[Otto Motor]

Here my measurements of the NiceHCK DT600. I really like the low end. As to the upper end, I feel it is really treble forward. Is that the 5-7 kHz peak or my imagination?

 

[Otto Motor]

And here a universal reversible mod: taping the nozzle 80-90% over with micropore tape removes nasty peaks approximately between 5 and 15 kHz (varies between earphones). Borrowed form the trick inventory of @james444.

However, even after the mod, this single DD still doesn't sound great: harsh because of a low-quality driver (I speculate).

 

[csglinux]

This isn't a criticism or a request for @crinacle to re-do his measurement database for a third time - just a point of discussion. I noticed something today in the first post of this thread - that we're trying, where possible, to adjust insertion depth in order to try and move any resonance peaks to 8 kHz. Is that fair to small, ergonomically-shaped IEMs that are designed for - and would allow - deeper insertion than that of large, clunky 36 BA-drivers-per-side monstrosities?

I get the ambiguity over insertion depth, but if an IEM can be inserted deep enough to cause a peak above 10 kHz (for example), but we artificially adjust it in the coupler to make that peak appear at 8 kHz, that's a whole different headphone and a very different listening experience. I don't know what the answer is with insertion depth when dealing with differently shaped ears and different IEM form factors, but as things stand, a reader who missed the fine print (as I did) might get an overly-negative impression of an otherwise good IEM. @crinacle - I liked your last sentence of that paragraph: "it's arbitrary and it's up to me." I think that's all the disclaimer you need. I personally wouldn't go chasing an 8 kHz resonance.

Thoughts? Criticisms? Death threats??


P.S. I don't know if anybody ever replied to this question:

I'm looking to put together a rig like Crin's. Where can you purchase IEC60318-4 compliant inner ear simulator for a good price? Or where can you purchase the same ear simulator like Crin's?

I wrote a small post that describes a number of affordable couplers, including the (very good) coupler that @crinacle uses. Follow the first spoiler link in the post below:

https://www.head-fi.org/threads/audio-measurements-on-a-headfi-budget.893084/

 

[csglinux]

There were some interesting comments about headphone unit variance in the latest (January 22nd) edition of the InnerFidelity podcast that featured Helmuts Bēms of Sonarworks. (Sonarworks have measured thousands of headphones.) Bēms makes two rather depressing claims: 1) they find the typical variance from unit to unit is around 4 dB and 2) that units sent out to reviewers typically sound - and measure - better than the bulk stock that's sold to the average unsuspecting victim customer. It's worth a listen. (Also, check out the cringe-worthy part at the end where it becomes obvious Bēms doesn't actually know what a harmonic is.)

I'd be curious to know if any of you agree with Bēms and Hunter that distortion is something you actually want from your headphones and speakers.

 

[castleofargh]

didn't listen(I have set that website free soon after Tyll left), but headphone variations are no secret. as far as I know only a minority of manufacturers bother to guaranty channel matching on a given pair(with whatever tolerance over whatever frequency range). I expect many others to have internal standards to avoid garbage results, but as I don't know them, I can't say much about that. it's sort of expected to see some variations between channels, and on average I'd expect even more variations between pairs(unless the manufacturer really checks nothing at all and just randomly picks the next driver in line).
the service Sonarworks proposes to measure our own headphone and create a curve for them, wouldn't make much sense if variations didn't exist on the headphone in the first place. with that said, we're talking variations over a given frequency range that needs to be properly defined(along with other measurement settings maybe) to compare apples to apples. if the headphones are super well matched from 100hz to 10khz but have 8dB variation at 19khz, that shouldn't really matter(plus how reliable would be a measurement at that frequency?).

I have no opinion on distortions, if anything that's not music or noise is distortion, I'd need some specifics before saying I welcome them with open arms ^_^. but in general, if I'm happy with the FR, I'm happy, so I don't object to distos as much as a true elite audiophile would, up to a point.

 

[crinacle]

This isn't a criticism or a request for @crinacle to re-do his measurement database for a third time - just a point of discussion. I noticed something today in the first post of this thread - that we're trying, where possible, to adjust insertion depth in order to try and move any resonance peaks to 8 kHz. Is that fair to small, ergonomically-shaped IEMs that are designed for - and would allow - deeper insertion than that of large, clunky 36 BA-drivers-per-side monstrosities?

