Atmos and TrueHD 7.1 playback on 5.1 systems - Tests, Results, questions, experiences

[~dave~~wave~]

I think the example set by the Abbey Road BD is the best case: simultaneous release of Atmos and properly mixed 5.1.

I think comparing the two mixes on a 5.1 system is very interesting and worthwhile.
I encourage everyone who's interested to take the time to just listen to the first minute of Here Comes The Sun, for example, and report what you perceive on YOUR 5.1 hardware with YOUR ears and YOUR brain.

Abbey Road, the single widely-owned example with high scores for both mixes anyone can easily compare without ripping.

 

[himey]

Their Atmos encoded disc.



When played back with decoding on an AVR, yes that’s my understanding.

I am having a problem understanding what the differences are in the files? It is probably something I am fundamentally missing in my reasoning, but if the processing is done in the hardware, depending on how many speakers you have in the settings, what differentiates these files between one another? Sorry if this is obvious but I can't wrap my head around it. Thanks.

 

[himey]

I think the example set by the Abbey Road BD is the best case: simultaneous release of Atmos and properly mixed 5.1.

Properly mixed?

I would be shocked if time was spent both mixing the Atmos AND the 5.1, separately. I was under the impression that the 5.1 mix, from Abbey Road, was derived from the Atmos mix, and outputted to 5.1 using the origional Atmos mixing software? Please correct me if I am wrong with that assumption.

 

[harync]

Properly mixed?

I would be shocked if time was spent both mixing the Atmos AND the 5.1, separately. I was under the impression that the 5.1 mix, from Abbey Road, was derived from the Atmos mix, and outputted to 5.1 using the origional Atmos mixing software? Please correct me if I am wrong with that assumption.

When I say properly mixed, I guess I should say properly encoded. In the course of trying out all the different DD+Atmos encodes, I can see where errors will unintentionally crop up if you play Atmos on a non-Atmos system.

I took Dave's suggestion above and listened to the Atmos and DTS versions of Here Comes the Sun on both systems. On the Atmos system, the Atmos version creates a "Dome of Sound" which is exactly how Atmos was conceived. DTS has discrete elements (Guitar LF, Keyboard LR, Strings RR).

On the non-Atmos system, the song essentially plays the same. The Atmos version has substantial overlap between the Fronts and the Rears, and no benefit of heights. Instead of "Dome of Sound" it's like extended stereo. The DTS is much more discrete.

I'll do some more testing and see if I can find another example of the unintended results of playing an Atmos track on a non-Atmos system.

 

[harync]

How an Atmos processor handles Atmos music, downmixed to 5.1 (or 7.1), vs a Dolby TrueHD processor, is an interesting comparison.

On paper it should be the same and the 3D speakers (Atmos) metadata are completely ignored?

I agree that on paper, the Atmos metadata should simply be ignored and you get 7.1. However, I think a lot of variability comes from how the mixer uses the Atmos software in terms of the Bed and objects. I'm pretty much a novice in terms of understanding how Atmos is mixed, but I can see where elements will not be discrete in the TrueHD core of an Atmos track depending on the choices of the mixer. In contrast, they might be discrete if the track was mixed in straight 7.1.

 

[himey]

I agree that on paper, the Atmos metadata should simply be ignored and you get 7.1. However, I think a lot of variability comes from how the mixer uses the Atmos software in terms of the Bed and objects. I'm pretty much a novice in terms of understanding how Atmos is mixed, but I can see where elements will not be discrete in the TrueHD core of an Atmos track depending on the choices of the mixer. In contrast, they might be discrete if the track was mixed in straight 7.1.

Separate but equally interesting conversation.

Hardware only conversation, does an Atmos processor/receiver ignore metadata if you set speakers to bed (4.0, 5.1. 7.1 ect.), non-3D only?

Human mixing element conversion, does the Atmos mixing engineer account for the bed as highly as the objects, especially when more bed only systems exist?

 

[bracelis]

Hardware only conversation, does an Atmos processor/receiver ignore metadata if you set speakers to bed (4.0, 5.1. 7.1 ect.), non-3D only?

As mentioned in a previous post, I tried it on a Denon/Marantz Atmos receiver, configured as 4.0 speakers and the receiver uses TrueHD decoding ( i.e. it behaves like a non-Atmos receiver with TrueHD decoding).

 

[harync]

Separate but equally interesting conversation.

Hardware only conversation, does an Atmos processor/receiver ignore metadata if you set speakers to bed (4.0, 5.1. 7.1 ect.), non-3D only?

Human mixing element conversion, does the Atmos mixing engineer account for the bed as highly as the objects, especially when more bed only systems exist?

As far as the latter point, I think that's really a music issue and 99% of the use cases are Dolby Atmos movie tracks. From my limited testing, ignoring the Atmos meta-data results in some bleed across the channels or putting something in the back only. For most films where any localized element (other than dialogue) will only last a few seconds it probably won't be noticed.

