[gerG]

I finally got my latest toy hooked up.

I have been intrigued for some time by the idea of equalizing in the digital domain, prior to conversion back to our analog realm. I had intended to try it with my old Behringer 8024, but the digital interface is an add on option, and it may or may not be compatible with consumer digital interface standards. So, when the 2496 came out with S/PDIF included, and it is less than half the size of it’s predecessor, I had to try it!

Initial impressions are very favorable, to say the least. The user interface is amazingly intuitive. I had the unit decoding a digital signal from a dvd player within 2 minutes, and I had an eq curve programmed, named, and stored within 5. This with the owner’s manual still sealed in plastic!

Initial impressions of the sound are in the agog category. This thing brings headphones to life, and I haven’t detected any degradation in the sound, yet. I haven’t done a full bypass test yet, but I will. Overall it is a stunner.

Pictures of the initial setup shown below. The gear is just stuff that was sitting around not being used, so there is no overarching system plan at work at this point. The amp is a headroom unit that sounds great, but has the odd quirk of having the connections and controls on the sides. It is only temporary for this application. Same for the source, which is a first gen DVD player. Since I could not find any glass fiber-optic cables around town (an I don’t have time to build one) I chased down a $10.00 plastic part from Fry’s. That search took over an hour, since they have similar parts scattered all over that immense store of theirs. What in the hell are they thinking?

Going forward I intend to get a short glass cable, and find a smaller source (suggestions appreciated). I also need an amp that fits the form factor, but can dish out the power. An amp and a source that each take up a half rack space would be ideal. That would make for a tidy transportable system that would fit in a briefcase, and could bring almost any pair of headphones to life!

gerG

 

[gerG]

Oh, and for you fellow analogue fans:

gerG

 

[lan]

Sweet.

Would you happen to have a K1000 settting for an EQ?

 

[wallijonn]

Quote:

What in the hell are they thinking?

Try finding a "rare" fuse in the fuse section.

It took me over an hour - I had to look at every fuse on the rack.

 

[goon-heaven]

Hi GerG,

I've had my eye on one of these when I rebuild my Hifi. It hadn't occurred to me that headphones would also benefit.

What EQ did the Behringer decide the HD600s need?

 

[TravelLite]

That's excellent, gerG. Beautiful!

I hope you'll be able to post further impressions, especially regarding the onboard DAC, after you've had an opportunity to finalize the system.

TravelLite

 

[gerG]

Hi Ian. I do indeed have a curve for the K1000s, but it is quite subtle. Mostly a notch at 2khz as I recall. I will do a new one soon.

Wally, more and more stores seem to be going to this relational organization, where parts are located near what they presumably belong to. It is absolutely maddening if you are looking for a specific item.

goon, i haven't tried the auto-eq yet. They don't seem to work well for headphones. The curve in the photo was derived from response measurements and a bit of tweaking. It works quite well. I do intend to try the RTA out. I had pretty good luck with the 8024 by measuring the response, then shipping the curve (inverted) over to the eq side. The only problem was that the number of samples was insufficient to get a good average response.

So far I only have one gripe about the 2496: the godawful orange display! I don't know what it is, but that color makes me itch like optical fiberglass. My Delphi XM tuner emits the same sickly glow. Maybe they got a clearance deal on orange LEDs right after halloween.


gerG

 

[halcyon]

gerG,

great work.

Can I ask something I may have missed in this long thread:

Do you tune by ear or to some idealized graph?

Would you be interested in doing an example correction for HD600 that simulates the idealized diffuse field equalized graph?

 

[Steve999]

Psst --

Halcyon, your question is over my head, but maybe this will help:

gerG's full-blown DEQ2496 thread, with a microphone-calculated HD600 curve posted

I'll say this, the DEQ2496 is IT for me. I am one happy owner.

Regards, Steve


Quote:

Originally Posted by halcyon gerG, great work. Can I ask something I may have missed in this long thread: Do you tune by ear or to some idealized graph? Would you be interested in doing an example correction for HD600 that simulates the idealized diffuse field equalized graph?

 

[gerG]

Hi halcyon!

Here is an even longer thread with slightly more up to date information:
(edit: removed link, since Steve beat me to it

)

I use a test mic and rta to get me started. My first pass is to tune as flat as I can at the reference plane (mic location). I then tune by ear from there. I use some heavy averaging in the above 2 Khz range, since the reasons for many of those fluctuations are not basic response issues. I also mic the phones uncoupled to look for unusual responses coming from the rear cavity. Some manufacturers go overboard with the rear reflections , presumably to get an ambience effect.

