HiFiMAN HE1000 Unveiled - Measurements & Official Discussion

[Resolve]

This is the place to discuss all things to do with the HiFiMAN HE1000 Unveiled:

Hifiman HE1000 Unveiled Headphones - Headphones.com

I’ve posted my impressions of it in a different thread already so I’ll add them here, along with the measurements.

B&K 5128

Raw

GRAS & HpTF Variation

Folks know that I hate the ‘unveiled’ concept, but on this one the implementation is a lot less anxiety-inducing. There’s actually a lot less surface area of the membrane exposed to the world compared to the Susvara Unveiled, meaning it’s less likely to get punctured by various things one might have lying around.

As far as the sound quality is concerned, this one is one of the more balanced egg-shaped HiFiMAN headphones out there. It has noticeably less treble compared to the Arya platform, and even the HE1000 platform. So the ear gain is kept in line a bit better with the rest of the response. Also importantly the typical HiFiMAN midrange dip around 1.5khz is a lot less here than on their other models.

I liked this headphone a lot, especially with binaural recordings from Allison Krauss that they had there (this is a wtf experience if I ever had one, my goodness). If I had one critique it’s that for me there might be a bit of excess 4khz, causing a kind of ‘clenched’ presentation to certain tones - but it’s really a nitpick since the rest of it was quite balanced.

If you can handle the ‘unveiled’ concept, I could see this being someone’s endgame headphone - and they also improved the more modal response of this series (the jaggies in the midrange are lessened), so it’s a great platform for some EQ as well.

EQ

You’ll note some extra treble peaks up there still, I could’ve gone a bit further here but this ended up sounding decent enough to me and I don’t want to use my HRTF to prescribe adjustments for other people. So you’d need to do that by ear. Remember that headphone behavior, especially in the upper frequencies, will shift from head to head.

The information below is far less impactful for sound quality, do not try to read the tea leaves from these metrics

Harmonic Distortion

95dB

100dB

105dB

This result is better than many moving coil headphones but not as good as what you find with certain other planar magnetics. Note that these large ovoid planars from HiFiMAN typically don’t have the best harmonic distortion performance.

It’s still below audibility at normal volumes and you won’t have any trouble making adjustments with EQ. The only thing to note here is that when you crank it, you do see some 3rd and 4th order products showing up, but it’s still below 0.5%.

Driver Fs:

This shows driver behavior when seal is broken. So no worries if you have glasses or don’t get perfect coupling.

[dixter]

@resolve I didn’t want to put your graphs and comments here as those are yours to post here…
my questions while you still have them in house…
how do the HE1000U compare to the Susvara U in sound sig… ?
can you post your graphs of the HE1000U and compare to the graphs of the SusU and the OG Sus?
if you have the OG Sus how does the new HE1000U sound sig compare…
and if you have had the time, do you have recomendations for EQ on the HE1000U or SusU
thanks

[Resolve]

Hey. It’s a tough call when it comes to sound quality. I can actually see an argument for HEK U over Sus U, if you’re the kind of person who wants more upper treble ‘air’ and sparkle, because that’s where HEK U seems to have more presence. They both have a bit of excess energy in the lower treble to my ear, just in different spots. So it kind of just depends on what you’re more sensitive to.

Overall I prefer the Sus OG. I mainly just prefer the HEK U over the other HEK series I’ve heard. Comparing it to the Susvara platform is hard to do because there are other reasons why I wouldn’t personally consider the Sus U.

[Nate]

How much L/R imbalance is concerning over the frequency spectrum? The curves show imbalances in some areas quite large relative to the preferance boundaries …50%+ in some areas and thats even after averaging. Mabe some frequencies reveal imbalances moreso than other frequencies, but I imagine it matters.

[dixter]

@Resolve … just curious as the graphs from the Canjam post and the ones here might indicate an issue with the L/R balance… looks like one of the drivers is having an issue in the higher frequency…

[Resolve]

No, this is actually totally reasonable channel matching. There’s a lot of misunderstanding when it comes to what’s perceptually relevant to do with this stuff, as well as the absolutely insane sensitivity to placement in the upper frequencies. Let me assure you, this is some of the better channel matching you’ll see for over-ear headphones, and any small differences would never be identified in a blind test.

[Nate]

Would you quantify what imbalance would be idetifiable?
This chart here shows an example better than the last post. Its your raw measurements plotted against the preferance boundaries. Looks like the new hifiman has wwway more imbalances than I’ve seen plotted for common HD600/650 for example.

[Nate]

Just trying to learn if 5-10+ extra decibls on one side can be identified and if not then what amout of imbalance would you say can be dectected? Given you’ve listened to tons of headphones and seen their frequency plots. Whats your take?

[dixter]

@resolve It just seems a little strange how much difference the two graphs look… one graph looks like the L/R are tracking well around the 12K area… and then another graph of the same headphone shows the L/R diverging for some reason… kinda makes you wonder which one is correct…

[Resolve]

This is largely just to do with how it was normalized. This is a good example of what we’ve been saying about the sine illusion and the problem of raw plots. This is the same data, just normalized slightly differently. Though I will note, I have L/R colors reversed there for some reason… may need to regenerate that.

Also keep in mind, these are two different rigs. The one showing better matching is the flat plate GRAS rig, while the B&K data is the full head, with contours to the side of the face.

As for what’s perceptually relevant, the only thing a person might hear is the excess around 4khz on one side for this unit. But that’s still highly unlikely. Folks need to stop reading the tea leaves in the treble.

[Resolve]

So, certainly this isn’t noticeable. It’s not perfect, but no headphone has perfect channel balance, and it’s also far worse in practice for the vast majority of headphones - so the FR at the ear drum in situ. When you look at plots, things are often visually more striking than they are perceptually relevant, and it’ll actually depend more on how wide the deviation is.

Here, updated with correct colors:

[Nate]

Can you comment on what level of channel imbalance you have precieved yourself as a concern?
I’m trying to ask an open question to learn. Reading the tea leaves is just how the open question came to mind. Im an degreed engineer that works with data every day. Its a fair question, but i sense you might not want to quantify your experience due to so many more question coming as a result of your answer?

[Resolve]

Well, in general people have a tendency to overfocus on things that really don’t matter. We haven’t even gotten into the massive variation in situ you get across different heads, and the delta with their HRTF due to headphones behavior changing with different ears/heads.

But as far as how channel imbalances propagate, with music, 4dB with a high Q (narrow) would be hard to identify. 4dB with a a low Q (wide) would be much more easy to identify. In this case, it’s unlikely to be audible. With test tones you could hear it for sure, but that’s a different scenario.

[dixter]

curious if you have a way to measure the L/R balance of your measuring device to see how well it tracks L to R compare… maybe the device has issues with high freq and that gets into the graph of a headphone being measured…

[Nate]

Interesting insight. Thanks

[Spearthrower_Owl]

Can anyone do measurements comparing HEKU to HEKv4 (Stealth) and HEKSE? Thank you!

[Spearthrower_Owl]

I just found this. If valid, looks as if the Unveiled has a bit more around ~1.5-2k and less at ~6k. Interesting . . . but it maintains soundstage? If so, very nice!