[gerG]

Thought so...



Sorry about the quality (of the image, I mean). Anybody know how to go straight from an Excel chart to a JPEG? Intermediate software is goofing things up a bit.

Ok, lets see, red trace with the accentuated midrange is the W2002, blue trace with the tilted mid and ripple at 3 to 4 khz is the AD10, and the relatively level pink one with the smooth step down from 3k to 4 k is the HD600.

There was a change in my technique. I used a TAH amp instead of the Corda, since I was mobile at the time. I will re-test the Sens with the same amp when I get a chance. I believe that the response pretty much matches what I heard during an admittedly hurried listening session.

Since I can't sneak out without making some comment about the sound, I will only say that I have concluded that AT cans do not appeal to my ears. See my AD10 review for problems and potential solutions.

As always, if at all possible, listen before you commit.

gerG

 

[Matthew-Spaltro]

I can't make heads or tails out of this graph.

 

[ServinginEcuador]

That bump in the high-end of the AD10s might explain some of that not liking them bit!!! Man, that is really quite bad when compared to the other cans!! It does look like your measurements accurately show the midrange recession on the Senns quite well!!! The low end really shows the strength of the 2002s as they bested both the AD10 and 600s, but the difference of the 2002 over the 600s is measurable, but not hearable tho!!

Great job gerG!!! Please continue to post your readings on the cans so we can see graphically what we hear audibly!!

 

[Number9]

Whoa ... maybe I'm colour-blind, but it looks to me that the W2002 has got the bump not the AD10.

 

[ServinginEcuador]

Quote:

Originally posted by Number9 Whoa ... maybe I'm colour-blind, but it looks to me that the W2002 has got the bump not the AD10.

Number9,

Actually, you're correct!! The high end bump goes to the w2002, and the low end drop goes to the AD10. Good catch!!! Guess I'm the color blind one here!

 

[Joe Bloggs]

Quote:

Originally posted by ServinginEcuador The low end really shows the strength of the 2002s as they bested both the AD10 and 600s, but the difference of the 2002 over the 600s is measurable, but not hearable tho!!

Why do you say that? Because of the bump between 10Hz and 20Hz? Ya can't hear below 20Hz, ya know

 

[wallijonn]

#9,

It's not the humps so much, but rather the dips. When you hit a dip the music disappears. In the case of my Ad10s there's about a 23 dB dip. Talk about losing flavour! What instruments play at about 7500 Hertz? since +3dB doubles the sound, or -3 dB halves the sound, a 23 dB dip translates to about a four fold decrease in sound.

To be fair, now that my AD10s have the Bayer DT831 pads on them, they are MUCH more sonically pleasing. I still have to pinch the pads down (flatten them) in the front so that the driver angle is increased, and the driver axis is more in line with the ear canal. I also have to put the pads in a more downward position - it opens up the midrange. But as I tend to this with my K1000s I have no problems doing the adjustments - along with the prefunctory large yawn to clear out the ears of internal pressure before listening.

The W2002 has the same 23 dB dip, but it occurs at about 12,500 Hz, well outside the main range. In this range I would expect an increased lack of decay detail.

I believe the W2002 also became brighter when the front of the pads were squished in, which points back to the angled driver. Perhaps it is because of the metal grille covering the cone, but likely it is a function of the cone material itself. I wish someone would make a cone where the dome is thicker at the base and very thin at it it's direct centre. This may produce a better frequency response. Uniform-thickness cones don't seem to be able to react fast enough to transients.

What that graph, like all graphs, fail to indicate is the sound one hears. The W2002 has a lot of high end energy; much more so than the AD10 or the K501. My AD10 and K501 sounded lifeless in comparison to the W2002. To make sure that it wasn't because of my Crown D60 amp, I made sure that I connected all headphones to the Melos $1000 amp AND the Crown D60. In all cases it was the headphones which had the greater differences, and not the amps. The $100 Crown made a good accounting for itself, but the Class A transparency was obvious; the mids and highs were much cleaner on the Class A equuipment.

