SUMMER MUSINGS: Surround sound to the masses! In support of 3D Audio / Dolby Atmos in music streaming (Apple "Spatial Audio")...

Hey guys and gals, this week I thought I'd take a break from measurements and observe what's going on in the audio, audiophile, and streaming world these days. I noticed some discussions and even concern among music listeners around Atmos and Apple Music's drive towards "spatialized" streaming.

As audiophiles, let's think and talk about this for a bit.

First, let's do a quick background recap so we're all on the same page...

I. Background, Tools, and Subjective Choice

Over the years, discrete multichannel audio like 5.1 or 7.1 where sound is designed to play from a specific multi-speaker configuration has been the traditional way that "surround" sound has been implemented. More recently, Dolby Atmos (in cinemas 2012, home theater by late 2014), dts:X (2015), Auro 3D (AuroMax with "objects" released 2015), and Sony 360 Reality Audio (MPEG-H 3D Audio, Sony demo ~2019?) have changed the landscape with systems that are able to encode the 3D sonic space as defined by "objects" relative to the listener's position. This frees the audio from the constraints of specific multichannel layouts to a certain extent although traditional multichannel 5.1, 7.1, etc. would remain as the "bed" channels - see here. It means that when played back, the audio system uses that extra 3D positional information in its DSP algorithms to determine what each target speaker, whether it be flat 5.1 / 7.1, with or without the extra height dimension (eg. 7.1.2 signifying 2 extra height channels), should play. In the same way, that information can be "virtualized" into binaural 2-channels using headphones to "trick" the brain to hear direction and depth based on HRTF modeling. Depending on which system one is talking about, there are limits as to the number of "objects" and the number of speakers the system can target to. These limits should not be a concern for personal / home use and can go up to big numbers like 128 objects, 64 speakers in cinema mixes.

Note that with object-based 3D audio, since the output is recalculated for the target audio system, there is really no longer such a thing/need for strict "losslessness", or "bit perfect" (other than simply accurate data transmission). In fact, multichannel playback almost always includes some form of processing anyways whether it be adjustment of relative front/rear/side speaker levels or bass management to suit the room at the most basic level. This is something that audiophiles need to appreciate and not fret over. For years, we've talked about digital room correction (like this) as another example of the sophisticated algorithms needed to achieve the results which will alter the data sent to your DAC. Not adhering to strict "losslessness" is actually representative of the power of the technologies and ability to meet the needs of individual listeners and must not be viewed as a negative within this context.

In the last week, I've noticed some negative talk about Apple Music and their promotion of Atmos remastering of music (Lefsetz Letter) which Apple also calls "Spatial Audio". Don't forget that Tidal and Amazon Music HD also offer Atmos material as well as limited MPEG-H / Sony 360 Reality Audio but only for restricted playback devices.

Regarding that Lefsetz Letter article, there are some points I certainly agree with. I would not support rushed jobs, poor remixing and deterioration in sound quality either! However, I would not call 3D audio a "gimmick". Is it any different than other studio techniques applied to affect the sound of the music? I'm sure at some point in the 1950's, cranky audiophiles probably also considered 2-channels gimmicky. In the modern era, Auto-Tune, and countless other ProTools / DSP plugins also can be considered "gimmicks". If used well, and artistically intended, such techniques have in fact defined so much of the music we hear these days. Cher's song "Believe" from 1998 showed off Auto-Tune in a big way! Yup, it was a "gimmick", but yet so much of modern pop recordings use the same trick now. Do we then write off the whole genre altogether? Or do we as listeners appreciate some of these changes and possibly even embrace the artistic intent? That is a choice each of us can make but I hope that the audiophile hobby is big enough to realize that musical tastes change and that the hobby is broad enough to include all music lovers and genres. Otherwise, the hobby might become nothing more than aging old dudes bitching about old hardware, and old recordings. ;-|

Realize that we're merely talking about "tools". Whether the creative music output is revelatory or debased gimmickry is a reflection of the artists' / studio engineers' abilities. Whether 3D positional audio, or some DSP is inherently "good" or otherwise is not the debate here.

[BTW, here's a good video by Rick Beato on Auto-Tune.]

No matter what, any kind of remixing and remastering of well-loved music will have its critics. Not everyone will like a multichannel / 3D mix. For example, I thought the Blu-Ray Sgt. Pepper (5.1) and Abbey Road (7.1 + Atmos) were good but certainly you're not going to find universal acclaim especially for landmark albums like these!

