Apple AirPods Max Review (Noise Cancelling Headphone)

[PeteL]

the frequency response of the headphone changes as the seal changes, Dynamic EQ is the strongest feature this headphone has.

So basically if you're comparing responses Above 1KHz THEN the seal makes a difference because the Dynamic EQ does not function above 1KHz.

So basically just match the graphs at 2KHz or something.

Thank you, but please take the time to look closely at the graph, You can match them at any frequency they will not stay matched for the rest, we have at leat 3 dB difference in the Upper mids regardless of where you match them of course if you match at 2K it will be closer, but the difference everywhere else will be significant!

 

[abdo123]

Thank you, but please take the time to look closely at the graph, You can match them at any frequency they will not stay matched for the rest, we have at leat 3 dB difference in the Upper mids regardless of where you match them of cours if you match at 2K it will be closer, but the difference everywhere else will be significant!

Yes but if you match the graphs above 1KHz you basically get the instrumentation error and unit variation. below that it's a gamble.

but luckily Dynamic EQ is tuned to Harman sub-bass so whatever correction you personally apply can be limited to the region above 1KHz and trust that Apple knows best.

 

[mononoaware]

Well measuring headphones below 800-1000Hz on one's head is the easiest thing to do since it uses the same principle as ANC HPs's feedback mic, and I'd say it works pretty darn well... if it weren't for the APM's peculiar headband to cup attachment design ruining things in extreme edge cases for me (when I very significantly tilt / rotate my head left / right). Basically the curves always land in the exact same place every single time I measure it at lower frequencies. ANC headphones with a feedback mechanism are by far the most consistent closed backs I've ever tried in that regard, APM included.
But since the APM's pivot has the potential to be sub-optimal for some people's morphology it might not be quite as excellent for everyone. That said that's less an issues related to Adaptive EQ per se and more to the APM's design.

While you may perceive it as consistent in real world use (on your head) I also suspect “measurement rigs” have the potential to throw off the algorithm being used, for example it does not account for a “perfect” seal.

 

[MayaTlab]

Basically, we are left with this because we can't rely on what mesurments say, despite a Whopping 4 dB difference in the hi mids, It's inside the measurments methods tolerance....

When I see so much variance I just reduce the "interpretational reach" I'm ready to indulge in. In all cases so far, the APM shows a pretty darn smooth response for a closed back below 4-5kHz, exceedingly low THD, some degree of deficiency in the ear canal gain region (which magnitude and exact shape is difficult to determine), and a lot of variation above 4-5kHz that makes the current roster of measurements not superbly illustrative. That's where I stop and then I listen .

 

[PeteL]

Yes but if you match the graphs above 1KHz you basically get the instrumentation error and unit variation. below that it's a gamble.

but luckily Dynamic EQ is tuned to Harman sub-bass so whatever correction you personally apply can be limited to the region above 1KHz and trust that Apple knows best.

Maybe, it does feel a bit arbitrary. It's my understanding that treble measurements are not as reliable, and Amir never matched to target at 1k or above. I believe we should have a standard, but I believe he wants to be agile to adjust that to every review to visually get more compliance (or more target matches and general tendencies) It's OK too. I get it, but my point remain, those two measurements show different frequency responses.

 

[staticV3]

A collection of Airpods Max measurements on GRAS systems:

• Dr. Sean Olive (45CA): link
• oratory1990 (45BC): https://headphonedatabase.com/oratory/headphones?ids=177
• AudioScienceReview (45CA): https://www.audiosciencereview.com/...-max-review-noise-cancelling-headphone.25609/
• Crinacle (43AG-7): https://crinacle.com/2020/12/19/apple-airpods-max-review-the-audiophiles-perspective/
• Resolve (43AG-7): https://forum.headphones.com/t/apple-airpods-max-new-over-ear-headphones-from-apple/9536/59
• 0dB (45CA-10): https://www.0db.co.kr/index.php?mid=REVIEW_0DB&category=182&document_srl=1886323

Also included are all measurements as csv.

