Audio Technica ATH-ADX5000 Review (Headphone)

[Jimbob54]

As a matter of fact, I just purchased those. Unfortunately, I never owned a Sennheiser HD6XX/600/650 and so I set out to to find a good deal to acquire 1 of those variants. In my quest to find the lowest price for a new set of the senns, I came across the ATH-R70X and was surprised by its specs, 470Ω, open-back,etc. and became intrigued enough that I ended up buying those instead. They may sound entirely different then the Senns. that I initially planned on getting but the higher resistance, dynamic design suited my needs since the set of cans is meant for a tube amp I was given awhile ago. Also, they are easier to drive than I expected, that's for sure. The sensitivity is 99dB/mW which is fairly high so that could account for it potentially.

Not a huge difference to the Senns - an easy listen for sure.

 

[phoenixsong]

I actually bought the R70x instead of the HD600 and HD650 after some comparisons in a non-ideal listening environment. It is not without flaws though; ended up selling it after a few months

 

[B4ICU]

Back to the testing method....
I see a pattern that kind of alerts me. The HP is an electro mechanical reproducer. The HP usually use one single full range device. it plays (on test as on our head) in a close environment. No room response, no axis angles (on axis, off axis etc') and it also has some isolating function (not much reported in Amir's tests!).
The response of such device should be continuous. I do not see why it should show such fluctuations and variations, with slight Fr. change.
Electronics, as Amp's and other pure SS devices, don't do that. Usually, once they variate, the attitude is kept. Roll off at high Fr.
or the same at very low Fr.
The test is kind of a sealed (black) box, when the HP goes on that head like dummy and than it is tested. I still would be happy to know what
HP Amp. is used, What Mic. amplifier is used and what kind of test pattern of AP is applicable.
When speakers are tested, there are some rules kept, as on axis and distance, to measure efficiency. Fr. is more complicated and need some controlled environment (noise, reflections etc'. Some of a speaker Fr. may come from crossover and multi elements structure. All that is not valid
with HP. No crossover, no crossover distortion, no elements overlap (-3dB) and many more "no"'s.

Look at 8.5kHz or even worse at about (?) 17kHz. The 17kHz is deviate by 20dB in about 1kHz (it's more like 5% Fr. deviation. Down 20dB and back up 20dB. 20dB equals to x100 amplitude.
Now let's compare the same at 500 Hz and down. Here, the response is with very smooth variations. The deviation is also no more than 7.5dB. That's about x5 only!
This chart is kind of making no sense to me. Especially that an electro mechanic device response should look more like the 500Hz rather than the 17kHz. Electro mechanical devices, are less "jumpy as Fr. rise. As the mechanic part has a mass, it is still a coil moving in an electromagnetic field
and this mechanism is practically tend to slow movement not to increase it. On this chart, the higher the Rf. goes, larger is the deviation!
This is against physics of standard mechanics or electromagnetism. Sr. Newton or Mr. Maxwell would be very unpleasant with those results.
Going back to the test, this dummy that the HP is placed on, what is it standing on? The same table with the AP test equipment / PC?
Does it vibrate? do you have a sensor (vibration) sticked to the table, that shows the output vs. Fr. from 20Hz to 20kHz?
My guts feeling are that the sudden Fr. change with amplitudes of up to x100, within very little change in Fr. is not the HP property.
If so, the entire HP testing should be revised. Something I remember from Physics lab, is that if the test results do not fit your expectations, you need to check your test methods before publishing the results.

 

[Maiky76]

Back to the testing method....
I see a pattern that kind of alerts me. The HP is an electro mechanical reproducer. The HP usually use one single full range device. it plays (on test as on our head) in a close environment. No room response, no axis angles (on axis, off axis etc') and it also has some isolating function (not much reported in Amir's tests!).
The response of such device should be continuous. I do not see why it should show such fluctuations and variations, with slight Fr. change.
Electronics, as Amp's and other pure SS devices, don't do that. Usually, once they variate, the attitude is kept. Roll off at high Fr.
or the same at very low Fr.
The test is kind of a sealed (black) box, when the HP goes on that head like dummy and than it is tested. I still would be happy to know what
HP Amp. is used, What Mic. amplifier is used and what kind of test pattern of AP is applicable.
When speakers are tested, there are some rules kept, as on axis and distance, to measure efficiency. Fr. is more complicated and need some controlled environment (noise, reflections etc'. Some of a speaker Fr. may come from crossover and multi elements structure. All that is not valid
with HP. No crossover, no crossover distortion, no elements overlap (-3dB) and many more "no"'s.

