Latest Thread Images

crinacle's IEM FR measurement database
Apr 1, 2017 at 7:08 AM Post #181 of 1,335
Fresh new measurements in from Addicted to Audio in Kew. Some new exotics that I'm pretty certain most people haven't seen in person, much less tried.
 
As usual, raw on the left and compensated on the right:
 
Westone ES60

 
 
Westone ES50

 
 
Westone ES30

 
 
Final Audio Design LAB II

 
 
Final Audio Design FI-BA-SS

 
 
Final Audio Design Heaven VIII

 
 
Final Audio Design F7200

 
 
Hifiman RE1000

 
Notes: Ignore the sub-bass bump on the ES50 graphs (and to a certain extent, ES30) as it is an artifact caused by excessive microphonics.
 
Not a lot of subjective impressions this round. None of them really excited or impressed me for their asking prices, to say the least. The ES60 does sound very good, but it's performing at what I'd consider the bare minimum within its price bracket (which in this case is the $1500 range).
 
Other FAD products were available but I decided to go with one of each one of their product "lines" due to having prior engagements. If anyone wants any specific graphs I can get them the next time I drop by the store.
 
Apr 7, 2017 at 5:43 PM Post #182 of 1,335
I'd love to see the FAD Heaven VI (and perhaps the IV), I can't seem to see it elsewhere online.
 
Apr 11, 2017 at 5:07 AM Post #184 of 1,335
Just received my UERMs back from reshell by AAW as well as a Flamenco loaner demo from Jomo. 
 
As usual, raw on the left and compensated on the right:
 
Ultimate Ears Reference Monitors

 
Bass and Treble

 
 
Bass
 

Treble
 

None
 

Comparisons

 
 
The UERM has essentially replaced my Etymotic ER4PT as the default reference, for my uses. Flamenco impressions over at the Jomo thread
 
Apr 11, 2017 at 5:26 AM Post #185 of 1,335
crinacle said:
   
Just received my UERMs back from reshell by AAW as well as a Flamenco loaner demo from Jomo. 
 
As usual, raw on the left and compensated on the right:
 
Ultimate Ears Reference Monitors

 
Bass and Treble

 
Bass
 

Treble

 
None

 
Comparisons

 
 
The UERM has essentially replaced my Etymotic ER4PT as the default reference, for my uses. Flamenco impressions over at the Jomo thread
 
Click to expand...

 
That is very flat, wow.
Thank you for the measurements.
 
Apr 15, 2017 at 8:24 AM Post #187 of 1,335
New stuff. Raw on the left, compensated on the right:
 
AAW A2H Pro (V1)

 
AAW W100

 
AAW W300

 
AAW Nebula One

 
Oriveti Basic

 
Oriveti New Primacy

 
Sennheiser IE800

 
Shure SE535

 
Shure SE846 (Blue filters)

 
Ran into some issues with measuring Shure IEMs. Really hate those thin nozzles, can never proper seat them in my coupler.
 
Subjective impressions: 
  1. Oriveti's new lineup is good, very good. Basic has excellent bass with decent tonal balance, Primacy makes a great neutral hybrid. At their asking price points, they're easily top in my recommendations list.
  2. A2H don't sound too good to me. I mean, it's crazy value for a 220USD dual driver hybrid custom, but eh... you get what you pay for.
  3. Nebula One kinda sounds bad next to the splendid heights of the Basic. Uncontrolled bass, smoothed over details and the semi-open design aren't winning me any favours.
 
Not much I can say about the 846 and IE800 that hasn't been said before. I like bass so... I like 'em, I guess.
 
noplsestar said:
@crinacle thanks for the new measurements. Can you please show the Flamenco and Samba in one graph for comparison if it's not too much work? Thanks.
Click to expand...

 
I'll work on it soon.
 
Apr 19, 2017 at 12:45 PM Post #188 of 1,335
DISCLAIMERS

I'll have to make this part extremely huge because certain uninformed individuals like to shove words I never said into my mouth. The objectivist crowd can be snooty but based off what I've endured in the last few weeks, the subjectivist crowd can be equally as obnoxious. Unfortunately, I'm straddling in the grey area in between and thus am bombarded by both sides of the fence constantly.

1. As stated above, a Frequency Response graph is not in any way an indication of good sound quality. Think of it as flavours, like in ice-cream. The FR is like the branding at the side of the tub telling you the flavour of the ice cream. It's not going to tell you if it's good ice cream, but at least you have a way of finding out if it's the flavour you want.

2. My database can only be compared internally. Really, I'm flattered that you would attempt to compare my graphs obtained from this tiny $25 condenser mic and makeshift coupler to stuff like Tyll's god-knows-costs-how-much dedicated measuring rig, but I think at this point I'll have to put my foot down and say that if you call my graphs inaccurate based on that comparison, you're doing it all wrong. Compare with data made from the same rig; that's the rule for seasoned measurers out there. (Also I've compensated my graphs based off higher end rigs like Tyll's, so I think one will find my curves much more similar than different).

3. Measurements are meant to complement subjective impressions, not form the basis for them. I get that certain people like to paint me as some sort of Rin Choi-like figure where all I do is look at graphs and determine quality sorely off of said metrics. 99% of the IEMs I've measured, I have listened to as well. If not for the graphs, one can also take this database as a list of IEMs I've tried and auditioned, and the graphs simply being proof of it.
 
