Apple AirPods Max
Apr 29, 2021 at 11:10 PM Thread Starter Post #1 of 55

bigshot

Headphoneus Supremus
Joined
Nov 16, 2004
Posts
26,954
Likes
6,386
Location
A Secret Lab
I'm curious how well the head tracking and dimensional audio work in the Apple AirPods Max. Has anyone tried them with multichannel content? Or with stereo content synthesizing true speaker soundstage?
 
Apr 30, 2021 at 3:08 AM Post #2 of 55
I also wonder how well this works and about the methods used.
Of course, I believe that discussions of the science or technology of any sound related topic (3D sound, recording and playback theories and implementation, acoustics, hearing perception, etc.) fits well in this section, even when appropriate subjective reports are included (like this thread).

You must realize that other people’s experience may not apply to you, since you’ve said several times that binaural doesn’t work for you. I think it’s clear that your HRTF likely falls outside the normal range. You’ve also dismissed spatial audio in headphones many times, since “spatial audio requires space”. Or have you accepted the error of your previous(?) opinion.

But the irony of you soliciting subjective experiences in the Sound Science section is hard to miss, since you love to try to shoo people away with your “here in Sound Science, we…”. I guess even gregorio recognized the hypocrisy of your switching from objective to subjective, depending on your agenda.

Back to the AirPods Max: you may find people who own them in the existing thread here.
Apple has a patent for the technology (patent here, and described here). Some less technical discussion can be found here.

EDIT: oops, the patent link above doesn't include the figures. Here is the PDF:
 

Attachments

  • US10880649B2.pdf
    955.7 KB · Views: 0
Last edited:
Apr 30, 2021 at 6:41 AM Post #3 of 55
Go away
 
Apr 30, 2021 at 10:13 AM Post #4 of 55
I did! I went to the other thread, and also other forums and Apple.

I found that had you asked a month ago, you’d find that you have to use an iPhone (7 or later) or iPad (similar age), and one of the specific apps that have specifically been programmed to use Apple’s “Spatial Audio”. At this point all of them are for playing/streaming video and Spatial Audio requires 5.1, 7.1 or Atmos sound in the video.

I’ve only read a dozen or so subjective impressions: all are positive. So far, no one has said the effect didn’t work for them, but I’m still reading. One cool thing is the sound is “centered at the device” and so if you move your head OR the device, the sound moves appropriately.

But the recent Apple rollout included a new Apple TV 4K without Spatial Audio (yet!), and the new M1 iMac with Spatial Audio. I have not found any direct statement that the AirPod Max will work with it though.

I’ll “go away” again and look into it some more. I might have to buy them to play with it… stay tuned…

EDIT: Here (link) is a list of apps that work (list nearly 2 mo.s old)

p.s. use the "ignore" feature to avoid my posts, if that helps.
 
Last edited:
Apr 30, 2021 at 3:10 PM Post #5 of 55
Thank you for discussing the topic and not just me. I’m much more interested in your comments on the topic.

In particular, I’m interested to hear how well head tracking works and how important that is for sound location cues. I’ve gone back and forth on that myself, because it’s hard to isolate the factors to determine relative importance. Here we have head tracking and some sort of microphones within the ear cups that give feedback to the signal processing. The latter is a completely different technique than the Smyth Realiser. I’m curious how it works and how well it works with various people.
 
Last edited:
May 1, 2021 at 1:37 PM Post #6 of 55
You're welcome... but you're so "discussable"...

Quick update and speculation: I have yet to find a negative impression of the head tracking, but I've also not found an in-depth discussion. I know from Realiser users and my own work that head-tracking can/will play a central role in the immersion/presence/realism of the illusion..
A lot of people have complained that there is no way to just listen to music (as opposed to watch/listen to a video) This story (link) prompts my speculation that it may be available in the future with Apple Music.
 
May 2, 2021 at 8:43 PM Post #7 of 55
Well, I took one for the team. I ordered a set. I have a new iPhone that supports all the DSPs the AirPod Max can do, so I'll give it a spin. I imagine I'll have trouble figuring out how it works. (Just as I am struggling to figure out how the multiple cameras on the back of my phone work...)
 