View attachment 134993

Look at 8.5kHz or even worse at about (?) 17kHz. The 17kHz is deviate by 20dB in about 1kHz (it's more like 5% Fr. deviation. Down 20dB and back up 20dB. 20dB equals to x100 amplitude.
Now let's compare the same at 500 Hz and down. Here, the response is with very smooth variations. The deviation is also no more than 7.5dB. That's about x5 only!
This chart is kind of making no sense to me. Especially that an electro mechanic device response should look more like the 500Hz rather than the 17kHz. Electro mechanical devices, are less "jumpy as Fr. rise. As the mechanic part has a mass, it is still a coil moving in an electromagnetic field
and this mechanism is practically tend to slow movement not to increase it. On this chart, the higher the Rf. goes, larger is the deviation!
This is against physics of standard mechanics or electromagnetism. Sr. Newton or Mr. Maxwell would be very unpleasant with those results.
Going back to the test, this dummy that the HP is placed on, what is it standing on? The same table with the AP test equipment / PC?
Does it vibrate? do you have a sensor (vibration) sticked to the table, that shows the output vs. Fr. from 20Hz to 20kHz?
My guts feeling are that the sudden Fr. change with amplitudes of up to x100, within very little change in Fr. is not the HP property.
If so, the entire HP testing should be revised. Something I remember from Physics lab, is that if the test results do not fit your expectations, you need to check your test methods before publishing the results.

You should post there:

https://www.audiosciencereview.com/...nt-thread-about-headphone-measurements.18451/
Also have a look at the characteristics of a Mylar diaphragm as an example and try to check what a comb effect is when you have reflexions.

 

[respice finem]

...No room response, no axis angles...

I wish it was that simple...

There is no room response in terms of the room one sits in, but there is always some response of the HP earcups, probably one of the reasons for "jagged" treble graphs.

Axis angles also do exist, even if you would wear your headphones with their acoustic center ideally aiming at your outer ear canal, there is no such thing as a straight and perpendicular ear canal (I've seen thousands of them).

Measuring headphones remains an Art of its own, and results will not be exact, Amir says that in his note (see review).

I'm afraid no setup will change it much, with the exception of IEMs, where the measurement might (one day) be performed right in front of the eardrum. And even then, the mic and its wires will change the volume and response in the small closed chamber between the IEM and the eardrum...

Nothing is ever perfect.

 

[B4ICU]

I wish it was that simple...

There is no room response in terms of the room one sits in, but there is always some response of the HP earcups, probably one of the reasons for "jagged" treble graphs.

Axis angles also do exist, even if you would wear your headphones with their acoustic center ideally aiming at your outer ear canal, there is no such thing as a straight and perpendicular ear canal (I've seen thousands of them).

Measuring headphones remains an Art of its own, and results will not be exact, Amir says that in his note (see review).

I'm afraid no setup will change it much, with the exception of IEMs, where the measurement might (one day) be performed right in front of the eardrum. And even then, the mic and its wires will change the volume and response in the small closed chamber between the IEM and the eardrum...

Nothing is ever perfect.

You could be right, if...
The deviation (I pointed) are of a 20dB (x100) at 17kHz! That a lot.
The only way it can come in, from the room response, is from the surface that the dummy head is placed on.
I would be happy if that would be a very massive marble qube, with a separate stand on a solid floor (tiles over concrete).
Having that dummy sitting on the table with the PC / AP unit and more, with a large surface standing on a hardwood floor is a big NO.
This HP model is closed ear and should have some isolation. To allow a 20 dB deviation is not likely. Also the 20 dB is a drop not a peak.
I am familiar with testing. I designed ATE's, integrated ATE's, wrote TRD's and more. Even had lunch with AP CEO in the 90's.
The results seems to show a problem, not HP related and of a significant magnitude.
I only say that there is a good reason to revise the test method and check from where those anomalies are coming from.

 

[amirm]

The test is kind of a sealed (black) box, when the HP goes on that head like dummy and than it is tested. I still would be happy to know what
HP Amp. is used, What Mic. amplifier is used and what kind of test pattern of AP is applicable.

The mic buffer is GRAS 12AX. External amplification is not needed due to sensitivity of the Audio Precision input. The headphone is driven by Topping A90 headphone amplifier. The test pattern is a log chirp designed specifically to measure speakers and headphones.