I would be very willing to try target curves other than flat. I did do an experiment where I set up a speaker for flat response in nearfield, then measured it in a room at a distance. The difference gave me a correction curve, albeit not quite diffuse field. I also repeated the test with the mic on my ear at the same location that I test headphones, giving me my very own hrtf. Interesting stuff, but I never did much with it. I am liking the near flat response better.

I haven't gotten to demonstrate my system to many hard core headphone fans yet. For me, when I switch on the processor the Sens come to life, the veil is gone, and the missing bass is instantly resolved. When I do an A/B with a casual headphone listener they don't even notice a difference. The exception is my Fostex TR50P cans. In stock form they have a response that peaks in the midrange, and falls off smoothly in both directions. Flatten them out and they go from muddy and dead to detailed and alive. Everybody gets that one.

This technique does not work on all headphones. Some that I have tried either can't do the job (Sony F1) or reveal awful traits that were previously hidden (AT W1000). Sens seem to do the best with eq, as do the Beyers. The AT W100 and AD10 ar both capable of amazing detail and heft given some adjustment.


gerG

 

[halcyon]

Thanks, I tried that long thread (actually I tried to refer to that as well, but missed).

I can't find what is your reference plane?

At the headphone driver cloth?

Inside ear canal?

Headphone coupled to head or not?

I'm very interested in your experiments and very tempted to get my first Behringer (ugh) product myself.

Thanks!

 

[gerG]

All good questions.

My desired reference plane is the imaginary surface defined by the outer surface of the pinea. Unfortunately, being a virtual surface, it is tough to hit consistently, even with a fixture. I have found that I can get consistent measurements by setting the tip of the test mic at the tip of the little point of skin that guards the ear canal (is there a name for that?) and pressing the barrel of the mic against the back of my jaw so that the little flap closes off the ear canal, and the mic barrel tucks into the gap behind my jaw. This is easy with an 1/4" mic (Earthworks M30) but was painful with a 1/2" mic (the Behringer). I did hack up a mic that sits within the bowl of the pinea with the sensor at my reference plane, but the bass response is not right yet. Very slick mic but needs some more tuning. It also needs better grounding because the metal mesh that I used for shielding zaps me every time I grab the volume control.

btw, for anyone trying to do measurements like this, make certain that your mic is very well shielded. Dynamic drivers generate strong magnetic fields. Mic leads make great antennas. Having a mic so close to a strong magnetic field makes for a high risk of noise intrusion. It is very tough to detect, because the noise is in the form of the test signal. Beautifully insideous, don't you think?

The headphones are coupled to my head for the test. The seal is prettey good. Since the mic fills in a natural gap behind my jaw, there is little to no compromise there.

Another test that I want to get back to is what I referred to as a "clamshell" test. I trap the mic between the earcups, clamp them together, and feed both sides the same pink noise signal. This simulates a perfectly reflective interface for a single cup, with the average of both sides measured. Not truely representative of the head, but very informative when I am trying to understand the behavior of a given headphone.

As for the DEQ2496, I am thrilled with it. It really did open up a whole new world of possibilities in the headphone realm. I can't think of any headphone related gear that I have bought that offered as much improvement for the dollar. otoh it is a complex piece of gear with an almost limitless range of sonic signatures, and it takes some work to get the sound "right". I doubt that my eq curves will work for everybody. Still, it is a starting point.


gerG

 

[halcyon]

Quote:

My desired reference plane is the imaginary surface defined by the outer surface of the pinea.

That's about the position of the concha, just after the flap you describe, which is tragus:

Sounds like a good point to measure, based on what I've read and what j-curve did with his binaural mics.

Quote:

I did hack up a mic that sits within the bowl of the pinea with the sensor at my reference plane, but the bass response is not right yet.

How do you figure this? Are you getting environmental noises at 50-60Hz and below? Are the results not in line with what you hear?

Quote:

This technique does not work on all headphones. Some that I have tried either can't do the job (Sony F1) or reveal awful traits that were previously hidden (AT W1000). Sens seem to do the best with eq, as do the Beyers. The AT W100 and AD10 ar both capable of amazing detail and heft given some adjustment.