When it comes to older recordings on streaming sites, I agree that those produced in mono/stereo must always be available in the original form. No "fold down" of surround-to-2.0 will achieve the exact sound of the original. While I can't say I am particularly driven to listen to older songs like Marvin Gaye's What's Going On regardless of whatever 3D version is available on Apple Music, I would hope that the "original" 2-channel stereo version with good dynamic range, will always be available for everyone interested in reliving what was heard when it first came out in 1971.

Some music lovers may be upset that it's "desecration" when classic recordings are remixed especially where the original artists are unable to "authenticate" the intent or vision of the music. However, let's not forget that remastering of most of our favourite recordings has been going on for years with or without the artist's direct input. If we buy new CDs of albums from the '70s and '80s, we're more than likely going to be picking up modern remasters, typically with a different EQ than the original, louder and more dynamically compressed. For example, I truly detested the UMG remastering of The Rolling Stones back in 2009/2010 - terrible sounding, distorted DR6 Sticky Fingers, and DR7 Exile On Main St. are examples of this trend. Is this any different than a multichannel Atmos mix? Will audiophiles be up in arms if these 2009/2010 remasters are the only versions available on their streaming service? (Arguably, audiophiles should be if these remasters are all that's available on their streaming platform!)

A deep catalogue these days IMO should include not just many songs and artists, but also the various remasters and remixes. In the same way, I hope that this is what happens to the 3D / Atmos mixes. That they're presented as another version in the large catalogue of albums that the listener can choose from.

Having multiple masters of the same album should not be a problem I hope. For example, here's Miles Davis' Kind Of Blue on Amazon Music HD:

Notice that there are 3 Mono releases on there (with 24/96 and 24/192 versions); so why not a 3D Atmos mix as well? (Heck, maybe also include the multichannel 5.1 SACD versions.)

II. Audible Differences and a Listen to Apple's "Spatial Audio"

In the last decade or so, the music industry has been courting audiophiles to buy "Hi-Res Audio", suggesting to consumers that these make a difference. I think most of us are familiar with the obvious truth - that "hi-res" audio sounds basically the same as lossless CD - and I think many audiophiles are "post hi-res" by now.

For years I've believed that surround sound is the future of audio. Yes, I know that analogue quadrophonic failed in the '70s and discrete multichannel SACD/DVD-A as music-only formats didn't do well in the early 2000's. Just because up to this point surround / multichannel hasn't made inroads as a major go-to format for music lovers doesn't mean that this will always be the case. Often, to make it stick, the technological foundation needs to mature - we might finally be there with 3D music.

As I expressed back in 2019, headphone virtualization of true 3D content can be a powerful factor that elevates its value. No longer tied to a multichannel home theater room, no longer forced into pre-defined multi-speaker arrangements, but rendered using modern DSP technology on smartphones and headphones, this will open up the content to many more listeners. I'm actually quite excited about the opportunity for artists and audio engineers to explore the extra spatial dimensions and for manufacturers to find creative ways to extract that information during playback across mobile and sound room environments for listeners. Furthermore, features such as head tracking can be added for yet another level of sonic exploration.

When listening to the music, you might like the remixed and remastered Atmos version, or you might not; but regardless, you will experience a difference! Much more so than just lossless or just "hi-res".

When it comes to the evolution of audio hardware, after all these years, I suspect 2-channel DACs and streamers have already reached maturity and market saturation with slowing growth and innovation potential. Multichannel will be good for the industry as it spurs future generations of hardware and refines the algorithms for quality rendering. (For example, I'd love to have a way to customize the HRTF to create a more convincing "surround" effect - RedScape and Creative at least offer ways to adjust the settings.)

Older music... New Dolby Atmos mix.

When an 800lb gorilla like Apple (finally) makes a move like this, there is the potential for significant changes ahead. As audiophiles, I believe it's better to embrace change so we can take part in discussions and help shape the future of these technologies. Sure, we might decide for ourselves to stick with vinyl or 2.0 playback. However as a body of audiophiles, let's make sure that we do not end up stuck in some kind of defensive position hanging on to anachronistic devices and ideas only. To shun advancements risks further marginalization of the "audiophile hobby" from mainstream music and the next generation of passionate music lovers who appreciate high quality hardware.

