Audio Measurements on a Headfi Budget

Would you buy an expensive headphone without hearing it or seeing any measurements for it?

  • No

  • Yes, I'd buy it on the spot if Steve Guttenberg says it sounds good. He is always right.


Results are only viewable after voting.
Jun 11, 2019 at 3:01 PM Post #46 of 114

csglinux

Headphoneus Supremus
Joined
Jun 4, 2014
Posts
3,603
Likes
2,760
I had some interesting communications with the folks at Earthworks recently. Their test methodology uses a free-field scheme and their M, QTC & TC series products are sold as free-field mics, but are actually all pressure mics. Some of their mics even come with a little 1/2" adapter. But of course life's never that simple :wink: In the same way that slight changes in eartip shape can make shifts in the upper frequencies, I see similar effects in the way these mics are mounted in the couplers. The best solution seems to be to seal the mic completely flush, but I think there are probably nuances here I'm yet to learn...
 
Jul 29, 2019 at 6:23 PM Post #47 of 114

csglinux

Headphoneus Supremus
Joined
Jun 4, 2014
Posts
3,603
Likes
2,760
I'm liking these Earthworks M23 mics :) They are more expensive than the other condenser mics I've been testing, but they still just about fit within the mandate of audio measurements on a head-fi budget (they're still way less expensive than GRAS or B&K rigs). The THD measurements I can get from the M23 basically match those from the GRAS RA0045 + 40AO (prepolarized mic) + 26CA (mic pre-amp) - plus the M23 takes standard phantom power, contains its own pre-amp and actually has a higher sensitivity than the 40AO+26CA combo. Here are the Xelentos, with Cp100 eartips, at 80 dB (the Sonarworks condenser mic - which is still miles better than the Taobao dynamic coupler mics - is the blue curve):

coupler_mics_thd.png
 
Oct 5, 2019 at 8:51 PM Post #48 of 114

csglinux

Headphoneus Supremus
Joined
Jun 4, 2014
Posts
3,603
Likes
2,760
Here's a quick explanation as to why my earlier measurements of the Xelento (https://www.head-fi.org/showcase/beyerdynamic-xelento-remote.22337/reviews?order=likes#reviews) didn't match those posted by @jude in the main Xelento thread (https://www.head-fi.org/threads/beyerdynamic-xelento.827372/page-78#post-14480482).

Firstly, my original measurements were diffuse-field compensated. This was done to match what InnerFidelity were doing at the time. As you can see, the agreement with @jude's measurements is poor, though some of the peaks and troughs appear to be roughly in the same location:
Xelento_DFC.png


Note the background color above. White = diffuse field compensated. I've tried to be consistent in that, when posting raw data, I always use a dark background. Here's the raw comparison (with a slightly different insertion depth):
X_cp100.png


Still not that great. The next issue was that I was using (and measuring with) SpinFit Cp100 silicone eartips. @jude had used the stock Xelento eartips. Eartips always make a significant difference. So here are my Xelentos measured with the stock Xelento eartips:
X_Xtips.png


Getting closer. The next issue was that I was using a conventional 711 coupler, whereas @jude was using the latest "hi-res" (RA0402/RA0401) coupler from GRAS. This intentionally puts a strong damping on the half-wave ear canal resonance, which, in the case of shallow-insertion IEMs, can have an influence at frequencies well below 10 kHz. Here are my Xelentos, using stock Xelento eartips, measured with the GRAS RA0402 coupler:
RA0402_Xtips.png


Still not 100% perfect agreement, but these differences could be more easily attributable to unit variance.

The morals of this story, as I see them:

1) Don't use diffuse-field compensation. @jude patiently spent an afternoon convincing me this was a bad idea, and I agree. Raw data is far easier when comparing with others, which is ultimately what we want to be able to do.
2) Always compare using the exact same eartip type and size (https://www.head-fi.org/threads/ety...and-your-couplers.908512/page-6#post-15108352).
3) Always compare with the same type of coupler. There are supposedly advantages of the RA0401/2 for manufacturers. I'm not entirely convinced that these "hi-res" couplers are ideal for end users, since ear canal resonances are a real phenomena, and (at least from my limited experience) the standard 711 couplers appear to get closer to what I hear than does the RA0402. But I'm open to persuasion if others have had the opposite experience...
 
Oct 6, 2019 at 4:16 PM Post #50 of 114

csglinux

Headphoneus Supremus
Joined
Jun 4, 2014
Posts
3,603
Likes
2,760
Have you considered doing a thread with just measurements?
Sure, I've considered it :) My ultimate goal would be to have all the members of the headfi measurement community be able to contribute to a thread like that. Technically there is already a thread on general measurements where folks can post their graphs: https://www.head-fi.org/threads/general-iem-measurements-discussions.903455/
The problem is we're quite limited as basic headfi members - we're not allowed to serve php scripts from an SQL database, and even just posting graphs is a problem because of the maximum limit of 25 images headfi imposes on any single post. I thought about creating a new thread and commandeering the next 26 slots (A->Z), but that wouldn't work either as you'd quickly fill up those slots with Sennheiser, Shure, Simgot, Shozy, Sony, Stax, etc. @crinacle had the right idea creating his own site: https://crinacle.com/graphs/graphtool/ The only problem is it's only @crinacle contributing to it. Despite the nice marketing slogan, there are plenty of IEMs that exist that aren't on the list. 99% of the time I go looking for a headphone, it's not there. But @crinacle (somewhat understandably) doesn't want to be taking measurements from sources he can't verify, or that don't follow the exact same measurement procedures. That's all laudable, but it leaves a gap with a lot of headphone measurement data going unpublished, and InnerFidelity no longer seem to be measuring anything since Tyll retired. I don't know what the best way to create a large, searchable, easily browsable database would be, but I think, for several reasons, headfi might not want to host it.
 