I get the ambiguity over insertion depth, but if an IEM can be inserted deep enough to cause a peak above 10 kHz (for example), but we artificially adjust it in the coupler to make that peak appear at 8 kHz, that's a whole different headphone and a very different listening experience. I don't know what the answer is with insertion depth when dealing with differently shaped ears and different IEM form factors, but as things stand, a reader who missed the fine print (as I did) might get an overly-negative impression of an otherwise good IEM. @crinacle - I liked your last sentence of that paragraph: "it's arbitrary and it's up to me." I think that's all the disclaimer you need. I personally wouldn't go chasing an 8 kHz resonance.

My current measurements are performed without adjustment of the coupler but I aim for 8k resonance whenever possible. You can see that a lot of FitEar custom demos have very early resonances and that's due to their stubby nozzles and ridiculously shallow fit. So far it's more or less arbitrary but not so arbitrary to the point where consistency is thrown completely out the window.

There were some interesting comments about headphone unit variance in the latest (January 22nd) edition of the InnerFidelity podcast that featured Helmuts Bēms of Sonarworks. (Sonarworks have measured thousands of headphones.) Bēms makes two rather depressing claims: 1) they find the typical variance from unit to unit is around 4 dB and 2) that units sent out to reviewers typically sound - and measure - better than the bulk stock that's sold to the average unsuspecting victim customer. It's worth a listen. (Also, check out the cringe-worthy part at the end where it becomes obvious Bēms doesn't actually know what a harmonic is.)

I'd be curious to know if any of you agree with Bēms and Hunter that distortion is something you actually want from your headphones and speakers.

I've been saying that for a while but consumers prefer to put their trust in manufacturers even against the presence of solid data, for whatever reason. It's a little disconcerting to be honest.

The distortion observation is also something that I've been investigating semi-recently. Distortion can be pleasant or at the very least, benign. That's the reason why I've been moving away from THD, CSD and impulse breakdowns; I've been doing those for a while and I struggle to find a correlation between those and any subjective metric of "audio quality". Particularly in the significance of THD; the "superiority" of 0.01% versus say, 0.5% is not consistent at all with my own subjective observations.

 

[ThomasHK]

There were some interesting comments about headphone unit variance in the latest (January 22nd) edition of the InnerFidelity podcast that featured Helmuts Bēms of Sonarworks. (Sonarworks have measured thousands of headphones.) Bēms makes two rather depressing claims: 1) they find the typical variance from unit to unit is around 4 dB and 2) that units sent out to reviewers typically sound - and measure - better than the bulk stock that's sold to the average unsuspecting victim customer. It's worth a listen. (Also, check out the cringe-worthy part at the end where it becomes obvious Bēms doesn't actually know what a harmonic is.)

I'd be curious to know if any of you agree with Bēms and Hunter that distortion is something you actually want from your headphones and speakers.

As some one who's done thousands of heapdhones measurements (most likely) and worked on many commercially available headphones for different brands, here's some info:

• Only few measurement setups can give you consistent measurements for end of line testing. I'm assuming the jig Sonar works uses is not one of this. Basically anything with "ears" and uneven surfaces for the headphones to sit on are out. So no dummy heads. For earphones it's basically what you guys are used to as being a 711, but with some additional jig to have consistent mounting every time
• All brands worth their salt will do frequency response and balance testing on each unit.
• Frequency response is compared to a Golden Sample, basically a statistical average of the final achieved production performance.
• Typical tolerances for FR are +/- 3dB, 100Hz to x kHz, depending on how smooth the HF is (sharp peaks and dips are very hard to measure consistently so typically wider bounds are used as you go up in frequency)
• Some headphones are specced with tighter tolerances in the mid range, especially ANC models (e.g.+/- 2dB), but that always comes at a cost, either because it requires component sorting or discarding of more sub assemblies
• L/R balance, for brands that care is typically measured over the same bandwidth as the FR and set to be <3dB

 

[james444]

I'd be curious to know if any of you agree with Bēms and Hunter that distortion is something you actually want from your headphones and speakers.