I would not be surprised if looking back in a few years, many of us will be pointing to the 3D version as the "definitive" album from some of our favourite artists just as there was a time decades back when stereo took over from mono.

For now, I anticipate that there could be pushbacks from more conservative music consumers. Let's see if the audiophile magazines say something about this! For me, so long as Apple and others remember to always keep the 2-channel version available for older recordings out of respect for the artists and engineers who first produced the content, bring on 3D audio!

I've been exploring the Atmos mixes over the last week using my iPad and wife's iPhone 11 through the AirPods Pro. As expected, the difference is clear and we'll each just have to listen and decide subjectively whether we enjoy them. For example, I didn't enjoy the Atmos version of Bob Marley's "Exodus" that much and thought The Weeknd's "Blinding Lights" had too much reverb. Likewise, what's going on with Lady Gaga's "Born This Way"? It sounded distant, mid-bass heavy, and unengaging. Gaga and Bradley Cooper's "Shallow" fared much better with tight accentuation of the solo guitar and voices, the sound of the audience appropriately pushed into the distance. Grover Washington, Jr.'s "Just The Two Of Us" sounded nice with better defined space between the instruments, great positioning of the percussion in the virtual soundstage. Kenny Rogers' "The Gambler" sounded clean, like it was recorded last week. Taylor Swift's duet with Brendon Urie on "ME!" was more fun in Atmos with greater subtlety to the sound. Swift's recent album evermore was remixed well which added a nice separation between some of the sound effects, instruments, foreground and background voices on a track like "no body, no crime". Want something heavier? Linkin Park's "In The End" from Hybrid Theory is pretty good.

Some of the classical music tracks sounded nice but the "concert hall" effect was of mixed quality. For example, I much preferred the standard stereo version of Kaori Muraji's Cinema - Movie Themes for Classical Guitar compared to what I heard as overly processed room reverb of the Atmos mix. However, I quite enjoyed Nadine Sierra's album There's A Place For Us on Atmos which did a good job isolating her voice front-and-center while spreading the instruments across the sides and behind her.

I'll wait for an Atmos remix of Rick Astley's "Never Gonna Give You Up" before rendering any final overall opinion on Apple Music even though Guns 'N Roses' "Sweet Child O' Mine" is pretty good. Not enough '80s Atmos stuff I say! While Norah Jones' Come Away With Me is a good start, until I can sample some Diana Krall albums in Atmos, I can't say if this is audiophile-approved. :-) 

III. An Opportunity for Better Sound? On Android, and Summary

Something I've noticed: the Atmos mixes are often not as loud as the stereo version, and they tend to be more dynamic. With this remix/remastering effort, there is a real opportunity here for artists and audio engineers to dissociate themselves from the music "Loudness War". A good multidimensional sound should include more dynamics, embody more subtleties, and distort less. I hope the industry doesn't blow it and consumers can be sophisticated enough to appreciate that louder often isn't better.

As you've likely seen on this blog, over the years we've discussed multiple instances of the problems with sound quality in music. Despite the idea that volume normalization in streaming should have reduced the tendency for "low-DR" productions, I think that the "culture" of how engineers push loudness as standard practice, how record labels expect albums to sound loud has been a hard one to beat. With 3D audio, let's hope that a new type of culture takes over. Perhaps the practices already laid out by cinema soundtrack engineers desiring to create a soundscape with retained dynamic impact, and in turn more subtle and desirable emotional impact can take over from the unceasing screams that have infected so much of the mainstream music output in the last few decades. This is something I hope audiophiles and the press can champion.

On a technical note, something I have not heard from Apple or seen in reports is a description of what form the streaming service is presenting the 3D mixes. For example, is Apple Music streaming this material as Dolby E-AC3 + Atmos? If so, at what bitrate (I'm going to guess that it's streaming between 640kbps to 1.5Mbps)? I'm sure the technically-oriented music lover / audiophile would be interested in knowing what we're listening to. Obviously, when listening with Bluetooth, the final data rate will drop significantly feeding the headphones. The Apple AirPods, Pros, and even the AirPods Max are all only AAC 256kbps devices.

Android users, for the time being if you want to hear Atmos on the Apple Music app, sign up to beta test here; make sure to turn on Dolby Atmos and might as well turn on Lossless Audio in Settings while you're there so long as no data cap issues.

