[crinacle]

  Is there any possibility to create and use calibration data in your software/app? Peaks seem to be correctly positioned, but there's a treble roll-off that could be compensated by the software, IMO.
I ventured to take your Westone W20 measurement and do some photoshop.
 

 
Yellow is your data while dark yellow/red is my data (calibrated vibro veritas) and the grey is taken from innerfidelity.
 
Resources:
- http://headflux.de/messungen/
- http://www.innerfidelity.com/headphone-measurements

 
I'm still in the process of analysing the compensation required for my IMM-06 (my measurements are very different from @nmatheis' after all) but your graph is very very useful, thanks!
 
Though I think the W20 comparison is a little exaggerated in the upper treble regions (5k-10k), but if you compare with my W60 readings (mine vs Tyll's) it seems more in line with my current observations. Here's what I noticed:
 

1. +/-5dB insensitivity from 1500Hz upwards (possibly earlier based off your comparisons)
2. Additional insensitivity from 5000Hz upwards (roughly 2-3 extra)

 
I should be able to make use of Spectrum Analyzer's calibration tools to create my own custom compensation, but unfortunately I'm still rather inexperienced in editing calibration files. If anyone can aid me please do, but here's the current compensation that I'm thinking about:
 

1. Gradual increase in sensitivity from 650Hz onward, peaking at +5dB at 1500Hz
2. Further increase in sensitivity from 5000Hz, peaking at a +7dB at 5500Hz

 
This should, in theory, match my curves point-on-point with Tyll's. As for re-doing the measurements... man, that sounds like a horrible headache...

 

[bartzky]

I guess Spectrum Analyzer needs something like a file that specifies frequency and gain like REW and Arta do?

I'd start building a calibration by setting a fixed point where your readings meet Tyll's grey ones. Something like 600Hz 0dB. Then I'd try to correct the whole downward slope by increasing a very high frequency like 20kHz. This should tilt the whole graph upwards starting at 600Hz.

Maybe you don't have to redo the measurements by simply applying the new calibration to the old measurements. At least REW can do this.

Is there any possibility to use sine sweeps instead of noise in your app?

 

[crinacle]

Been working on a compensation curve for the past day or so. I'm making progress, but still don't have anything truly solid as of now. I'm getting close, so that's rather exciting...
 
Etymotic ER4PT for reference
 

Spoiler: Previous compensation attempts

First attempt
 

Slight tweaking
 

Adjustments to the upper-treble compensations
 

Failed attempt to smooth

 
Here's what I think may serve as the basis for all future tweaking:

 
And as a fun attempt, my AAW A3H custom (which lines up with the manufacturer's curve rather well now)

 
 
Do let me know if I should do anything extra, or if I should continue etc.. Feedback is most welcome.
 
EDIT: due to the nature of my previous measurements (screenshots, raw data was NOT saved unfortunately) I am unable to simply apply any new compensations to them. A re-measure is definitely necessary. sigh...

 

[bartzky]

Great progress!

3 kHz should be a little higher while 7-8 and and the peak at 10 kHz are slightly to high.

I'd recommend to use different IEMs for calibration at some point. Every individual IEM differs and so do the measurements.

And can you change scale of the y-axis? A smaller range and 5 dB steps would be a great improvement for readability.

 

[bflat]

Nice work! Sticky worthy.

 

[EagleWings]

Also some additional low-end:
 
+2.5 db at 20Hz
+2 db at 50Hz
+1 db at 100 Hz

 

[twister6]

Great effort @crinacle.  Just curious, since you are still working on calibration of your setup (believe me, I know, went through the same with Veritas until I got help, thanks to @Brooko), will you go back and remeasure everything again to correct FR curves?

 

[1TrickPony]

nmatheis- need a big favor friend, do you think you could post your measurements on the titan 5 and 3?

 

[ohotonge]

WOW, good information, thanks

 

[crinacle]

Great progress!

3 kHz should be a little higher while 7-8 and and the peak at 10 kHz are slightly to high.

I'd recommend to use different IEMs for calibration at some point. Every individual IEM differs and so do the measurements.

And can you change scale of the y-axis? A smaller range and 5 dB steps would be a great improvement for readability.

Unfortunately the app that I'm using now does not have the option to modify axis values. Think I mentioned it earlier in the thread.

The compensation is still a work in progress of course, currently only have the ER4 as a proper reference. When my Harmony 8.2 comes back from Piotr I can probably compare with his IEC standard rig.

Also some additional low-end:

+2.5 db at 20Hz
+2 db at 50Hz
+1 db at 100 Hz

I don't think the bass sensitivities are of any concern based off my comparisons with speakerphone's and Tyll's measurements, but thanks for the suggestion. In fact I think I may need to decrease sensitivity in that area...

Great effort @crinacle
.  Just curious, since you are still working on calibration of your setup (believe me, I know, went through the same with Veritas until I got help, thanks to @Brooko
), will you go back and remeasure everything again to correct FR curves?

Current plans suggest no. I'll be starting classes soon so there won't be a lot of opportunities for big updates like before, probably a few scattered here and there occasionally. But who knows, maybe I'll revisit them one day but as of now I'll have to determine a compensation curve.

nmatheis- need a big favor friend, do you think you could post your measurements on the titan 5 and 3?

I'll chime in here and say that due to the design of the Titans (and on that note, earbuds) it would be impossible to get an accurate measurement using just the tube coupler due to how these earphones use concha acoustics to operate properly, so unless there's a dummy ear used it would not be representative of the FR we perceive.

 

[EagleWings]

I don't think the bass sensitivities are of any concern based off my comparisons with speakerphone's and Tyll's measurements, but thanks for the suggestion. In fact I think I may need to decrease sensitivity in that area...

 
Got it. I just understood you are asking for suggestions to compensate the raw measurements on your rig. I assumed you were asking for recommendations to build a negative compensation curve to show compensated measurements of IEMs. 

 

[nmatheis]

nmatheis- need a big favor friend, do you think you could post your measurements on the titan 5 and 3?

I can, and I understand crinacle's reservations regarding measuring the semi-open Titans, but they're an IEM that can change based upon each individual's outer ear anatomy. As such, I think a comparative measurement is fine as long as everyone understands the caveat that the overall FR curve might not be representative of what you're hearing. You will however be able to see the differences between the two.

 

[nmatheis]

crinacle: Can you shoot me a PM with your compensations? I'd like to learn see how my measurements look with that curve applied...

 

[1TrickPony]

I can, and I understand crinacle's reservations regarding measuring the semi-open Titans, but they're an IEM that can change based upon each individual's outer ear anatomy. As such, I think a comparative measurement is fine as long as everyone understands the caveat that the overall FR curve might not be representative of what you're hearing. You will however be able to see the differences between the two.

Thanks crinacle nmatheis
Appreciate the intention and quality of your work. I initially wanted those graphed provided my familiarity as reference, and Nik and myself heard them the same way like many other seasoned head fi members even though every now and then I'd have another member tell me a different story. Fair game I say. This is a very useful tread indeed! Keep up the great work -- exciting times here indeed!

 

[ezekiel77]

Front page congrats! One of the most useful threads here!