Oct 6, 2019 at 10:27 PM Post #51 of 114

fsi22

100+ Head-Fier
Joined
Dec 8, 2014
Posts
274
Likes
187
Location
Australia
Sure, I've considered it :) My ultimate goal would be to have all the members of the headfi measurement community be able to contribute to a thread like that. Technically there is already a thread on general measurements where folks can post their graphs: https://www.head-fi.org/threads/general-iem-measurements-discussions.903455/
The problem is we're quite limited as basic headfi members - we're not allowed to serve php scripts from an SQL database, and even just posting graphs is a problem because of the maximum limit of 25 images headfi imposes on any single post. I thought about creating a new thread and commandeering the next 26 slots (A->Z), but that wouldn't work either as you'd quickly fill up those slots with Sennheiser, Shure, Simgot, Shozy, Sony, Stax, etc. @crinacle had the right idea creating his own site: https://crinacle.com/graphs/graphtool/ The only problem is it's only @crinacle contributing to it. Despite the nice marketing slogan, there are plenty of IEMs that exist that aren't on the list. 99% of the time I go looking for a headphone, it's not there. But @crinacle (somewhat understandably) doesn't want to be taking measurements from sources he can't verify, or that don't follow the exact same measurement procedures. That's all laudable, but it leaves a gap with a lot of headphone measurement data going unpublished, and InnerFidelity no longer seem to be measuring anything since Tyll retired. I don't know what the best way to create a large, searchable, easily browsable database would be, but I think, for several reasons, headfi might not want to host it.

If you could get together and create a datum. Maybe have a few iems that can be used to calibrate test rigs, so everyone who is contributing, will then be within that tolerance level.
 
Oct 7, 2019 at 6:12 AM Post #52 of 114

castleofargh

Sound Science Forum Moderator
Joined
Jul 2, 2011
Posts
9,522
Likes
4,918
Oct 16, 2019 at 6:34 AM Post #54 of 114

TLDRonin

500+ Head-Fier
Joined
Nov 5, 2017
Posts
514
Likes
287
Location
California
If I wanted to buy the 711 coupler and dynamic mic set, what else would I need to start measuring? I have an iPhone, laptop with usb and a amp/dac
 
Oct 16, 2019 at 5:00 PM Post #55 of 114

csglinux

Headphoneus Supremus
Joined
Jun 4, 2014
Posts
3,603
Likes
2,760
If I wanted to buy the 711 coupler and dynamic mic set, what else would I need to start measuring? I have an iPhone, laptop with usb and a amp/dac
The simplest way would probably be to just use your iPhone with a dynamic/electret mic coupler. If you have a newer iPhone, you'll need that dongle. You'll also want a splitter (3.5mm trrs male to two 3.5mm trrs female), so that you can drive pink noise into the IEM from one sound-generating app and simultaneously record the averaged input from the coupler using the AudioTools app.

The AudioTools app is very good for FR. In my experience, it matches the results from REW extremely closely.
 
Oct 17, 2019 at 2:37 AM Post #56 of 114

TLDRonin

500+ Head-Fier
Joined
Nov 5, 2017
Posts
514
Likes
287
Location
California
The simplest way would probably be to just use your iPhone with a dynamic/electret mic coupler. If you have a newer iPhone, you'll need that dongle. You'll also want a splitter (3.5mm trrs male to two 3.5mm trrs female), so that you can drive pink noise into the IEM from one sound-generating app and simultaneously record the averaged input from the coupler using the AudioTools app.

The AudioTools app is very good for FR. In my experience, it matches the results from REW extremely closely.
So if I get this https://item.taobao.com/item.htm?spm=a1z09.2.0.0.289c2e8dyh5pF7&id=37135316408&_u=63rk7cgh0386
the splitter and the app I should be good to go?
 
Oct 17, 2019 at 2:37 PM Post #58 of 114

hakuzen

Headphoneus Supremus
Joined
Feb 9, 2016
Posts
1,795
Likes
2,656
Last edited:
Oct 30, 2019 at 3:13 PM Post #59 of 114

csglinux

Headphoneus Supremus
Joined
Jun 4, 2014
Posts
3,603
Likes
2,760
There was a nice video posted to the front page of headfi today. It's a little basic, and perhaps a little biased toward one particular manufacturer, but it gives a good history of ear simulators:

 
Nov 6, 2019 at 9:50 AM Post #60 of 114

WinterCharm

100+ Head-Fier
Joined
Feb 24, 2012
Posts
214
Likes
19
Location
USA
Wow.

The amount of information in your very first post is both stunning and incredible. Amazing work, and I'm really loving the level of dedication you guys have to headphone measurements. I stumbled in here as I was trying to learn more about accurately measuring headphones for a review I'm working on, of the AirPods Pro vs other well known wireless earbuds.

I'm considering going with an IEC 711 adapter + UMIK1 for measurements, as apparently the THD of these mics is far lower, and allows for proper measurement of stuff like ANC. Do any of you foresee any issues with this?
 

Users who are viewing this thread

Top