This to me could suggest that you are not necessarily measuring/equalizing to a "natural" reference, but are bringing up detail that is lost in normal non-headphone listening as well, when using some desired curve.

However, this desired curve changes based on how the headphone is coupled (and perhaps other characteristics). I'd figure that F1 is a case onto it's own, but it's interesting that the closed W1000 cans don't fit your approach so easily.

Of course, the above is just pure conjecture and not meant as a criticism (esp. when we are talking about subjective pleasure, not objective reference).

I know from experience that it is easy to move normal masking tresholds down by equalizing a headphone just by using a cheap 10 band software equalizer. Of course, a lot of "additional detail" is gained perceptually, but as you don't hear those details naturally at all (without the headphones), it seems to me that the detail is then "artificial" (i.e. hyperreal). Augmented hearing, if you will

One more question:

How flat do you aim beyond 2kHz or do you just smooth out things beyond 2kHz (at your mic recording position)?

I'm just wondering about your personal HRTF. Have you tried to measure your own HRTF with a speaker setup (I can't recall and search is down, so can't check your earlier posts).

Your stuff is very interesting indeed.

I'd like to play around with this stuff myself, but other current purchases are putting this on hold for me, so I have to pester you about the issues

regards,
halcyon

 

[gerG]

Halcyon, your "pestering" is quite educational and helpful. Please keep it up.

I will post a pic of my ear mic this evening. Very pretty, and an obvious hack using commonly available parts. It is mounted into an "ear-gell" (courtesy Jabra and Radio Shack) which fills in the concha and both primary grooves radiating therefrom. btw, thanks much for the diagram. I didn't even know I had an anti-tragus.

The problem with the mic is that the bass response rolls off prematurely, relative to the M30. I attribute this to how I mounted the capsule. I also need to run some magnetic field tests on it to make certain that my shielding is sufficient.

My biggest problem with the W1000 was when I tried to correct for their bass rolloff. Boost in the low end made it apparent that they have a resonator at work in there. Results were very boomy, sloppy bass, even though the response was coming in flat. More proof that spectral response is only part of the story. It is certainly an artificial result relative to the uncorrected sound of the cans. otoh once it became apparent with correction, I could still hear it uncorrected. At least, I believed that I could. Funny how adaptive the brain is when it comes to processing sound.

Beyond 2K is very treacherous, and takes some careful evaluation. Some of those dips and peaks are attributable to the basic headphone dimensions. I look for 1/4 wave and 1/2 wave reflective distances. These correspond to 180 degree and 360 degree phase shifts when the path length is doubled. These can be equalized some, but not completely. I have an issue with equalizing to correct reflections (in rooms and cars, as well as headphones). It is a bad thing because it only treats the level, but not the phase or direction error. I would much rather treat the cause of the reflection. If serious enough, it may be worth modifying the headphone to tone it down a bit.

There is also usually a hf peak due to out of phase energy from the back of the driver reflecting off the shell or screen, and coming back in through the windows around the driver. This can be sealed off mechanically, but tends to play hell with perceived soundstage, and will do some odd things down in the bass range. Equalizing that peak retains the soundstage, and helps with the sometimes overbright nature of the particular headphone. The DT931 is my favorite example. It has very flat response except for a narrow hf peak that seems to be a deliberate shell reflection. Taking out that peak with a simple parametric takes away the brightness and yields a wonderful sound.

Other hf variations I haven't tackled yet. One of my rules of thumb is smooth eq curves, and the hf stuff would require way too much yanking around. I just shoot for an overall level that gives me the detail that I want without being harsh or sibilant. After my initial cut at flat response I usually just smooth out the curve and back off the corrections. This helps retain some of the character of the specific headphones. They wouldn't be as much fun if they all sounded exactly alike.

I did measure my own hrtf, but with my old mic and A/D. I will be setting up a speaker system this weekend, so it is a perfect chance to try it again.


gerG

 

[gerG]

btw, this is (almost) what the setup looks like now:

The Grace has since been centered up and set with HD velcro. The 963 is shock mounted with foam, so I can pull the Grace and cords, put the cover on the case, stand it on end, and carry it with a single hand (leaving the other free to carry my scotch, of course).

The mic I was talking about is a complete hack-up of the Behringer mic. Only a prototype, and more work required:

The tip was trickier than it looks:

gerG