 

[solderdude]

The results seems to show a problem, not HP related and of a significant magnitude.
I only say that there is a good reason to revise the test method and check from where those anomalies are coming from.

When/if Amir would decide to use extreme smoothing and divisions like 10dB instead of 2.5dB the plots would look very different.
It would be better if one would be selling headphones.

There is nothing weird/wrong with the measurements. Just overlay them with other peoples measurements using similar/same industry standard test fixtures and apply similar smoothing and they would look the same.
The near 10kHz 'anomaly' is caused by the pinna and is seen in all similar test fixtures. It's a something to keep in mind.


B.t.w. the measurements would not look any different when Amir would use any other decent power, low distortion, capable amplifier b.t.w.

 

[amirm]

Look at 8.5kHz or even worse at about (?) 17kHz. The 17kHz is deviate by 20dB in about 1kHz (it's more like 5% Fr. deviation. Down 20dB and back up 20dB. 20dB equals to x100 amplitude.
Now let's compare the same at 500 Hz and down. Here, the response is with very smooth variations. The deviation is also no more than 7.5dB. That's about x5 only!

You need to reach for that high-school physics lessons to understand this, namely the simple formula for computing wavelength. At 17 kHz, the wavelength of sound is 0.8 inches or 2 centimeters. This is easily within the distance of the driver and the artificial ear/surround "cheek" of the dummy head. As such, you get "modal" response that causes dips and peaks just like it happens in low frequencies with speakers in rooms. The distance can be such that you can even get full cancellation although in reality that doesn't quite happen due to losses and such.

You can also get reflections from cup edges of the headphone, internal and complex ones form the pinna, etc.

These are the obvious ones. The less obvious ones require expertise in acoustics and beyond high-school physics. Namely these are not ideal drivers and generate resonances both internally and from structure of the headphone itself. These "little headphones" generate their own sound waves that mix with the main sound and create more variations (due to vector sum).

There are also severe non-linearities where the headphone no longer acts pistonic and hence generates harmonics which can change the frequency response as I showed in this instance:

The smooth and "pretty" graphs you see in my preference curve and measurements from others is due to heavy filtering and smoothing. I have applied some smoothing but I don't want to go too far as it then masks faults in the driver/headphone that would get lost otherwise. Here is an example of that in speakers: https://www.audiosciencereview.com/...ds/perceptual-effects-of-room-reflections.13/

This is the high-resolution in-room measurements allowing reflections just like headphone measurements above:

You see massive dips far worse than headphone measurement due to many reflection paths. We can filter all of that:

Now all the dips are gone but so has the detail we may need to judge the speaker hence the reason simple filtering is not a replacement for anechoic measurements of the same speaker:

Bottom line, you are not understanding the nature of sound acoustics and measurements. Please ask questions when you don't understand something rather than running off with a bunch of assertions about the measurements being in error.

 

[amirm]

Even had lunch with AP CEO in the 90's.

I had coffee with Robert Redford at Sundance Festival where he and the show organizer wanted us to sponsor the festival. While it was a pleasure to meet in him person (a super nice guy), it didn't make me an actor!

 

[respice finem]

I know it probably won't be possible, but, if only for demonstration purposes, it would be interesting to see the treble response, without the metal grille on the back of the earcup. Any reflexions causing high interference must IMHO be from the driver's direct vicinity, because otherwise they would be of too low amplitude.

 

[solderdude]

I know it probably won't be possible, but, if only for demonstration purposes, it would be interesting to see the treble response, without the metal grille on the back of the earcup. Any reflexions causing high interference must IMHO be from the driver's direct vicinity, because otherwise they would be of too low amplitude.

Below the HD58X with complete rear (screen + holder) removed and with screen. That screen is a lot less 'transparent' then the AT one.

Can you guess/see which measurement is with and without a screen ? Take into consideration that while mounting the screen back on one inevitably moves the headphone a bit.

 

[respice finem]

Proves how difficult it is. Seemed plausible to me because of the vicinity, but "as we can see, we can't see a thing"...
This might point to "sins" of the driver itself, but this is "technically too high" for me, I didn't even have lunch with a physics teacher

 

[cranfan95]

I briefly owned these, they sounded like crap for the used price I paid let alone $2,000

big fail from AT. big ups to pimp daddy Amir for the spot-on review.