 

[MayaTlab]

While you may perceive it as consistent in real world use (on your head) I also suspect “measurement rigs” have the potential to throw off the algorithm being used, for example it does not account for a “perfect” seal.

I don't just perceive it, I measure it .
I do agree that the measurement process may throw off the algorithm. The wiggles Crinacle measured below 100hz for example may be a by-product of the rig / methodology.
But I'm not sure that it's easy to determine exactly why / how.
Crinacle didn't get these wiggles when he measured the APM using RTA and pink noise (cf. his review).
Apple mentions that they played pink noise to "initiate" the system before measuring the APM's THD on a 5128 for promotional purposes :

https://www.apple.com/newsroom/2020...gic-of-airpods-in-a-stunning-over-ear-design/
But is that really needed to avoid throwing off the "system" ? As far as I'm concerned regardless of whether I'm playing pink noise or not before running sweeps I get the same on-head response.
Resolve mentioned instead that he got a varying response with different amount of clamp at low frequencies :
https://forum.headphones.com/t/apple-airpods-max-new-over-ear-headphones-from-apple/9536/59
And it seems that Amir struggled with this a bit as well.

In the ear canal gain region personally I'd rather bet on the APM's headband to cup attachment + spring loaded cups (the combination of which applies the clamping force around the ear in a different way from a traditional yoke design) + peculiar earcup design not playing nice with the typical test rig (or some human heads, perhaps - it's entirely possible that some people may experience the APM's ear canal gain region as it shows on test fixtures after all !), which geometry around the pinna isn't a particularly good match for the average human head for most of them. But this is just a hunch, and I'm likely wrong.

Above that, no idea. It's a mess.

Jude at Head-fi wrote a series of posts regarding his difficulties in feeling confident about his measurements of the APM :
https://www.head-fi.org/threads/airpods-max.949152/post-16058062
https://www.head-fi.org/threads/airpods-max.949152/post-16071445
https://www.head-fi.org/threads/airpods-max.949152/post-16091082

With headphones such as the HD 650, HD560S, Sundara, etc. in general the available measurements gave me the "gist" of what I was about to experience listening to them, but the APM (and more recently the Hi-X65) are headphones for which that didn't quite work for me.
One of the reasons I'd be quite interested in learning more about headphones transfer functions using real-ear measurements.

Anyway, for various reasons I find them really hard to love, but I do find them absolutely fascinating in so many ways for better or worse.

 

[Tks]

But nobody would experience them that way.

I would? Its how I use my Airpods Pro anyway.

 

[Robbo99999]

The response in Crinacle's graphs is a little more elevated at, for example, 100 and 1000Hz, than in Amir's. If you were to normalise there the difference at 3kHz I think would diminish a bit. Crinacle's rig is quite different from Amir's it seems other than the plate / coupler (and pinna ?) :
View attachment 146292
https://crinacle.com/2020/12/19/apple-airpods-max-review-the-audiophiles-perspective/

In regards to the APM in particular I think this is a form of over-interpretation. I'd take measurements of the APM with a pinch of salt in terms of the exact magnitude of the ear canal gain response and the way the FR is shaped above 5kHz. Oratory uses the 5001 pinna (is that the same as Crinacle's ?) but on a HATS and gets pretty different results in the trebles, particularly in terms of magnitude.

Crinacle now uses the same, in terms of compatibility with the Harman Target Curve, equipment as Oratory & Amir. Yes, he's only measuring one headphone cup at a time, but he flips the headphone around and slots in the other ear to measure the other channel. I think he made the rig himself but instead making sure he was buying in all the critical GRAS parts that are compatible with the Harman Target Curve, including the pinna (ear). So I think we can say Crinacle's measurements are the same as Amir's & Oratory's apart from the measurement protocol and how they decide to place the headphone on the rig and how they average their results are different to arrive at a final curve, but the equipment is all compatible.

