Dan Clark Stealth Review (State of the Art Headphone)

[Yorkshire Mouth]

Amir, do you think you compromised your measurements by listening first? Would it be even better to not know which headphones you're measuring, to eliminate any bais? Thanks

It's difficult, though not impossible for bias to compromise measurements. More likely the other way around.

If the item measures well, you're certainly likely to be influenced into thinking it sounds good. But if it sounds good, that can't affect the readings, which are being done by a machine.

Now, yes it's possible that, if you know you have a superb-sounding set of headphones, you're possibly going to be more careful to ensure the headphones are placed carefully and correctly on the instrumentation, but if Amir's 'normal' is to be a bit of a perfectionist with this anyway (which I suspect he is), then that'd be difficult to improve on, no matter how much more careful you were.

By the way, Amir would be the wrong person to ask. Bias really doesn't work like that. Amir can't eliminate his own bias (as none of us can) by thinking about it, or being extra careful. Amir's skill and expertise could not possibly influence his own bias. Amir is a skilled and experienced electronic engineer in the field of audio. If you want to find out about bias, ask an expert in bias.

That would be the scientific thing to do.

Absolutely no offence intended to either you, Amir, or anyone else.

 

[solderdude]

For someone struggling with this material, is there a simple test one can do to determine if their equipment can push enough current? For example, I ordered the Stealth intending to use it with my E1DA 9038S Gen2. If the volume stays the same from 60-100 (arbitrary), is that current limit being hit at 60?

i use headphones exclusively with my phone, no exceptions. I'm fine w buying a portable amp but really prefer dongles.

The E1DA 9038S (don't know about gen 2 but @IVX will know) can supply 3.3V in 23 ohm = 109dB peak SPL. Should be loud enough for somewhat loud listening levels. Needs a balanced cable though.

 

[VMAT4]

So... What is Dan Clark about as a company? Is it a boutique business or large corporation?

It may be still possible to buy stock in the company. See StartEngine.com

 

[Doodski]

It may be still possible to buy stock in the company. See StartEngine.com

Interesting web page. https://www.startengine.com/danclark

 

[edahl]

Interesting web page. https://www.startengine.com/danclark

It says "offering closed" to me!

 

[Lord Victor]

But can you feel the bass hitting your ears?
this headphones hit hard as no other can do?

By measurements this can be one of the best headphones

I share that concern - it has been one of the major shortcomings of past DCA models. Hoping the changes here have solved that.

some have suggested this less physical, more blunted sound is more correct though…

 

[Hifihedgehog]

Price points aside...

 

[preload]

If Amir ever gets to test the Aeon 2 Noire, I would be very curious about a blind test with the Stealth to see if he could notice a difference between them, seeing as they measure fairly similarly. I think a blind test of he listens to headphones AA, BB, and AB in a random order and tries to guess which one is AB. For the pair he guesses to be AB, he writes his impressions and guesses which headphone is which.

I would guess he wouldn't be able to distinguish which pair is AB from sound alone,

My prediction would be that the two headphones would be easily and readily distinguishable. Also the question isn't really that interesting (in my opinion). We already know that small differences in FR are identifiable. And even if they measured identically on the gras45, the 2-dimensional FR plot doesn't explain the whole sound picture. Try equalizing a pair of HD650s to match the curve of the HD800S. They won't even sound remotely similar.

 

[edahl]

I share that concern - it has been one of the major shortcomings of past DCA models. Hoping the changes here have solved that.

some have suggested this less physical, more blunted sound is more correct though…

Someone mentioned they thought og it as misdirected sound energy, makes sense to me

 

[musicarcade]

It may be still possible to buy stock in the company. See StartEngine.com

It states the offering ended March 26th this year.