Distortion is commonly used in music production, for pretty much the same reason they mention in the podcast:

I like to create what I call “false harmonics.” (Meaning: they're real harmonics, but they weren't there in the original recording.) More simply, I use distortion to create the impression of a full capture by adding harmonic energy.

https://www.waves.com/tips-to-make-vocals-sound-huge

So no, I wouldn't want my headphones or speakers to add extra distortion on top of that. If I were curious how it sounds, I'd use software like V4A to try it out, but not hardware on which it can't be undone.

 

[csglinux]

So no, I wouldn't want my headphones or speakers to add extra distortion on top of that. If I were curious how it sounds, I'd use software like V4A to try it out, but not hardware on which it can't be undone.

My thoughts exactly. It sounds like a sketchy sales pitch for wonky hardware: "Ok, it measures like crap, but trust us, it sounds great! Not only that, but if the recording is created with the same distortions, and then you add more of those distortions via DSP, it will sound exponentially better!" (Not just three-times better, because, obviously, the combined effect is non-linear )

 

[Otto Motor]

I am close to releasing my own blog...which will feature all my measurements. One item I will add is a correlation of graph and sound perception and its limitations...which will help people in graph interpretation.

 

[james444]

One item I will add is a correlation of graph and sound perception and its limitations...which will help people in graph interpretation.

I can see that "Killed after membranes was pierced with a needle" will help people in graph interpretation.

 

[Otto Motor]

I can see that "Killed after membranes was pierced with a needle" will help people in graph interpretation.

What's wrong with visualizing disaster? It will help people getting rid of old junk while discovering their sado-masochistic side. Hey, your air-vent mod works extremely well for the Sennheiser Momentum in-ear...reducing the bass brings out the recessed midrange without a downside. The driver is obviously good. You could try this on our ie800.

The air-vent mod alone does only work soso for the NAD HP20 and B&W C5 S2 bass monsters -- these now need a bit of upper midrange/lower treble taming with micropore...or transpore or paper-transpore (no surprise, see below).

I am presently compiling an overview of reversible modding (with source citations and credits) for my forthcoming blog: it cannot do any harm having all principles compact in one place -- links allow the interested reader to get the nitty gritty.

 

[Brooko]

First thing I'd do for your rig is get a calibration curve organised. See if you can find someone with an IEC711 compliant rig, both measure a series of IEM's, and use his correct measurements to calibrate yours. There is no way the stuff your measuring has that magnitude of dips and peaks - unless they are truly awful.

+25db sub-bass over the mid-range, and a drop of 25db from 5 kHz to 10 kHz ........ extreme yuck. That can't be real!

Here's a post from Tinyman - another DIY rig - https://www.head-fi.org/threads/nad-viso-hp20-in-ear.685854/page-2#post-10227176

Bass matches yours - but lower treble (while it has same shape) doesn't have same magnitude. FWIW my Veritas also had the typical drop in this region - which is why I had to correct it with a calibration file.

 

[Otto Motor]

First thing I'd do for your rig is get a calibration curve organised. See if you can find someone with an IEC711 compliant rig, both measure a series of IEM's, and use his correct measurements to calibrate yours. There is no way the stuff your measuring has that magnitude of dips and peaks - unless they are truly awful.

+25db sub-bass over the mid-range, and a drop of 25db from 5 kHz to 10 kHz ........ extreme yuck. That can't be real!

Here's a post from Tinyman - another DIY rig - https://www.head-fi.org/threads/nad-viso-hp20-in-ear.685854/page-2#post-10227176

Bass matches yours - but lower treble (while it has same shape) doesn't have same magnitude. FWIW my Veritas also had the typical drop in this region - which is why I had to correct it with a calibration file.

Thanks, will look into it. I have the same setup as biodegraded and our curves match pretty much (see below). And when an iem has a neutral signature, my measurements look neutral.
The NAD and B&Ws are truly bass canons.

 

[Brooko]

Nice - biggest concern is the lower treble drop. Just suggesting that if it was me I’d be getting it checked. What else have you measured? If you can match with something Crin or I have measured, it’ll tell you if there is something funky with your set-up