Jump to content
IGNORED

Article: Apple Music Lossless Mess Part 2: AirPlay


Recommended Posts

16 minutes ago, Rijckholt said:


Hi @tgp-2,

 

Thanks for your reply. While the icons are indeed a tell tale of the protocol that is being used (AirPlay 1 at 44,1 kHz ALAC ) it does not necessarily mean that what is inside that ALAC wrapper is the original (lossless compressed) content. Based on my network traffic I suspect that the Apple Music and classical app are compressing in lossy way BEFORE it is handed over the the AirPlay service of the phone. Resulting in sending less data wrapped in the ALAC AirPlay1 stream (like playing an MP3 via Airpay 1 in the old days).
 

This would have been introduced in the either iOS 15 or 16 as in iOS 14 (if the testing done as described in this article is correct) the input and output are identical. 
 

I am suspecting Apple made changes and this is no longer the case. However I cannot test it in my setup. Except by looking at network traffic that sits well below the 1000 kbps that I would expect when a lossless compressed music file (CD quality) is streamed. 
 

In order to confirm this you need to know what is sent and what s unpacked in the DAC and compare the two. Maybe @The Computer Audiophile or @Marco Klobas can help here.

 

cheers

Onno

If I can find some time I’ll look into it. I may wait for iOS 17 as it’s right around the corner. 

Founder of Audiophile Style | My Audio Systems AudiophileStyleStickerWhite2.0.png AudiophileStyleStickerWhite7.1.4.png

Link to comment
14 hours ago, Rijckholt said:

Based on my network traffic I suspect that the Apple Music and classical app are compressing in lossy way BEFORE it is handed over the the AirPlay service of the phone

 

can you describe what you're seeing in the network traffic? ... i can't imagine why a lossy encoding step would be needed right before the Airplay transcode to ALAC

Link to comment

@tgp-2 it definitely not needed :-) but apple might have decided to do it in order to preserve battery life of the phone/iPad. Why not give the user the option to stream lossless via AirPlay 2? Apple made sure that doesn’t work and now maybe also closed the route via AirPlay 1 form their own apps (music and classical). Just need to make sure what is going on.  As for my network traffic: it is just minimal more in line with lossy compressed streaming than with lossless streaming. Just don’t see enough data flowing from my phone to the AirPort Express. 

Link to comment

@Marco Klobas thank for your reply. Yes hoping for losses via AirPlay 2 or Roon integration would be nice indeed. As for your testing, not sure about your answer. I don’t think you need more advanced tool but do same test again but now with latest version of iOS and see if your results are stille the same as when you tested with iOS 14 2 years ago (this is also only focused at mobile iPhone/iPad streaming from Apple Music/classical not about streaming from a Mac computer) 

Link to comment
5 hours ago, Rijckholt said:

@Marco Klobas thank for your reply. Yes hoping for losses via AirPlay 2 or Roon integration would be nice indeed. As for your testing, not sure about your answer. I don’t think you need more advanced tool but do same test again but now with latest version of iOS and see if your results are stille the same as when you tested with iOS 14 2 years ago (this is also only focused at mobile iPhone/iPad streaming from Apple Music/classical not about streaming from a Mac computer) 

 

I repeated the test, recording on a Mac (macOS 13.5.1) a song played via AirPlay 1 from an iPad (Music app – iPadOS 16.6) and comparing it with the original track with inverted phase.

 

I recorded the audio with Audition using the same 2 methods already mentioned in the past posts to AirPlay the song:

  • Via Shairport Sync
  • Via AirServer software receiver (on intel Macs the AirPlay receiver built in macOS isn't available anymore)

A third recording has been made with Shairport Sync via its stdout option, again as already made.

 

The results are unfortunately even more puzzling. I expected the same behavior. Instead I got a different outcome. These inconsistent results are really starting to annoy me...

 

In my last recap post I mentioned:

  • The ShairPort Sync iPad to Mac AirPlay 1 (track #4) shows no difference (as expected because it's nulled) – Bit perfect
  • The AirServer iPad to Mac AirPlay 1 (track #3) shows no difference (as expected because it's nulled) – Bit perfect
  • The ShairPort Sync iPad to Mac AirPlay 1 (track #2) recorded via stdout shows a difference at around -190 dB. In this case it has a more irregular range (going from -140 db to -240 db). Sometimes it seems that the tracks null each other for a fraction of time – No bit perfect

Here's the complete post:

 

Now, the third point (the stdout recording) is still valid (at least one certainty...).

 

The other two recordings (Audition via ShairPort Sync and AirServer) are not.

