[xnor]

What I like about these compensation curves is that you can always go back to the raw data and apply different/improved curves in the future. 

 

[udauda]

I just checked the pdf files you've uploaded- how about adding this kind of distortion(THD+N) graph? The 500Hz referenced data Headroom was using does not tell much about what matters the most to listeners: the audibility threshold of distortion. As long as the distortion stays below 1%, it will highly unlikely to be audible.(with music)
 
And unless the drivers are aligned like loudspeakers(free-field: AKG K-1000 & Stax Sigma), diffuse-field compensation is the right one since that is the true representation of headphones' localization properties, as suggested by G. Theile in 1986. Someone said something about Indepent-of-Direction compensation here, but ID only removes non-directional components of HRTF, so it should not be preferred. It'd be ideal for measuring car acoustics, in which the localization component is the only necessity of the test.
 
Also, I hope you'd measure frequency-dependent attenuation caused by impedance interactions between amplifiers and headphones. For this, you need to measure headphones driven with different amplifiers with various source impedance. (or simply use many serial resistor adapters as dummy loads) It'd be a quite tedious job, yet we would be able to see how headphones react to different sources. (sometimes these interactions make clear audible difference)

 

[xnor]

Quote:

I just checked the pdf files you've uploaded- how about adding this kind of distortion(THD+N) graph?

Isn't that exactly what he posted?

 

[Tyll Hertsens]

Hiya folks, kinda buried in testing some headphones for HeadRoom, but over the weekend there will be a big push to finish up the test sequence.
 
However!!! I did run a burn in test on a pair of brand new K701s. These headphones were removed from the head each time and I burned them in on a separate system.
 
Here's the pdf of the burn in test.
 
Obviously the FR repeatability error due to placement of headphones, even when averaged over five placements each time, is greater than any burn in effect on FR.
 
Lots more things to test on the effects of burn in, but bottom line this data shows it's not large --- at least wrt the FR.

 

[xnor]

Nice. Couldn't take a closer look yet, but it's nice to see another one of those burn-in measurements, with the same results. ("at least wrt the FR")

 

[SP Wild]

Quote:

Lots more things to test on the effects of burn in, but bottom line this data shows it's not large --- at least wrt the FR.

I am curious if the THD levels might differ significantly with burn-in?

 

[Tyll Hertsens]

Quote:

I am curious if the THD levels might differ significantly with burn-in?

Me too. When I've got more time I'll do it. Likely in the fall.
 
Just to properly set the expectations around here, I'm taking some time this summer as I'm sort of between jobs to have a mid-life crisis ride on my motorcycle.
 
Meet Daisy:
 

 
She and I will be on the road together from July 1 to Sept 15. Going to WA, OR, CA, CO, TX, LA, FL, NC, NY, MA, NH, IL, then to Burning Man, and all points inbetween.  About 10,000 miles of helmet time. I need it.
 
So, come mid September I'll either have my head on straight and will be full on developing the Headphone Hobby Lab .. or I'll be in Mexico heading south never to be heard from again.
 
(Don't worry, I'm enjoying getting a serious headphone measurement capability together for the hobby, I'll be back .... pretty sure.)
 
Hope to see a lot of you guys at CanJam.  I'll have the system there measuring headphones, if you've got some exotic cans that don't need an amp bring 'em by.
 

 

 

[SP Wild]

Cheers matey, enjoy that breathtaking scenery!

 

[Lunatique]

One thing I always wished to see was headphone measurements of models/brands that HeadRoom didn't carry, such as Stax or many other brands/models. Now that the testing is a separate operation, I'm really looking forward to seeing measurements from all the models/brands HeadRoom never carried.
 
Just out of curiosity, is it possible to test soundstage/stereo separation with headphones? In the audio world soundstage seems to be the second most talked about thing behind frequency response, and it would be totally awesome if it could be tested.

 

[xnor]

@Tyll: Enjoy the trip with Daisy! Phew, that's quiet a long tour you've planned there. I hope that Daisy's back is soft and comfortable. 

 
@Lunatique: I don't know of any soundstage measurements and think it wouldn't easy to do or even impossible, especially with headphones. With headphones the soundstage is artificial imo, left is always separated from right except you choose to crossfeed.
The only related thing you can measure would be how identical the left + right drivers behave (volume, FR, resonances ...).

 

[rsaavedra]

Enjoy the ride Tyll! Looking forward to those measurents from CanJam. 

 

[Jamey Warren]

Tyll,
 
I hope this doesn't take away from the purpose of your thread.  Since people are talking about the idea of a difference graph here, and you're the one doing the measuring for us, I thought it would be a good place to continue the discussion.  If you prefer a different thread, let me know.  Anyway, check out this 'difference' graph tool. 
 
http://www.headphone.com/learning-center/technical/difference-graph.php
 
Choose your reference headphone, maybe the best one you own, then choose a headphone you're curious about and see what the differences in frequency response is. 
 
This tool will need some refinements over time so I'm looking forward to everybody's feedback.

 

[Roger Strummer]

That is a very interesting tool, from a quick glance I see that the difference graph is constructed by assigning the reference headphone as a frequency response equaling zero in every point and then graphing the difference between the headphones around that constant line (I hope I'm being clear in what I'm trying to say), am I right?
 
If the answer is positive my first (which I highly suspect it is), a small suggestion is to add a constant line at db=0, because when one is seeing the differences in the highs (where I can see more peaks generally) it is easy to lose what line is the db=0, which makes one have to look to the left and lose the point one was examining.

 

[udauda]

Dear Tyll, I just remembered something I wrote to you a year ago. Since then I've been waiting for any fixes, yet the ER4 data still remain unchanged. Now you've taken charge of the measurement setup, I do not think it's irrelevant to discuss this matter here. I obtained the raw measurements of ER-4S and 4P from PCMAG(they use the same HEAD-acoustics HATS system too), and here I compare them to the headroom's data:
 
ER-4S:

You can see something's definitely wrong. Along with the insertion depth issue in the high frequency, the mid-high frequency is boosted way too much for ER-4S. It's either ER-4B that was measured, or the seal was broken.
 
And 4P:

We can see the high frequency deviation again, which is resulted by a way-too deep insertion.
 
Since you've taken charge of data, I'm sure these issues will be gone in no time.

 

[Head Injury]

Quote:

Tyll,
 
I hope this doesn't take away from the purpose of your thread.  Since people are talking about the idea of a difference graph here, and you're the one doing the measuring for us, I thought it would be a good place to continue the discussion.  If you prefer a different thread, let me know.  Anyway, check out this 'difference' graph tool. 
 
http://www.headphone.com/learning-center/technical/difference-graph.php
 
Choose your reference headphone, maybe the best one you own, then choose a headphone you're curious about and see what the differences in frequency response is. 
 
This tool will need some refinements over time so I'm looking forward to everybody's feedback.

OMG awesome tool.