(s). Also, how do you know it's a resonance unless you've run a cumulative spectral decay plot on the headphone?

You obviously know a lot more than me so I'll take your word for it. All I know is that I can not prevent the dip at 13.5 kHz (at some of the peaks at other frequencies and troughs at others) no matter what I do to the EQ. The phones just won't respond well there. I interpreted this as a resonance of some sort. If I stated that wrong then sorry. In the end it still does not retract from the fact that I have very much improved my phones performance.

[Dreadhead]

Arius sorry if I came off harsh. No harshness intended.

[Dreadhead]

Heard back from Berhinger and there is no proximity effect so I should be set

[hungrych]

Arius sorry if I came off harsh. No harshness intended.

You obviously are new here.

[Dreadhead]

You obviously are new here.