 

[amirm]

I know it probably won't be possible, but, if only for demonstration purposes, it would be interesting to see the treble response, without the metal grille on the back of the earcup. Any reflexions causing high interference must IMHO be from the driver's direct vicinity, because otherwise they would be of too low amplitude.

Not possible. I returned the unit to its owner already.

 

[B4ICU]

You need to reach for that high-school physics lessons to understand this, namely the simple formula for computing wavelength. At 17 kHz, the wavelength of sound is 0.8 inches or 2 centimeters. This is easily within the distance of the driver and the artificial ear/surround "cheek" of the dummy head. As such, you get "modal" response that causes dips and peaks just like it happens in low frequencies with speakers in rooms. The distance can be such that you can even get full cancellation although in reality that doesn't quite happen due to losses and such.

You can also get reflections from cup edges of the headphone, internal and complex ones form the pinna, etc.

These are the obvious ones. The less obvious ones require expertise in acoustics and beyond high-school physics. Namely these are not ideal drivers and generate resonances both internally and from structure of the headphone itself. These "little headphones" generate their own sound waves that mix with the main sound and create more variations (due to vector sum).

There are also severe non-linearities where the headphone no longer acts pistonic and hence generates harmonics which can change the frequency response as I showed in this instance:

The smooth and "pretty" graphs you see in my preference curve and measurements from others is due to heavy filtering and smoothing. I have applied some smoothing but I don't want to go too far as it then masks faults in the driver/headphone that would get lost otherwise. Here is an example of that in speakers: https://www.audiosciencereview.com/...ds/perceptual-effects-of-room-reflections.13/

This is the high-resolution in-room measurements allowing reflections just like headphone measurements above:

You see massive dips far worse than headphone measurement due to many reflection paths. We can filter all of that:

Now all the dips are gone but so has the detail we may need to judge the speaker hence the reason simple filtering is not a replacement for anechoic measurements of the same speaker:

Bottom line, you are not understanding the nature of sound acoustics and measurements. Please ask questions when you don't understand something rather than running off with a bunch of assertions about the measurements being in error.

Sorry Amir.
Filtering input graph is a well practiced technique. It shall first be clear if that is the actual thing, or it is just a display problem. On that very same graph, the low Rf. are snoth. No sudden or extreme over or under shoots. So it is not about filtering. it is about data this graph provides.
I still looking for an answer about the placement of that dummy and what kind of vibrations it may be susceptible to?
For reflections, I said nothing. It is what it is on head or on test dummy. Irrelevant to this discussion. I did suspected that maybe some of that
Fr. graph, is holding data of external vibrations. What about that?
I assume, that if I would write a remark as you did with Robert Redford, you would block me as you did on the Marantz AV receiver, over the fan noise discussion. My comment about AP CEO, Mr. Bob Metzler (passed away), was to share an event, when he visited our facility (Elbit Systems)
as we were about to equip with AP gear for some VOA stations we win to build it's audio and control subsystems, including self test.
In other words, AP is no strange to me. Mr. Robert Redford and your comment are.
I also pointed out, that measuring equipment, need to be tested and calibrated periodically. Please be kind to answer this question.

 

[amirm]

I still looking for an answer about the placement of that dummy and what kind of vibrations it may be susceptible to?

The measurement fixture is extremely heavy. Probably 20 to 25 pounds. I also have it on an isolation pad. Not that any vibration would disturb the frequency response as measurements are loud enough to dwarf any such thing. But it helps with distortion measurements.

 

[Rockfella]

@amirm .. waiting for M40x review hehe. What you said comes out true almost all the time. Speaker business is 80% marketing. I think that applies to every industry.

 

[amirm]

In other words, AP is no strange to me.

The Audio Precision plays no significant role in headphone frequency response measurements. I use it because I have it and prefer its interface to other software. Otherwise you could use REW or any of myriad of other software and get exactly the same results. Your insistence on showing familiarities with it indicates no understanding of the actual process involved here.

 

[amirm]

For reflections, I said nothing. It is what it is on head or on test dummy. Irrelevant to this discussion.

It is? It is the thing that explains the dips you were worried about. Not some magical weakness in measurements. Take a $5 mic, put it in a flat plate and measure some headphones and then come back to complain when you know something about this process. Your constant protests on these threads is becoming so annoying. You are in serious danger of being shown the door. The review threads are not for empty complaints.