[Danny Richie]

Tyson,

The only way to really see what the speaker is doing (and not the room) is to take a gated time window measurement from about a meter away (no room reflections). That would be a semi-anechoic measurement.

Near-field measurements of the drivers really won't tell you anything either because that takes away a critical phase relationship between your drivers that very critical.

Room correction can be somewhat useful but only in ranges below 200hz. It should never be used above that. The reason being is that it really doesn't correct a problem. It turns down a problem area. In other words, let's say that you have a peak caused by a side wall reflection. You can turn down the amplitude in the frequency range to compensate but now you have really decreased what you hear on axis compared to the reflection to even it out. It is still heard by the ear as a disruption in the sound stage, and the imaging and sound stage cues are compromised.

The only real way to correct room related problem is with room treatment.

However, I did not bust out the mic just because I felt like measuring my speakers for the "fun of it".  I pulled it out because the top end started sounding a bit hot and spitty.  And that corresponds to the change in overall FR shape between the graphs.  So, I adjusted my EQ to compensate for the dip about 1khz a bit more and it started sounding tolerable again.

The "really" odd thing is that a few hours later, the system started sounding way too soft and subdued in the highs.  I didn't bother to remeasure (especially since my wife was now home), but removing the extra compensation from the 1khz range brought back a normal balance.  Weird, eh? 

That is really not that weird. It sounds like you are chasing room reflection issues. I bet I can move the mic around in your room a little and find a place that is too hot, too soft, too bright, bass heavy, thin in the bass, etc. In fact I bet I can find +/-10db swings in the bottom end real easy.

And there is nothing wrong with a well designed active system, but what you are using is not so much an active crossover as a digital crossover. So you are going to be really held back by the D/A conversion process of that unit. From an entry level stand point those things are great. They will even allow you to hit good performance levels at a mid-fi level. But from a high end audio stand point those can be a bottle neck.

I spoke to some people many times about really doing a high end version of a digital crossover. In fact I spoke to Eric from db Audio Labs about it not too long ago. He said that it is very possible and not that hard to do with their DAC's. The problem is that for a three way design I'd need three Tranquility DAC's ($4,785 retail), then six channels of pre-amplification with a common controller for all six channels... Then the six amps... And I just don't see it being worth it. It could be done and it would be really great, but it is just not practical.

[Danny Richie]

One tangential question I do have, which maybe Danny can answer, or anyone else for that matter, concerns the fact that measurements can (and do) vary based on small movements of the mic.  But, on the other hand, no matter where I sit, my speakers sound pretty much the same.  In other words, my speakers always sound fairly consistent, at no point do I think "Wow, they sound like Dynaudios from here", and this shift half a foot and think "Wow, and now they sound like B&W speakers from here".  The shifting in FR as measured by a mic would seem to indicate that there should be concurrent tonal shifts that are quite dramatic.  But, for the most part, my speakers sound pretty much the same regardless of sitting position, at least tonally speaking.

No problem.

When you measure one speaker from one point in space (including room reflections) you do so with a linear output (the same at all levels).

When you listen to your speakers you are hearing two of them playing and with a non-linear output (the music varies all over the place).

Small narrow band peaks and dips are easy to measure. Just look at your measurements and you can see them all over the place.

Your ear hears more of an average level of those. In other words you can't distinguish between them too well. For instance, look at your first measurement. You can tell me from listening that you hear a peak at 1.35kHz and a dip at 1.1kHz even though they represent a 15db swing. So a narrow band change is tough to hear, but a broad band peak or dip is pretty easy to hear.

Now if you get real used to listening to an MLS response from a single speaker (I have) then you can eventually hear the response and draw the graph without seeing the computer screen, but it takes a lot of practice. Even still the narrow band changes are tough to catch.

[Quiet Earth]

Interesting thread. Not sure why the emphasis is on making so many measurements, and I don't believe that making substantial adjustments is warranted during a break in process, but still, it's interesting all the same.

 I personally believe the only thing you really need to measure (and possibly adjust) during a break in process is your patience. 

[Tyson]

I think maybe you are confusing the DEQX with a DCX?  Here is what I have:

http://www.deqx.com/products.php

Pic:



The DEQX has a huge number of advantages over something like the DCX (which is not a bad unit, but it's pretty mediocre parts quality out of the box and really only lets you mimic passive crossovers in the active domain).  I know that Steve at Empirical loves the DEQX and did a bunch of upgrades/mods to the 2.6 unit, which was later incorporated by the DEQX engineers in to the 3.0 version, which is what I have. 

But, I am GLAD we've had this discussion, because you've given me the impetus to do some 1m gated measurements.  I've done them before, and they were "OK" as a means of doing basic driver measures, but I found that I liked nearfield better for the initial setup/correction measurements on the drivers.  However, after I did the nearfield measurements, I went straight to farfield/room measurements, and I realize now that is a mistake.  I think the 1m gated measurements will let me dial in the baffle/box/driver integration a bit better than what I've been able to do so far.

My wife and daughter are almost home now, so this will have to wait, but I'll try to re-measure sometime this weekend.

Oh, and one cool trick I learned about 1m gated measurements - the standard approach is to lift your speakers 1m off the ground and then do the measures.  But you can get a similar affect by simply tilting your speaker back and raising your mic to be on the tweeter plane (or midrange plane, or woofer plane).  My speakers are HEAVY, so not having to lift them on to a ladder or stool makes doing gated measurements MUCH easier.

[Steve]

Interesting thread. Not sure why the emphasis is on making so many measurements, and I don't believe that making substantial adjustments is warranted during a break in process, but still, it's interesting all the same.

 I personally believe the only thing you really need to measure (and possibly adjust) during a break in process is your patience.

I had an interesting experience, some years ago a friend purchased a pair of madisound monitors using MTM design, two scanspeak woofers plus tweeter.

He brought them over and I installed the crossover and listened for awhile, in my marked speaker spot. The bass was tight and sounded wonderful.

He took them home and a couple of weeks later stated the bass was not right, too bloated. So he brought them back and I tried them again. Wow what a difference. The bass was extremely bloated. I tried moving them around and still way too bloated.

So bloated in fact that he took the speakers and lined the inside with lead sheeting on the walls, dampen the cabinet wood, and it was still way bloated. Then he packed the entire cabinet with fiberglass. Still way too bloated. Eventually he gave them to his son to use.

Interesting experience indeed.