 

[Robbo99999]

If Amir ever gets to test the Aeon 2 Noire, I would be very curious about a blind test with the Stealth to see if he could notice a difference between them, seeing as they measure fairly similarly. I think a blind test of he listens to headphones AA, BB, and AB in a random order and tries to guess which one is AB. For the pair he guesses to be AB, he writes his impressions and guesses which headphone is which.

I would guess he wouldn't be able to distinguish which pair is AB from sound alone, but this is just a guess I have little confidence in. The stealth is 415 grams and the Aeon 2 Noire is 328, for this reason a blind test may not be possible, unless both headphones feel very similar and the tester deliberately does not pay attention to weight. There are other issues such as volume matching, getting the appropriate dacs and amps, creating a scenario close to a double blind test, and of course guessing randomly will correctly identify the headphones ~17% of the time.

This would be an annoying test to set up for a somewhat meaningless result, but I do think it would be interesting nonetheless.

My prediction would be that the two headphones would be easily and readily distinguishable. Also the question isn't really that interesting (in my opinion). We already know that small differences in FR are identifiable. And even if they measured identically on the gras45, the 2-dimensional FR plot doesn't explain the whole sound picture. Try equalizing a pair of HD650s to match the curve of the HD800S. They won't even sound remotely similar.

The main difficulties in trying to EQ one headphone to the frequency response of another, or indeed when EQ'ing 2 headphones to the the same target curve and comparing them is that you are relying on the accuracy/repeatability of the measurement process used (ie. the tester & their measurement methodology) and also there is the element of unit to unit variation (including pad wear) that cannot be accounted for (assuming you haven't measured those specific units on the GRAS). Also frequency response above 10kHz is unreliable in the measurements, so you can't match the 2 headphones above 10kHz anyway. So those 3 variables conspire to mean that there are always some differences between 2 headphones that have been EQ'd to the same target curve. There is one other variable as well - differently designed headphones will react to the GRAS differently than they will with your own head, the Headphone Transfer Function is different, so even if you could equalise and account for the earlier 3 mentioned variables you'd still be left with different Headphone Transfer Functions which would result in some differences in frequency response received at your own eardrum between 2 different designs of headphones that had been thoeretically perfectly EQ'd to match each other on the GRAS unit. So that's around 4 variables in total that conspire together to create audible differences between headphones that have been EQ'd to the same curve on a GRAS unit.

The fact that the overall cup design shape looks quite similar between the the Aeon 2 Noire and the Stealth then it's possible that if they're EQ'd to the same curve that they may well indeed sound more similar to each other than 2 headphones that have large design differences between the cups (because the the headphone transfer function of the Aeon 2 Noire & the Stealth are likely to be more similar if the physical cup design is similar), but still you're left with the other variables of difference I mentioned in the previous paragraph. There's quite a lot of variables associated with headphone listening! (Also not mentioned so far as an additional variable: HRTF differences between individuals which creates differences in the subjective listening experience between individuals, theoretically we all have our own ideal Target Curve - which is not a factor in speaker listening. Having said all of that I am a fan of headphone listening & the Harman Curve, so I don't wish to paint a negative picture, just outlining the variables associated with headphone listening).

 

[GaryH]

That's... Literally what I'm trying to do, or at least aspiring to do. Seemingly that gets missed in all of this.

If you want to fully investigate 'macrodynamics / macrocontrast / slam', I reckon the answer will (at least partially) come from measuring the acoustic impedance curve of headphones, as I mentioned in this thread.

 

[Phos]

I generally like seeing impulse response tests because good headphones tend to look one way and less good headphones tend to look not that way. Plus, it's nice for seeing if they've reversed the phase for some reason. This one looks remarkably like an open headphone, which seems to be the point.

 

[ifloatoveryou]

My prediction would be that the two headphones would be easily and readily distinguishable. Also the question isn't really that interesting (in my opinion). We already know that small differences in FR are identifiable. And even if they measured identically on the gras45, the 2-dimensional FR plot doesn't explain the whole sound picture. Try equalizing a pair of HD650s to match the curve of the HD800S. They won't even sound remotely similar.