EDIT: @crinacle , I better mention you here in case I got something wrong, since I don't want to speak of your equipment wrongly.

 

[eddantes]

Seems to me that while diffrences in measurements in the bass region are a cause for concern (fixtures, clamping force, pads, etc, etc) the differences to the right of that green line are what is glaring to me. There are parts of these charts that are 6db apart...

 

[abdo123]

View attachment 146318

Seems to me that while diffrences in measurements in the bass region are a cause for concern (fixtures, clamping force, pads, etc, etc) the differences to the right of that green line are what is glaring to me. There are parts of these charts that are 6db apart...

Wow. I wonder how other 'popular' headphones do in comparision, something hand-tuned like high-end Sennheiser.

 

[Robbo99999]

@PeteL and @MayaTlab , the differences you're debating between Amir's result and anybody else's is gonna be down to what @GaryH said here:
https://www.audiosciencereview.com/...-noise-cancelling-headphone.25609/post-873389
One of the limitations with Amir's testing protocol is that he doesn't do many many measurements and create an average, so his published measurement in terms of positional variation could be sitting almost anywhere on that spectrum of variation that GaryH is showing in Oratory's graph. That's one of the problems with comparing Amir's measurements with anyone else's. Most other people do a number of measurements and average them which removes the element of positional variation, so it's probably a bit easier to compare results of people of who do averages rather than those who publish a single measurement. There's also the unit to unit variation variable to explain some differences as well.

 

[MayaTlab]

how they decide to place the headphone on the rig

That's the one bit where Crinacle's rig differs a bit from hammerhead style fixtures. Unless I'm mistaken in the ear canal gain region (1-4kHz or so) HPs still operate under "pressure conditions" (I don't know the right terminology, ie basically SPL is the same at any point in space inside the front volume) - that's what that article suggests for example : https://www.aes.org/e-lib/browse.cfm?elib=20605
So anything that affects the front volume's, well, volume, may affect the response, and perhaps in the case of the APM disproportionally so in the ear canal gain region. And the APM's headband to cup attachment + spring loaded earcup may add quite a bit of difficulty to the equation. Perhaps the ways pads compress as well : on a rig with a flat plate around the pinna it's probably quite a lot more evenly compressed than on a real human's head (particularly since the APM's pivot + spring cannot by design ensure an even pressure around the ear for most people).
Above 4-5kHz no idea what's going on.
Amir's results don't seem to diverge any more from the average of the APM's results on a GRAS rig than the other individual results. It's a generalised crapshoot past 1kHz or so.

There's also the unit to unit variation variable to explain some differences as well.

As I said I've measured four APMs, that's eight earcups in total (since they're nearly perfectly front to back symmetrical, unlike other headphones, it's somewhat valid to reverse them and measure them on the same ear), and the response didn't vary that much (if taking into account the typical seatings to seatings variation I see in the ear canal region on my head with the APM there may have been no more than around 1dB of sample variation in that range, with one cup being an outlier and the others even more tightly grouped). Cf graph for four earcups earlier.

 

[Robbo99999]

That's the one bit where Crinacle's rig differs a bit from hammerhead style fixtures. Unless I'm mistaken in the ear canal gain region (1-4kHz or so) HPs still operate under "pressure conditions" (I don't know the right terminology, ie basically SPL is the same at any point in space inside the front volume) - that's what that article suggests for example : https://www.aes.org/e-lib/browse.cfm?elib=20605
So anything that affects the front volume's, well, volume, may affect the response, and perhaps in the case of the APM disproportionally so in the ear canal gain region. And the APM's headband to cup attachment + spring loaded earcup may add quite a bit of difficulty to the equation. Perhaps the ways pads compress as well : on a rig with a flat plate around the pinna it's probably quite a lot more evenly compressed than on a real human's head (particularly since the APM's pivot + spring cannot by design ensure an even pressure around the ear for most people).
Above 4-5kHz no idea what's going on.
Amir's results don't seem to diverge any more from the average of the APM's results on a GRAS rig than the other individual results. It's a generalised crapshoot past 1kHz or so.



