Headphone CSD waterfall plots
Aug 22, 2011 at 1:12 AM Post #31 of 937
Purrin, very nice!! The sr80 mod is interesting. I need to check these mod threads...

From back of the hand calculation the 2kHz peak is likely the first 1/4 wavelength resonance of the undamped back chamber. Some foam in there, esp. At the grill location should help a lot. 

On the other hand, I don't see how the rs models can sound better given the same basic design unless the driver is somehow tuned to alleviate such acoustic res. Or maybe the chamber is quite different dimensions, but still these undamped modes and bound to occur above 2k, regardless of the shape...
 
Aug 22, 2011 at 1:52 AM Post #32 of 937
The RS1 cups are a different dimension. The RS2 cups are the same or similar. A few years ago, I noted a similar peak ~2kHz in my RS2 FR measurement (the RS1 didn't have this peak.) You can hear it too as kind of a "honk" in the midrange. It's possible the driver may contribute too.
 
At this point, I'm just happy how the foam is really helping with damping down the ringing on the HD800.
 
Aug 28, 2011 at 10:29 AM Post #34 of 937
Also, for those who got some data and would like to use Excel to check these CSDs, I made a spreadsheet with a macro to calculate the ffts (you need to enable macro and load the analysis toolpack).
 
Not as clean as scilab but probably more convenient:

 

 


 
PS: many many thanks again to Purrin for letting me play with his data!
 
Aug 28, 2011 at 6:55 PM Post #38 of 937
Thanks all, credit all goes to Purrin for taking such clean data and Tyll Hertsen for pointing me to Excel :wink:.
 
I will process a few other cans that Purrin sent me the data for. Also, I have corrected the SPL scale, the Excel graphs above are not properly normalized...
 
cheers,
arnaud
 
Aug 31, 2011 at 4:46 AM Post #39 of 937
I re-did the SET1 and SET2 measurements. The new plots use the same data, but presented more correctly. The original plots were on based on acoustical pressure in pascals instead of being converted into decibels. The new plots are interpolated between the data points and have tick marks on the axes.
 
http://www.head-fi.org/t/566929/headphone-csd-waterfall-plots/15#post_7698438
http://www.head-fi.org/t/566929/headphone-csd-waterfall-plots/15#post_7699881
 
Here is a "perfect" (well not actually perfect, but an electrical loopback of the impulse response) to give us an idea how far off the ideal these headphone drivers are.
 

 
I'm be posting another set of a few more other headphones soon.
 
Aug 31, 2011 at 11:04 AM Post #40 of 937
Marv, your custom made graphs look very professional, esp. the detail of the settings in the upper left corner! Somehow, I got a nagging feeling you got a bit inspired by Clio, didn't you? :wink:.
 
I am not sure it's appropriate for me to post the same data processed in Excel as it feels like a contest, but I was requested by someone earlier... Only added the HP1000 as I felt compelling to compare it to the HF2. Is that just me or the HF2 is a naughty in the 2D view, as if it got irritated that I show it in that light :wink:
 



 



 
Aug 31, 2011 at 11:49 PM Post #41 of 937
 
Quote:
Marv, your custom made graphs look very professional, esp. the detail of the settings in the upper left corner! Somehow, I got a nagging feeling you got a bit inspired by Clio, didn't you? :wink:.
 


Yup :) But honestly, I like that information there so I can look back and know what I did. I also need to keep up with you since you keep pulling rabbits out of your hat - like doing CSDs with Excel.
 
 
Sep 1, 2011 at 1:38 AM Post #42 of 937

Set 3

Fostex T50RP and Sony MDR-SA5000
 

Fostex T50RP Left and Right

  
 

Sony SA5000

 
 
 
Sep 1, 2011 at 1:48 AM Post #43 of 937

Set 4

Six different HP1000 drivers on same pad.
KEL and KER have the best response. ISL and ISR are badly matched on the same headphone.
 
 
 

 
Sep 1, 2011 at 10:41 AM Post #44 of 937


Quote:
Originally Posted by purrin /img/forum/go_quote.gif
I also need to keep up with you since you keep pulling rabbits out of your hat - like doing CSDs with Excel.


hahaha, you win purrin I can't keep up with the volume of data you're generating :wink:. Glad the issues with Clio got circumvented with your home-brew!
 
 
Sep 2, 2011 at 2:43 AM Post #45 of 937

Set 5:

Koss ESP950 (electrostats), Sennheiser HD558, Sony XB500 XB700

 

Sony XB500 XB700

Always good to test something you know will suck to validate these measurements.
 

 

Sennheiser HD558 with the 598 mod

Not bad.

 

Koss ESP950

Smooth FR. Fast decay from 2.5kHz upward with spurious ringing. I really like how these headphones sound.
 
 
 
 

Users who are viewing this thread

Back
Top