Last edited:
Apr 19, 2017 at 2:04 PM Post #190 of 1,335
Nice work @crinacle!  I tried to reproduce what you're doing (using your white-noise file), but I think my (Android) version of AudioTool is having a few problems at the low end:
 

 
Sorry for the tiny legend in the above (this is default REW output). I believe the AudioTool app for Android is not the same as AudioTool for iOS (different software authors). I'm going through a two-step procedure here with AudioTool - raw data compensated once to give the loop-back input white noise flat a completely flat FR, and then again to add a diffuse-field compensation (I'm just used to looking at DFC curves in REW). The problems I've had are: 1) I don't have a calibration curve for my mic 2) there aren't enough frequency bins or user-definable EQ taps at the low-end of AudioTools FFT to accurately adjust the low frequency response 3) I found I needed to massively hike up the low end in Android 6 to compensate for some weakness in AudioTool... then my phone updated itself to Android 7 and the low-end went up through the roof, i.e., something appeared to change in the AudioTool app under Android 7.  So, I'm a believer in the concept, but I'm struggling to get consistent results from white-noise + my phone, relative to the exact same microphone/coupler (Vibro Veritas) used with REW.
 
@crinacle, is your compensation curve just to drive your white noise via a loop-back into your smartphone to a flat frequency response, or is it something else? (I think I'm correct in assuming none of your graphs are diffuse-field compensated?) Do you have your raw data stored (as ASCII files) from all your measurements, in case you later wanted to add a different compensation curve? The Android version allows you to do this, but I didn't see that option in the iOS version.
 
Apr 19, 2017 at 3:55 PM Post #191 of 1,335
crinacle You should change the term "compensation" to "calibration". Basically your results are (visually) calibrated to match uncompensated IEC-standard measurements. Compensation is a whole nother subject. Sorry for nitpicking..

I really like how your calibration has turned out.
 
Apr 19, 2017 at 4:15 PM Post #192 of 1,335
or don't even try. I don't see the benefit of providing 2 graphs of the same measurement. I understand the desire for honesty, but if the compensation feels better to you, just use that and forget about your actual raw. I don't believe it offers any extra information.
 
Apr 19, 2017 at 11:28 PM Post #193 of 1,335
orions said:
Could you do one for the AKG N40?
Click to expand...

 
I'll keep it in mind for the next session.
 
csglinux said:
  Nice work @crinacle!  I tried to reproduce what you're doing (using your white-noise file), but I think my (Android) version of AudioTool is having a few problems at the low end:
 

 
Sorry for the tiny legend in the above (this is default REW output). I believe the AudioTool app for Android is not the same as AudioTool for iOS (different software authors). I'm going through a two-step procedure here with AudioTool - raw data compensated once to give the loop-back input white noise flat a completely flat FR, and then again to add a diffuse-field compensation (I'm just used to looking at DFC curves in REW). The problems I've had are: 1) I don't have a calibration curve for my mic 2) there aren't enough frequency bins or user-definable EQ taps at the low-end of AudioTools FFT to accurately adjust the low frequency response 3) I found I needed to massively hike up the low end in Android 6 to compensate for some weakness in AudioTool... then my phone updated itself to Android 7 and the low-end went up through the roof, i.e., something appeared to change in the AudioTool app under Android 7.  So, I'm a believer in the concept, but I'm struggling to get consistent results from white-noise + my phone, relative to the exact same microphone/coupler (Vibro Veritas) used with REW.
 
@crinacle, is your compensation curve just to drive your white noise via a loop-back into your smartphone to a flat frequency response, or is it something else? (I think I'm correct in assuming none of your graphs are diffuse-field compensated?) Do you have your raw data stored (as ASCII files) from all your measurements, in case you later wanted to add a different compensation curve? The Android version allows you to do this, but I didn't see that option in the iOS version.
Click to expand...

 
Just a word of note, refer to Tyll's raw curves (grey) rather than his compensated red/blue ones. My compensation curve is just done using a calibration file to keep them up to spec to IEC-standard measurements as @bartzky just mentioned. I have raw data stored in the AudioTools app but not in the FFT app that is used in my database.
 
castleofargh said:
  or don't even try. I don't see the benefit of providing 2 graphs of the same measurement. I understand the desire for honesty, but if the compensation feels better to you, just use that and forget about your actual raw. I don't believe it offers any extra information.
Click to expand...

 
I'm actually thinking of scrapping the compensated graphs altogether since the rest of the data are all raw (and they're all screenshots and not actual data, so I can't simply apply the calibrations to old graphs). As I've said, my efforts were to bring my graphs up to spec to commonly used IEC couplers and to make my graphs slightly more readable, but it seems more and more unnecessary the more I think about it.
 
Eh, I'll probably still do IEC-calibrated curves for my own uses anyways, and provide them to whoever requests them. That seems like a decent middle ground.
 
Apr 20, 2017 at 3:28 AM Post #195 of 1,335
that's pretty much what I do ^_^. I got used to a given compensation for myself but I usually don't post it. no matter how many warnings I will write, most just don't read or don't understand why it's a problem to put them side by side with innerfidelity. or that a flat line on the graph does not mean what they think it means.
in the end those who learn a little about graphs don't need our warnings and compensation efforts, and those who need to be warned don't care.
biggrin.gif
  you've measured a great deal of IEMs so one graph of each for direct comparison within your own database seems perfectly relevant(whichever you chose). at least as much as it can be with our cheap gears and methods.
 
You must log in or register to reply here.

Users who are viewing this thread

Total: 6 (members: 0, guests: 6)
Back
Top