May 2, 2021 at 9:59 PM Post #8 of 55
I'm curious how well the head tracking and dimensional audio work in the Apple AirPods Max. Has anyone tried them with multichannel content? Or with stereo content synthesizing true speaker soundstage?
I’ve had them and tried it, also have it on my AirPods Pro. It’s the same thing. Verticality is what it’s lacking, and it isn’t a particularly “real” experience even when moving your head left and right. It isn’t bad, but not even close to a true multichannel system. Maybe next iteration will be better.
 
May 2, 2021 at 10:27 PM Post #9 of 55
Do you have a recent iPhone with dimensional audio support?
 
May 3, 2021 at 5:12 AM Post #11 of 55
I'm researching how to access Atmos content and view it. The two sources I have for it are the Apple TV app and the HBO Max app on the iPhone itself. (I don't have an Apple TV set top box, and I won't buy one if this isn't worthwhile.) The only way I see to get the picture on a bigger screen than my iPhone (again no set top box) is to mirror the phone on my iMac using Quicktime with a lightning to USB cable. I'll give that a try and see how it works.

Another thing I'm researching is spatial audio using stereo music. Apparently it is possible to switch on spatial audio with any stereo source using the Headphones settings. I'm guessing this is similar to Dolby Pro Logic II channelled for spatial audio in the AirPod Pros. Interested to see how that works too, and if it is something like a Smyth Realiser Lite.

The reviews I've read say that the soundstage is focused on the location of the phone, so if I plug it into my iMac and set it underneath the screen, it should focus the multichannel on the screen. The adaptive EQ based on the array of microphones inside the ear cups sounds interesting too.

We will see. I can always return the cans to Amazon if they are a fail. I'm very happy with my new iPhone Pro Max, so hopefully the headphones will be good too.
 
May 3, 2021 at 11:36 PM Post #12 of 55
OK. I've spent the afternoon and evening with the AirPods Max. I'm going to keep them. They are definitely massively overpriced, but they have definite advantages over other cans.

First of all the spatial audio and head tracking... No. Not at all like multichannel. There is good control of the left and right panning when you turn your head, but there is absolutely no depth nor rear channel information. The rear channel info shows up for me down by my waist, and it doesn't have any left or right- it's kind of like Dolby Stereo with just one channel behind (or in this case by my belt.) Maybe it works better for other people. I listened to some Atmos programming on Apple+ and HBO Max. It sounded very good, but the voices were still inside my head, not up front in the center channel. It works if you hold your phone 8 or 10 inches from your face, but it isn't soundstage by any stretch of the imagination. Head tracking works if you hold the phone still, but if you move the phone at all, it defaults to the middle. Spatial audio does absolutely nothing for music.

However, there is something going on in music. Apple calls it "adaptive EQ" and it appears that the DSP is analyzing the music and breaking it down into pieces and handling it separately, much like the new iPhone cameras take several pictures and paste them back together into a combination that is better than any of them alone. It works remarkably well in rock music, especially Pink Floydy stuff with lots of ping pong stereo and phasey effects. It places sound and pans it across the channels really well. When something pans from left to right, it's extremely effective. The panning element is super clear and consistent from side to side. It's also good for guitar solos, making them sound clearly separate from the music around it. The effect is kind of like HDR for music. Hard to describe. It can be fooled though. In Back In The USSR, it has no idea how to handle the jet sounds, muting one channel almost completely and leaving tape hiss on the other. I also ran into a problem with a classical recording where the adaptive EQ grabbed onto the room tone and sculpted it into a clear sound, rather than a smooth fade up. I don't like what the adaptive EQ does to orchestral music. It tends to mass the strings and make them sound like a movie soundtrack. I definitely won't be using these for classical music.

The response curve is very good. A little adjustable in the settings, but the default was best for me. There are a couple of little tweaks I would make (and probably will if I can figure out how normal EQ works with adaptive EQ.) It has a bass hump between 80 and 160, and the upper range of vocals is a little hot around 3 or 4kHz, making male vocals sound a little thin, Not terribly far off, but compared to my Oppo PM-1s, it is clearly inferior. The Oppos sound a hair clearer at loud volumes too. Much better than the AirPod Pros. The noise cancellation is state of the art. You won't find better.