What exactly is not explained by frequency response? If two headphones has the exact same frequency response as measured in a listener's ear, the listener would hear the exact same thing, including sound stage. Of course, the frequency response as measured in a gras45 is different than what would be measured in your ears, which would be different than my ears. This is mostly because adjusting the headphone slightly will affect treble noticeably and an imperfect seal from glasses, a beard or even a narrow job will affect bass noticeably.

The noir and stealth differ mostly in these two areas where frequency response changes from user to user, which is why this hyptoehtical experiment would be interesting to me. Of course, if it is easily discernable which pair is being worn due to how they feel on the head, then this experiment could not be performed.

 

[Kyle / MrHeeHo]

What exactly is not explained by frequency response? If two headphones has the exact same frequency response as measured in a listener's ear, the listener would hear the exact same thing, including sound stage. Of course, the frequency response as measured in a gras45 is different than what would be measured in your ears, which would be different than my ears. This is mostly because adjusting the headphone slightly will affect treble noticeably and an imperfect seal from glasses, a beard or even a narrow job will affect bass noticeably.

The noir and stealth differ mostly in these two areas where frequency response changes from user to user, which is why this hyptoehtical experiment would be interesting to me. Of course, if it is easily discernable which pair is being worn due to how they feel on the head, then this experiment could not be performed.

they do not have the same frequency response though, it takes only a quick look at measurements from other sites to confirm this

 

[preload]

What exactly is not explained by frequency response? If two headphones has the exact same frequency response as measured in a listener's ear, the listener would hear the exact same thing, including sound stage.

Not necessarily. A simple frequency response plot does not contain any information about the angle of incidence of the sound (hint: neither a headphone transducer nor the cup chamber can be considered a point source). The plot also does not contain information about distortion. And it contains no information about the time domain, particularly resonances.

 

[ifloatoveryou]

Not necessarily. A simple frequency response plot does not contain any information about the angle of incidence of the sound (hint: neither a headphone transducer nor the cup chamber can be considered a point source). The plot also does not contain information about distortion. And it contains no information about the time domain, particularly resonances.

If the angle or position of the headphones change, the frequency response will change as well - any effect the angle of incidence will have on sound will be shown in the frequency response. The same goes for resonances - any effect the stealth's meta material resonator has on sound is reflected in the frequency response.

You are right about distortion. I should have qualified my statement - if two headphones have no audible distortion and have the exact same frequency response as measured in a listener's ear, the listener would hear the exact same thing.

I am not exactly sure what you meant by explained by frequency response. You would be correct if you mean that the frequency response will not tell you a headphone's angle of incidence or the specific resonances present. The soundstage, of an HD800S or any other headphone, is a function of frequency response. If you were able to EQ the HD650s to be the measure exactly the same as the HD800S in your ears, they would sound the same. Just applying oratory1990's EQ, they will sound different.

 

[preload]

If the angle or position of the headphones change, the frequency response will change as well - any effect the angle of incidence will have on sound will be shown in the frequency response.

I'm not referring to the position of the headphones. I'm referring to the fact that sound waves strike the outer ear at a multitude of varying angles of attack, which are a function of the transducer radiating sound as a surface and at various angles and reflections within the cup. A simple 2-d FR plot doesn't characterize any of this.

The same goes for resonances - any effect the stealth's meta material resonator has on sound is reflected in the frequency response.

Again a 2-d FR plot does not have a time axis. Ringing from a resonance would persist for a duration of time. A simple 2-d FR plot doesn't characterize any of this.

If you were able to EQ the HD650s to be the measure exactly the same as the HD800S in your ears, they would sound the same.

I love how you have to make such an inane assertion just to make the rest of your theory work.