 

Here are the three screenshots highlighting that none are bit prefect:

 

spacer.png

spacer.png

spacer.png

 

Acoustically, all three recordings are nulled (dead silence) leading to think that it's bit perfect. If you look at the frequency analysis, you see at -150dB or even lower that it's not true.

 

Granted, many things have changed since my last tests. iPadOS, Music app, macOS, AirServer, Shairport Sync, Audition: all have been updated since. Only the tested album is the same. 🙂

 

I know that my null test I risky because the songs have to be synchronized as perfectly as possible (matching their tempo) plus the convoluted recording procedure (AirPlay software receiver -> Loopback audio interface -> Recording software). It's the only method at my disposal.

 

I don't know why the AirServer and Shairport Sync recording sessions don't null completely this time. I tested twice. Apparently the procedure was the same – the results don't match, though.

 

I'm sorry @Rijckholt if I'm only confusing you. 😉

 

I suggest you to enjoy your Dynaudio Focus 50 speakers. Even via AirPlay 2. Who cares! As an owner of the Contour 20i I'm sure they sound great! 😀

 

Maybe in the future I'll give a try again. I'll wait some changing signs from Apple first...

Link to comment
5 hours ago, Rijckholt said:

 As for my network traffic: it is just minimal more in line with lossy compressed streaming than with lossless streaming. Just don’t see enough data flowing from my phone to the AirPort Express.

 

getting a good look at the network traffic can be difficult ... it's easier when either the sender or receiver is a general purpose computer such as a mac, raspberry pi, etc. ... where possible i use the utility "bmon" (i'm sure there are other good options) ... the receiving end of an Airplay1 stream will look something like this ...

Screenshot 2023-08-20 9.48.57 AM.png

Link to comment

Out of curiosity I tried something different: no Apple Music involved.

  • Loaded the original song on Plex server
  • Played the song through iPad Plex app via AirPlay to both AirServer and Shairport Sync
  • Recorded both sessions with Audition as usual
  • Compared the recordings with the original song (phase inverted)

Result: both recordings show the same "activity" at -150db or thereabout. Nulled acoustically, not through frequency analysis. Not bit perfect.

Link to comment
20 hours ago, Marco Klobas said:

Acoustically, all three recordings are nulled (dead silence) leading to think that it's bit perfect. If you look at the frequency analysis, you see at -150dB or even lower that it's not true.

 

Granted, many things have changed since my last tests. iPadOS, Music app, macOS, AirServer, Shairport Sync, Audition: all have been updated since. Only the tested album is the same. 🙂

 

I know that my null test I risky because the songs have to be synchronized as perfectly as possible (matching their tempo) plus the convoluted recording procedure (AirPlay software receiver -> Loopback audio interface -> Recording software). It's the only method at my disposal.

 

I don't know why the AirServer and Shairport Sync recording sessions don't null completely this time. I tested twice. Apparently the procedure was the same – the results don't match, though.

 

I'm sorry @Rijckholt if I'm only confusing you. 😉

 

I suggest you to enjoy your Dynaudio Focus 50 speakers. Even via AirPlay 2. Who cares! As an owner of the Contour 20i I'm sure they sound great! 😀

 

Maybe in the future I'll give a try again. I'll wait some changing signs from Apple first...

 

Hi @Marco Klobas

 

Wow, what an effort you put in! Highly appreciated! I was hoping it would be easier to determine if things are lossless or not ;-). I have also asked Otto Jørgensen from  @DYNAUDIO in another forum if they are able to confirm if Apple Music >> Ariplay1 >> Airport Express >> Toslink >> Dynaudio Focus is bit perfect or not.

 

After reading and trying to comprehend your approach I was wondering what your graphs would look like when you compare a song that is certainly being lossy compressed via AAC using normal airplay 2. You would obviously not see bit perfertness but what about the nulling? Would it be very clear the song is now compressed? 

 

I really thought this was just a very easy thing to determine so I could decide to put the airport gen 1 in the chain or not ;-) You very rightfully point out that it would be better to just enjoy the Focus speakers. To that I must say I absolutely love them and my ears most likely are unable to hear the difference between ALAC and AAC indeed. The Contour 20i are superb (i know from a friend)!!! 

 

Cheers,

Onno

 

 

 

Link to comment
19 hours ago, tgp-2 said:

 

getting a good look at the network traffic can be difficult ... it's easier when either the sender or receiver is a general purpose computer such as a mac, raspberry pi, etc. ... where possible i use the utility "bmon" (i'm sure there are other good options) ... the receiving end of an Airplay1 stream will look something like this ...