As I said I've measured four APMs, that's eight earcups in total (since they're nearly perfectly front to back symmetrical, unlike other headphones, it's somewhat valid to reverse them and measure them on the same ear), and the response didn't vary that much (if taking into account the typical seatings to seatings variation I see in the ear canal region on my head with the APM there may have been no more than around 1dB of sample variation in that range, with one cup being an outlier and the others even more tightly grouped). Cf graph for four earcups earlier.

I can't speak for the specific detail you mention re front volume, etc, as I don't have a deep understanding of that, but I do know that different measurement protocols between two people using the same equipment will result in two different results, so that's the variable I mentioned & that I pointed out, which was one of @GaryH 's points......so your point you bring up would be in addition to that if I could vouch for what you're saying.

Ok, you had quite low unit to unit variation with this headphone then.

EDIT: you talk about validity of reversing a headphone and measuring the other earcup.......I don't know if you're referencing Crinacle's methodology here or not.....if you are then you should be aware that he inserts the other ear into the fixture before flipping the headphone, so he's correctly measuring both left & right channels with the correct symmetry (seems you might have missed that point)...so he's never measuring the headphones on backwards, lol.

 

[MayaTlab]

different measurement protocols

I do feel indeed that protocols may have a role to play, but maybe not protocols alone to that extent. If that were the case we'd see this sort of gross variation in measured response for all headphones I think.
I feel that there's something else at play here.

Ok, you had quite low unit to unit variation with this headphone then.

All I'm saying is that... I'm very skeptical that sample variation can explain such large discrepancies between industry standard rigs, at least to the degree that @eddantes highlighted. It's really that, a question of degree, which in the case of the APM is staggering. To the point of being worrying if that's what real humans experience (but maybe not...).

EDIT: you talk about validity of reversing a headphone and measuring the other earcup.......I don't know if you're referencing Crinacle's methodology here or not.

No it's just that as my head is asymmetrical I usually can't measure L/R channel imbalance. But since with the APM I can reverse them and they remain relative to one ear or the other acoustically nearly identical, it's quite valid to do it this way .

BTW Apple entertained for a while the idea to make them left / right reversible and it's likely that the mostly front to back symmetrical design may be an offshoot of that original goal :
https://appft1.uspto.gov/netacgi/np...85".PGNR.&OS=DN/20200280785&RS=DN/20200280785

The final earcup shape was already there at least as far as 2018 :
https://pdfaiw.uspto.gov/.aiw?PageN.../1/2019%26RS=(AANM/apple%2BAND%2BPD/20190801)

The APM is unusual in Apple's development process as it's one of the few Apple products for which we can get a decent glimpse of its development history through four years of rather explicit patents.

 

[Robbo99999]

I do feel indeed that protocols may have a role to play, but maybe not protocols alone to that extent. If that were the case we'd see this sort of gross variation in measured response for all headphones I think.
I feel that there's something else at play here.



All I'm saying is that... I'm very skeptical that sample variation can explain such large discrepancies between industry standard rigs, at least to the degree that @eddantes highlighted. It's really that, a question of degree, which in the case of the APM is staggering. To the point of being worrying if that's what real humans experience (but maybe not...).



No it's just that as my head is asymmetrical I usually can't measure L/R channel imbalance. But since with the APM I can reverse them and they remain relative to one ear or the other acoustically nearly identical, it's quite valid to do it this way .