Consider this:
1) do you really think ppl haven't tried to eq their hd650 to match the hd800?? I have for fun. It's not even close. Not even a little bit.
2) if you think you have the eq curve that will make the hd650 match the hd800 identically, by all means post it and I'd be happy to eat my words, and I'm being sincere
3) Harman's own research, whereby a pair of AKG701 headphones are eq'd to mostly match the FR of various other headphones, and where they subject listeners to blind listening tests, found that the most preferred curve was that of a Sony MDR-something. I purchased that headphone. Amir reviewed it. And others have bought it as well. The unanimous impression is that these headphones are "just okay."
4) if it were THAT easy to copy headphones, why wouldnt headphone manufacturers take the cheapest low distortion headphone possible and simply eq it to match the most preferred $$$ model? I mean why not eq your hd650 into an Orpheus?

 

[ifloatoveryou]

Again a 2-d FR plot does not have a time axis. Ringing from a resonance would persist for a duration of time. A simple 2-d FR plot doesn't characterize any of this.

Resonances do show up in the frequency response. The frequency response is a fourier transformation of the impulse response. If you wanted time to be an axis of the graph, you would perform an inverse fourier transformation of the frequency response.

I love how you have to make such an inane assertion just to make the rest of your theory work.

Consider this:
1) do you really think ppl haven't tried to eq their hd650 to match the hd800?? I have for fun. It's not even close. Not even a little bit.
2) if you think you have the eq curve that will make the hd650 match the hd800 identically, by all means post it and I'd be happy to eat my words, and I'm being sincere
3) Harman's own research, whereby a pair of AKG701 headphones are eq'd to mostly match the FR of various other headphones, and where they subject listeners to blind listening tests, found that the most preferred curve was that of a Sony MDR-something. I purchased that headphone. Amir reviewed it. And others have bought it as well. The unanimous impression is that these headphones are "just okay."
4) if it were THAT easy to copy headphones, why wouldnt headphone manufacturers take the cheapest low distortion headphone possible and simply eq it to match the most preferred $$$ model? I mean why not eq your hd650 into an Orpheus?

You don't need to call my assertions inane and be dismissive. I am only repeating what the science states, none of this is a matter of opinion. I can link you to explanations made by oratory1990 if that would be helpful.

I did not say it was easy or possible to EQ the HD800S to sound like the HD650, or vice versa. In fact it is not possible to do so just by looking at measurements made on a gras45. I said if you could eq two different headphones to have the exact same frequency response in your ear, they would sound the same. I chose the HD650 and HD800S since you mentioned them, but it really does not matter which two headphones I choose, as long as distortion is not audible and the frequency response is the same.

I can't tell you precisely the reason people don't eq and spend $$$ on their preferred headphone model. I would image a lot of it is because they do not know about the science behind headphones. People also value comfort, build quality and aesthetics. Cost has no correlation with sound quality.

 

[Robbo99999]

Resonances do show up in the frequency response. The frequency response is a fourier transformation of the impulse response. If you wanted time to be an axis of the graph, you would perform an inverse fourier transformation of the frequency response.



You don't need to call my assertions inane and be dismissive. I am only repeating what the science states, none of this is a matter of opinion. I can link you to explanations made by oratory1990 if that would be helpful.

I did not say it was easy or possible to EQ the HD800S to sound like the HD650, or vice versa. In fact it is not possible to do so just by looking at measurements made on a gras45. I said if you could eq two different headphones to have the exact same frequency response in your ear, they would sound the same. I chose the HD650 and HD800S since you mentioned them, but it really does not matter which two headphones I choose, as long as distortion is not audible and the frequency response is the same.

I can't tell you precisely the reason people don't eq and spend $$$ on their preferred headphone model. I would image a lot of it is because they do not know about the science behind headphones. People also value comfort, build quality and aesthetics. Cost has no correlation with sound quality.

If you could EQ an HD650 and HD800s to match the same frequency response at your own eardrum (not on a GRAS) then indeed as you say they would sound the same (bar any differences in distortion measurements).