Screenshot 2023-08-20 9.48.57 AM.png

 

My network runs on Unifi and i am using their dashboard. I am not sure how accurate and real time their stats are. I agree with you that analyzing network traffic can be done a lot better a computer endpoints. For my test I really want the sender to be the iphone or ipad using Apple music because that's what i am using in real life. The receiver at this point is a Dynaudio Focus speaker that is receiving a PCM signal via Toslink. So i cannot really debug that or install software (I have asked Dynaudio the same question btw). So i need to prepare/setup a computer that can capture the PCM signal from a toslink connection and then I can compare both ends. I thought this was al easy but learning it is more complicated than at first glance.

 

 

 

Link to comment
2 hours ago, Rijckholt said:

After reading and trying to comprehend your approach I was wondering what your graphs would look like when you compare a song that is certainly being lossy compressed via AAC using normal airplay 2. You would obviously not see bit perfertness but what about the nulling? Would it be very clear the song is now compressed? 

 

I've uploaded here a video comparing a few seconds of Norah Jones's song Sunrise. The top track is an AIFF (uncompressed, lossless and phase inverted during the playback), the bottom track is an AAC (lossy, converted from the AIFF).

 

When the AAC track is unmuted you hear a partially muffled sound – it reminds me, to some extent, a sound played in a metal tank.

Link to comment
  • 1 month later...

- Has anyone compared choosing Airplay via iOS Control Center versus Apple Music app Airplay choice?

- Also, what about the Control Other Speakers & TVs (COST) option in iOS? Does choosing an Airplay receiver via Control Center COST button, not the Airplay list check option, result in a lossless 16/44 connection?

I’d prefer to keep my Airport Express gen 2s as Airplay 2 plus, I have other Airplay 2 AV gear, speakers, and Apple HomePod minis.

If simply always choosing Airplay from the macOS or iOS Control Center (or COST) and never via Apple Music app fixes the problem, then, we are done, right? Well, done for Apple Music app users.

Link to comment

Just for giggles I downgraded to 7.6.9 on an Airport Express gen. 2 that I only use when in office and — after a strange warning, I did a power off by unplugging (because decades of using Apple stuff has taught me to never trust warnings like that and to just restart the device after any updating or downgrading glitch like that) — sure enough it came back on my network fine, and Airplay is the non circle check mark via latest iOS. So it is lossless 16/44 from Apple Music from iOS (latest) to this ancient relic.

To downgrade, open Airport Utility. I did it on mac mini on latest macOS but it would also be possible in iOS. On mac, click on the Express icon. Hold down option key while over the version. A drop down arrow appears. Click on that. Choose 7.6.9 and it will downgrade to Airplay 1. An unplug restart of the express may be needed if you get an odd failed warning. If like mine, it should just be fine.

I also tested via macOS latest version on M1 mini to it by choosing the AE from the Finder menu. It seems way snappier! With Airplay 2, it took the OS a few seconds to confirm the link from that menu. With the downgrade, it was instant. So something did happen for sure. 

Not that my next comment is reliable but, to my ears, it sounds like CD quality. I play CDs on that same system and, they sound great. Apple Music with this downgrade sounds great too. 

Link to comment

link to an updated version of what I shared in this thread earlier this year in pdf format ... added sections on AirPlay from a Mac and remote controlling lossless Apple Music playback on Apple TV ... note, for the Apple TV use case, my only HDMI connection is to a TV that I think only does 48kHz, so I'm not able to check for any sample rate conversion etc (will leave that to others here who are more experienced and better equipped) 🙂

 

Link to comment
11 minutes ago, tgp-2 said:

link to an updated version of what I shared in this thread earlier this year in pdf format ... added sections on AirPlay from a Mac and remote controlling lossless Apple Music playback on Apple TV ... note, for the Apple TV use case, my only HDMI connection is to a TV that I think only does 48kHz, so I'm not able to check for any sample rate conversion etc (will leave that to others here who are more experienced and better equipped) 🙂

 

The COST button comes in two flavors on iOS. One is the one you show. But it is the circle one that is accessed via Control Center that is the right one to use. After using the round one and selecting a device, launch Apple Music and a device will have been automatically selected as the device you control. This method seems to be a way to skip Airplay from playing from the iOS device and keeping an iPhone just as a “remote” so to speak. It is this that I wonder, maybe, that is not AAC when controlled this way. I dunno… This is not really even Airplay but is interwoven into the UI. Messy…

Note that, I have noticed if a device has not been used recently, it may not show up on this round COST option also, Apple confusingly tossed in a device that is getting data via the iPhone (Airplay) with devices that can independently play Apple Music (which for COST, seems right, not a mix). So, the device list can be a mix that changes. Ugh.