BTW Apple entertained for a while the idea to make them left / right reversible and it's likely that the mostly front to back symmetrical design may be an offshoot of that original goal :
https://appft1.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PG01&p=1&u=/netahtml/PTO/srchnum.html&r=1&f=G&l=50&s1="20200280785".PGNR.&OS=DN/20200280785&RS=DN/20200280785
View attachment 146324View attachment 146325
The final earcup shape was already there at least as far as 2018 :
https://pdfaiw.uspto.gov/.aiw?PageNum=0&docid=20190238968&IDKey=61A77CF7D6D9&HomeUrl=http://appft.uspto.gov/netacgi/nph-Parser?Sect1=PTO2%26Sect2=HITOFF%26u=%252Fnetahtml%252FPTO%252Fsearch-adv.html%26r=8%26p=1%26f=G%26l=50%26d=PG01%26S1=(apple.AANM.%2BAND%2B20190801.PD.)%26OS=aanm/apple%2Band%2Bpd/8/1/2019%26RS=(AANM/apple%2BAND%2BPD/20190801)
View attachment 146326
The APM is unusual in Apple's development process as it's one of the few Apple products for which we can get a decent glimpse of its development history through four years of rather explicit patents.

I think the placement variation graph from Oratory that GaryH posted showed a lot variation itself, so if comparing Amir's single measurement it could be sitting almost anywhere amidst all those possible curves. It's accepted by Oratory & Crinacle & I believe Amir that the Harman Target Curve is valid for all of their different GRAS measurement systems they use, of which Oratory uses two slightly different ones although they are the same re pinna (just one is a full dummy head and the other is more like Crinacle's).....so if the Target Curve is valid between them then by extension that also means that the measurements would be comparible & valid between them (dependant on user protocol though still of course like we mentioned). So I think perhaps we're going down a rabbit hole for no reason.

 

[MayaTlab]

I think the placement variation graph from Oratory that GaryH posted showed a lot variation itself, so if comparing Amir's single measurement it could be sitting almost anywhere amidst all those possible curves.

So I think perhaps we're going down a rabbit hole for no reason.

The problem isn't that Amir's results are different from the others (eyeballing it I don't think that they seem to diverge from the average any more than the others), it's that none of them are similar enough . Without taking into account Amir's measurements, even with spatial averaging involved for the other methodologies, results remain quite substantially divergent in a way that I'm not sure we have a decent enough explanation for.

 

[JJB70]

I thought Crinacle was on target with these. If they were priced equivalent to the Sony 1000XM4 and Bose 700 they would be a rock solid recommendation for anyone invested in the Apple system. However at the price, and considering it is not difficult to further find useful discounts on alternatives then they look very expensive. But, regardless of argument over measures I really found that they do what they do very well if people like a bit of bass and brightness (and many do).
I am not sure any of it matters. The only people that matter to Apple are those who might buy them. And for those thinking of buying I don't think the stuff being discussed here is relevant and price probably isn't an issue (people don't usually buy Apple is they're looking for cheapness).

 

[Robbo99999]

The problem isn't that Amir's results are different from the others (eyeballing it I don't think that they seem to diverge from the average any more than the others), it's that none of them are similar enough . Without taking into account Amir's measurements, even with spatial averaging involved for the other methodologies, results remain quite substantially divergent in a way that I'm not sure we have a decent enough explanation for.

You're just using linguistic gymnastics there. There's still user protocol variation that can sit within the scope of variation that GaryH showed with Oratory's range of individual measurements which were used to arrive at an average for that particular headphone....lots of variation for that particular headphone.

EDIT: it might be useful for @crinacle or @amirm or Oratory (if he were here on ASR) to chime in on as to how valid their respective equipment is for comparison between each other, although I'm sure they would accept measurement protocol, unit to unit variation, and pad wear as acceptable/understandble variations between any comparable measurements.....but I think they would agree the that if all those were equal then they would arrive at the same result.....I remember Oratory saying that his two GRAS measurement systems are directly compatible/comparable for instance as the core fundamentals are the same, even though one is a full dummy head & the other is an "ear with a clamp". So it's no coincidence that they all use the same Harman Target Curve - because it's valid on that equipment.

 

[GWolfman]

Thanks