As the science evolves we may be able to micro analyise the graph and correctly correlate it from one headphone to another. But as yet they are but indicators. The W2002, as any $1000 headphone would, sounds superb. One may just have to choose wisely the type of music one listens to. In the case of the W2002, it should lend itself to vocalised music quite well. Whether one prefers to listen to male vocals or female vocals is another matter. From what I can see, Jazz vocals may sound outstanding on them, wheras female opera singers may not.

It would be interesting to see how different "neutrally sounding" mesh materials may affect the sound. Perhaps a tightly woven material may diffuse the sound more and help to equalise the sound energy eminated. i think a small layer of cotton batting may change the whole graph. But the inherent weakness is still in the Audio Technica driver itself, and I fear that ALL Audio Technicas suffer the same defencies, to some degree or another. If the A1000s are supposed to be the W2002s without the wood, I would expect the same response from them. It would be interesting to test A900s and A9Xs to see if, in fact, they show the same tendencies. (Are you listening Pikawel?)

Greg - why wasn't my favourite DT831 added? I found that the DT831 highs were nicely tamed with the Crown D60, whereas they were a little too bright with the Class A equipment. Clean and clear, but just a little too much high end energy.

 

[ServinginEcuador]

True Joe, but you can feel it!! But, in reality the difference on the graph looks like a mere couple 3db or so.

I also put in my original post that it would not be hearable tho!!

They seem pretty equal throughout the entire bass spectrum, within about 1-2db of each other all the way through. Funny how the big drops of one can at the high end is a peak for another can, while a peak is the drop of the other!! Seems to happen 5 times between the 600 and 2002. Whereas some of those peaks are similar between the 10 and 2002. None of which tells me which set of cans I would like the best, just good info!!

 

[ServinginEcuador]

Quote:

Originally posted by wallijonn [BWhat instruments play at about 7500 Hertz? since +3dB doubles the sound, or -3 dB halves the sound, a 23 dB dip translates to about a four fold decrease in sound. [/B]

FYI - 23db is a 200 fold increase or decrease in sound. Every 10db is ten times, while 3db is doubling. So, the first 20db would be 100, the last 3db twice, making the total 200 times.

 

[Joe Bloggs]

I think those peaks and dips at the upper frequencies are mostly interactions between the phones, the enclosure and the acoustic coupler gerG used. May not tell you much about how the phones sound on your ears

Funny, I thought gerG's handle was Greg Freeman. Has he managed to change that somehow? Or has my memory gone bad?

 

[Number9]

Joe Bloggs....
I was referring to the big rise at approx. 2000 hz (if I read the scale properly) for the W2002.

This would typically add a midrange forwardness in vocals.

If you look at the smoother plot over the midband of the HD600, it is considerably flatter in the range from approx. 200 Hz to 3KHz.
This is more ideal.

Dips and valleys at the very low frequencies, and beyond 7Khz, our ears are not as sensitive too.

 

[wallijonn]

Joe,

Greg is gerG (just put the emphasis on the end instead). gerG is Greg spelled backwards. Someone must have been watching the Three Stooges that day, "...Sunev. That's Venus spelled backwards."

Yes, the microphone is probably issueing the response at the top end. But if we can only get the Senns to bounce back at 6000 Hz... btw, Greg's working on just that very problem.
---------
SE,

I don't think that the measurement can adequately trace the Senn midrqange recession. I tend to think that it has more to do with the voice coil construction itself, or the damping ring. After all, if it was just the frequency response, one could conceivable get rid of the veil through equalisation, or an elaborate equalisation circuit. (See the Headwize site for the DT770 equalisation circuit for comparison).

Basically I would like to see a microphone tip that has a 90 degree angle to it so that it can be directly aimed at the transducer. As it is, it's picking up sound waves as they pass the microphone voice coil, and so fail to adequately reflect direct energy, even if it is pink noise. Just as changing the angle of the transducer on the headphone causes changes in the sound quality, changing the angle of the microphone may cause a change in the frequency response. What is undisputable is the fact that what gerG hears and what is represented closely resemble each other. So that validity can not be questioned. And I have seen gerG go through the process of optimising the energy response, so these are best effort represenations (of the headphones themselves).