But the surprise is how damn well it works. Pairing is a breeze. Playing music and controlling playback is simple. It turns itself on and off. No power button. And it senses when you put them on and turns them on and pairs. Like all Apple products, it is well designed and simple to use. But these are easier to use than my AirPod Pros. They're comfortable, lighter than my Oppos and easy to toss in a bag and go. That's what I'm going to use them for. These and my iPhone are all I need, and I have the AirPod Pros if I need ear buds.

I'm going to keep playing with the dimensional audio and try to figure out the adaptive EQ. I've had some bluetooth connection problems too. Gonna have to figure that out.
 
Last edited:
May 4, 2021 at 1:40 AM Post #13 of 55
However, there is something going on in music. Apple calls it "adaptive EQ" and it appears that the DSP is analyzing the music and breaking it down into pieces and handling it separately

It's a lot simpler than that, it's just a way to deliver a constant FR below a certain frequency with the help of the inward facing microphone, regardless of seal / pads compressions / etc., by adjusting in real time the FR :
https://twitter.com/oratory1990/status/1343323936801644544/photo/1
Other ANC headphones also do this, for example :
https://diyaudioheaven.wordpress.com/headphones/measurements/bose/qc35-ii/
The results being, in general, significantly better consistency across a varied range of users below 1khz or so, unlike most other closed backs (well with the exception of one individual for the APM in Rtings' tests) :
https://www.rtings.com/headphones/1-4/graph#16092/4007/565
vs your typical closed backs :
https://www.rtings.com/headphones/1-4/graph#1671/4007/669
(Rtings tests bass response on five real humans with five seatings each, not on a dummy head, and merge their measurements with their HATS past a certain frequency).
Apple has some patents to make Adaptive EQ work even when ANC / transparency modes are off (https://patents.justia.com/patent/9515629#history).

It's unlikely to be effective above 1khz, that's the next frontier for headphones design.
I've recently read this paper which seems to propose an automated way (but not in real time) to apply some form of personalised EQ above 1khz and below a few kHz by using only one single mic, of the kind most ANC headphones already have :
https://www.aes.org/e-lib/browse.cfm?elib=20605
The N90Q from AKG also tried to provide a way to correct the FR past 1khz for each individual, also not in real time, this time with two mics inside the cups.
 
Last edited:
May 4, 2021 at 1:44 AM Post #14 of 55
Ooo! Let me try moving them around on my head... Yeah! There isn't a sweet spot. I can even pull the cups off my ears a cm or two and it still sounds the same. Interesting.

Do you know anything about the spacial audio? There's some sort of phase thing going on. The sample audio in the settings has dry sound in the stereo and lots of secondary depth cues in the spacial audio sample. Can it actually separate out sound sources and add phase or reverb effects to it?
 
May 4, 2021 at 1:56 AM Post #15 of 55
Ooo! Let me try moving them around on my head... Yeah! There isn't a sweet spot. I can even pull the cups off my ears a cm or two and it still sounds the same. Interesting.

I'm not sure of it but I believe that below 1khz over-ears still operate mostly in pressure chamber conditions, so the exact front / back / up / down position of the headphones isn't that critical. Adaptive EQ rather helps with varying front volume (volume of air between your ears and the driver), pad ageing / compression, break of seal, etc.
The idea isn't new (Bose has been doing it for years) but it's possible that Apple's implementation is more refined.
In any case given the rather dumb idea of applying the headband's clamp force at the top of the earcups, which then forced Apple's engineers to make the most sophisticated piece of mechanism ever designed in a pair of headphones (the spring loaded joint : https://www.ifixit.com/Teardown/AirPods+Max+Teardown/139369) to compensate for that - partially, as it can't be fully successful given the invariable spring rate of the joint mechanism, unlike the dumb, simple yoke mechanism that naturally exert clamping force in the ideal place -, Adaptive EQ definitely helps a lot people like me with seemingly edge case morphologies which tend to introduce breach of seals or uneven pads compression.

Do you know anything about the spacial audio?

Nope, only that for me it's yet another surround sound simulation that failed to escape the uncanny valley (I find it OK for left / right distinction but that's pretty much it, I'm not getting much in the case of front / back or up / down differentiation, or depth). It uses one generic HRTF profile so maybe that's one of the reasons it's not that successful for me. Or maybe not. IDK.
 
Last edited:

Users who are viewing this thread

Back
Top