Using wired and circumaural headphones gave me the best 3D impression and overall sound quality:

This Samsung Galaxy Tab S6 Lite with my Drop + HiFiMan HE-4XX streaming with Apple Music 3.6.0-beta app sounds noticeably better than through the AirPods Pro. Subtleties were better resolved and space on some of the merely "okay" Atmos tracks like Taylor Swift's "You Need To Calm Down" (her voice a bit too recessed IMO), The Weeknd's "Blinding Lights" and Billie Eilish's "Bad Guy" came through with better clarity. Of course, with such headphones, there's the issue of portability, and a better DAC/headphone amp would be preferred.

While headphones are nice, the best sound for me at home would be in the multichannel home theater feeding all the speakers. I'm guessing at this time the best way to pipe Atmos music into the AV receiver is with an AppleTV 4K and the Apple Music app using HDMI. Windows iTunes currently does not support Atmos. Knowing the way Apple is with keeping a tight ecosystem, I would love to see Amazon Music HD add more 3D content, and even more importantly, support more hardware, particularly a simple way to send the full resolution bitstream to an Atmos-compatible receiver. The fact that Amazon Music HD currently restricts the 3D content to playback using its Echo Studio speaker is a bit of a joke. Maybe Amazon partnering with Sony and their 360 Reality Audio / MPEG-H 3D for more content might be a nice move. Even better, I think for audiophiles, Roon + Amazon Music HD content plus 3D would be a dream!

Obviously, at this point the small players are going to be squeezed badly. Tidal needs to overhaul their price structure. They need to toss out their pseudo-hi-res. This is not looking good. At least Tidal's 3D offerings can be played back on more devices than Amazon Music.

For those who have not tried it, here's Dolby's Atmos Visualizer Demo - have a listen with headphones, see what you think. Also, I think it's a bit confusing when Apple uses the term "spatial audio" too broadly. Here's an article talking about some differences between Atmos and Apple's "Spatial Audio" as it applies to headphones. For now, "dynamic head tracking" with Apple Music playback is unavailable with a target of release in the Fall.

Fun fact: Dolby Atmos was announced in mid-2014 for home theater and the first Blu-Ray with Atmos - Transformers: Age of Extinction - came out by late September 2014. This is around the same timeframe as MQA's announcements and I wrote about them in early 2015. It's interesting tracing the paths that these CODECs took over the years and how they're now converging in the music streaming world. Obviously for me, one is not like the other and I do hope they take divergent paths toward some form of evolutionary success and some form of extinction.

Enjoy the summer (in Northern Hemisphere), friends. I'm aiming for some R&R ahead.

And of course, enjoy the music ;-).

PS: I see Mark Waldrep also wrote an article this week about Apple "spatial audio". Great stuff!

He makes an important point. If 3D audio technology were available back in the '60s, wouldn't the Beatles, Rolling Stones, The Beach Boys, Hendrix, The Who, Zeppelin, The Doors be right in there playing with these tools in the studio? Likewise, I bet in the '70s you'd have Steely Dan, Pink Floyd, The Eagles, Queen, Bowie, Curtis Mayfield, Rush, Yes, ELP, ELO, etc... doing the same.

PPS: Based on the discussions in the comments, I gave it a try with my wife's M1-based MacBook Air running the latest macOS Big Sur 11.4, Apple Music app connected to my Yamaha receiver by HDMI:

Nope, not going to get clean Atmos bitstream out from the MacBook. The best I could do with the Audio MIDI settings is 7.1.0/24-bit/48kHz output which you can see on my TV screen. I only have a 5.1.2 Atmos set-up but no Atmos so just 5.1.0.

It kinda works. The Atmos version will be decoded by the computer to the "bed" multichannel output. No problem with playing sequential tracks on a playlist, but it doesn't seem to want to stay on the Atmos stream even when I turn the setting to "Always On" Atmos. Once awhile I have to close the Apple Music app and restart to get back to the multichannel stream.

Anyhow, even without any fancy Atmos height channels, the 5.1 effect is nice to hear in the home theater. You can tell the difference between tracks with how many bed channels they're based on. For example, Taylor Swift's "Willow" is 4.0 (no center info nor use of LFE for sub), Kenny Rogers' "Gambler" is 5.1 with his voice in the center channel, Norah Jones' "Don't Know Why" sounds like a 5.0, and GNR's "Sweet Child o' Mine" appears to be 4.1.

Fun to try, but too much of a hassle with the Mac. Clearly, Apple wants us to use the AppleTV.