 

[halcyon]

Nice graphs, but do they tell us anything useful?

Headphone measurements should be done with head coupling and calibrated test gear to tell us something useful about how they might sound. Otherwise the graphs may not tell anything about how the phones may sound. This is a sad, but a known fact.

Especially in this case, when we are comparing open and closed headphones. The closed headphone will measure totally differently with an artificial head, depending on microphone position and ear type used. Even the open headphone (HD600) will measure differently (see graph below).

The results after 2 kHz can be so varied as to be considered totally misleading. The data after 8 kHz can be discarded altogether as it propably has no relevance to what we are actually hearing

Compare the above HD600 results to these measurements done with an artificial head and proper calibrated test equipment and good methodology (not done by me, they are from a Master's thesis by Toni Hirvonen).

Do the HD600 graphs look similar (if you correct for the fact that they are plotted to different scales)?

I don't think these measurements serve a lot of useful purposed for drawing conclusions on how headphones might sound.

They probably have a useful application when modifying headphones and trying to understand what the mod did to the sound pressure characteristics (can be different from how it sounds!).

Please understand that I'm not trying to start a pissing contest about who has the best measurements, but I try to let people to know that they shouldn't make the same mistakes as I did, when I first started looking at graphs like this

Best regards,
Halcyon

PS Graphs have power. We project our beliefs, fears and hopes into them. Let's not try and start using these graphs as an argument as to how a certain headphone might sound. Trust your ears.

 

[Tomcat]

Quote:

Originally posted by halcyon Nice graphs, but do they tell us anything useful?

I am afraid halcyon is correct. Just look at the virtually identical bass roll-off all three headphones exhibit below 150 Hz or so. How likely is that? I'd say this a strong indication that the test set-up has been less than ideal.
Quote:

Graphs have power. We project our beliefs, fears and hopes into them. Let's not try and start using these graphs as an argument as to how a certain headphone might sound. Trust your ears.

Amen to that, halcyon.

 

[gerG]

Dang, you guys got ahead of me.

I thank the supreme being (Jude) for changing my tag, at my request. I messed up when I first registered here.

The plots tell me lots. (For details on my crude technique check my AD10 review). They certainly are not the whole story. From these curves I would draw the following conclusions:

The W2002 have a very emphasized upper midrange.

The AD10 have an emphasized upper mid, with an audible peak around 4000. Curiously the peak sounds much worse to me than the measurement would indicate.

The ripple in the response curves are caused by various standing waves in the enclosed cavity. btw, I was using a dummy head (my own

). I can change the amplitudes and location a bit by moving the microphone around. The important thing to note here is how well the HD600 control these interactions. That aside, I agree with the observations that HF response plots are tough to get meaning from. Up to 5k is pretty indicative of the sound character, though. Troughs are forgivable, but peaks cause bloom that I find very distracting.

All three of these cans have very similar bass response to my ear. They are all thin, and the traces indicate the same. With the AT it is easier to lose the bass line under the mids than with the Sens, although I don't think that bass response is a strength of the HD600. All 3 cans had stock cables.

Most importantly, the sound and measured response of the AT cans will be greatly affected by the distance between the pinea and the driver. In my case the driver is actually in contact with my ear, which results in terrible peakiness and diffuse HF. A pad mod will work miracles (even on the W2002).

I put these traces up simply as information that I wanted to share. I also share my own interpretations. Please understand that they are only my opinions, and as such are insufficient information on which to base a decision. Please collect as much info as you can, preferably by listening for yourself.

Turbulent discussion of my results are encouraged! I am in this to learn, and the feedback helps guide the learning. I will continue to share what I find.

Sorry to cut this short, but I have to cover my ears and get back to work.


Stay tuned.

gerG