It is also all too easy to reflexively choose a device on the main layer/list of any of these. That is Airplay. Don't do that. Click on the COST button, then choose a device from the totally different looking stacked list of devices.
I think Apple should start over with Airplay and COST UI and make it simpler but clearer and not so many ways to find it and use it. Just one and done. Also, it should show unavailable stuff by making them grey. Not removing them entirely from the UI. 

Screenshot 2023-10-11 at 1.41.21 PM.png

Screenshot 2023-10-11 at 1.41.56 PM.png

Link to comment

sideremark: what makes me personally upset, that Apple is able to play lossless using the new Airpods 2 Pro (USB Charger) to Apple Vision Pro. They have the technology obviously ready.

 

Well, when I dig into this, I decided to use Airport Express Gen2 for 44.1/16 lossless. But you need a certain firmware to my experience not the latest. Did post above here. By using old Firmware with Airport Gen 2 you force the higher quality. Simply because it can not do Airplay 2 but only Airplay 1.

Link to comment
6 hours ago, MgP2804 said:

sideremark: what makes me personally upset, that Apple is able to play lossless using the new Airpods 2 Pro (USB Charger) to Apple Vision Pro. They have the technology obviously ready.

 

Well, when I dig into this, I decided to use Airport Express Gen2 for 44.1/16 lossless. But you need a certain firmware to my experience not the latest. Did post above here. By using old Firmware with Airport Gen 2 you force the higher quality. Simply because it can not do Airplay 2 but only Airplay 1.

- Tim Cook talked about this Vision Pro and audio in an interview and, I forget why but, it was expected and not upsetting. I think it boiled down to the Vision Pro was robust enough and close enough to Airpods for it to happen. The tech is ready but, is your wallet?

- Yes, I post above that that is what I did. Downgraded to 7.6.9 to get 16/44 lossless. But, that Toslink to AVR with MusicCast allowed me to hack it to send it via MusicCast. Enabling me to play 16/44 from Apple Music to Express to AVR to my MusicCast speakers, which can get and play up to 24/192. Which otherwise, as thier own Airplay receivers, are just Airplay 2 and probably not 16/44 but AAC.

If only Apple would let its Airplay companies like Yamaha, etc. do the same, have an Airplay 1 option, so to speak, we could skip this extra step with old Airport Expresses.

Link to comment
13 hours ago, actuel audio said:

- Tim Cook talked about this Vision Pro and audio in an interview and, I forget why but, it was expected and not upsetting. I think it boiled down to the Vision Pro was robust enough and close enough to Airpods for it to happen. The tech is ready but, is your wallet?

- Yes, I post above that that is what I did. Downgraded to 7.6.9 to get 16/44 lossless. But, that Toslink to AVR with MusicCast allowed me to hack it to send it via MusicCast. Enabling me to play 16/44 from Apple Music to Express to AVR to my MusicCast speakers, which can get and play up to 24/192. Which otherwise, as thier own Airplay receivers, are just Airplay 2 and probably not 16/44 but AAC.

If only Apple would let its Airplay companies like Yamaha, etc. do the same, have an Airplay 1 option, so to speak, we could skip this extra step with old Airport Expresses.

it is definitelly not my wallet. thatfore I am upset. it is just a chip! +++ yep 7.6.9. sounds great with airport gen 2. this is my experience too. :)

Link to comment
2 hours ago, actuel audio said:

So somehow, Apple Music is casting to an Airplay 1 device which can only do Airplay 1 whilst also casting to a couple of independent MusicCast stereo mode speakers.

 

yes, it seems adding an AirPlay 1 receiver to a multi-room group is supported when streaming from Apple Music on a Mac (it's not an option when streaming from iOS), and this forces the use of real-time (AirPlay 1) streams to all receivers in the group

 

but it looks like the Mac's Apple Music app isn't receiving lossless in this case ... the lossless indicator disappears, and a rough check of the network traffic also suggests a smaller/lossy file downloaded to the Mac ... so it's a real-time / ALAC stream, but from a lossy source (and in this case it's not clear why the lossy source would be necessary) 🤔 

 

compare this to selecting only AirPlay 2 receivers ... the lossless indicator returns, and there's more data downloaded to the Mac ... but the buffered AirPlay 2 streams are lower bitrate / lossy

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...