This really only becomes a problem with discrete surround audio mixes, which we can all unfortunately admit is a niche hobby. So I'm guessing that no, a typical mixing engineer for a movie soundtrack isn't worrying too much about what happens with different bed only systems ignoring meta-data. I think the Tidal Atmos offerings also demonstrate this point as they are quite inconsistent in their approach to beds and objects. However, Tidal's hardware requirement for an Atmos receiver does suggest Tidal is trying to make sure the overall experience is somewhat consistent and they're trying to avoid unintended consquences of non-Atmos playback.

 

[HomerJAU]

I am having a problem understanding what the differences are in the files? It is probably something I am fundamentally missing in my reasoning, but if the processing is done in the hardware, depending on how many speakers you have in the settings, what differentiates these files between one another?

Firstly, you are right that the final speaker output is created based on you AVRs speaker configuration. But it is based on the original mix. If a sound is authored to left back top and your AVR doesn’t have that speaker it’s routed to another nearest to the request position (decision is coded into the Atmos processor). So the priority is the mix position, actual output is determined on speaker configuration.

Those Atmos test files came from different Titles on the disc. Each Title has a different Atmos mix (and different video if you watch the video).

 

[HomerJAU]

Hardware only conversation, does an Atmos processor/receiver ignore metadata if you set speakers to bed (4.0, 5.1. 7.1 ect.), non-3D only?

I‘d say: It does not ignore metadata. EDIT: But the metadata objects get rerouted to the bed.

I don’t know if this is correct but there’s no reason the decoder shouldn’t or couldn’t do this since that’s how Atmos decoders work for all other speaker configurations.

 

[HomerJAU]

I believe most players will not recognize a TrueHD codec in a MP4 container -- you will need to use mt2s.

The original m2ts files are huge. Up to 1GB each as each channel signal in each file runs for 1 minute.

 

[AYanguas]

I remember reading about how a Dolby Atmos track is encoded. I will try to explain how it works:

• First, the track is a Dolby TrueHD (or Dolby Digital Plus) multichannel track. (5.1 or 7.1) (You could convert it to MCH FLAC, for instance).
  • With a receiver that decodes Dolby TrueHD, either because it has No Atmos functionality or because you select a Sound Option that “ignores” Dolby Atmos, YOU GET ALL CONTENT mixed in the multichannel track.
  • In some cases, even a core AC3 could be embedded into the track to be backwards compatible with old decoders.

• Then, for an encoded Dolby Atmos track, metadata is “added” to the 7.1 track, that describes the content and the location coordinates of each “object”.

• During decoding of the Atmos Track, the content of the “sound objects” is "subtracted" from the floor channels/speakers and “added” to the corresponding speakers, either floor speakers or height speakers, to generate the “sound object” from the location mixed by the mixer engineer.
  • The Atmos decoder decides in real time what speakers to use, depending on the AVR configured speakers, to generate the better possible image of the location of the sound object.
  • If the decoder does not support Atmos, simply the metadata is ignored and the full multichannel track is reproduced.

In summary, NO CONTENT is never Lost from a Dolby Atmos track when played by a decoder that does NOT support Atmos.

Someone says that then, the Non-Atmos decoder “downmix” to the available speakers.

But as far as I understand it is the opposite. You have a “core” multichannel 2.0/5.1/7.1 as the basis of the track with ALL content in it. It is the Atmos decoder who “UPMIX” with the metadata the object sounds to the available speaker configuration.

So, when a mixer artist produces a Dolby Atmos track, the final render production process already generates multichannel 5.1/7.1 track that contains all sound, apart from the metadata that contains the “objects”. The mixer engineer can assign location coordinates to an object or can assign directly a “bed channel” or speaker to that object. The terminology "bed channel" (or speaker) is a 'physical channel/speaker' either on floor or height. Some confusion exists thinking that bed channels are only floor channels as opposed to heights.

 

[ar surround]

I think comparing the two mixes on a 5.1 system is very interesting and worthwhile.
I encourage everyone who's interested to take the time to just listen to the first minute of Here Comes The Sun, for example, and report what you perceive on YOUR 5.1 hardware with YOUR ears and YOUR brain.

Abbey Road, the single widely-owned example with high scores for both mixes anyone can easily compare without ripping.

When I say properly mixed, I guess I should say properly encoded. In the course of trying out all the different DD+Atmos encodes, I can see where errors will unintentionally crop up if you play Atmos on a non-Atmos system.

I took Dave's suggestion above and listened to the Atmos and DTS versions of Here Comes the Sun on both systems. On the Atmos system, the Atmos version creates a "Dome of Sound" which is exactly how Atmos was conceived. DTS has discrete elements (Guitar LF, Keyboard LR, Strings RR).

On the non-Atmos system, the song essentially plays the same. The Atmos version has substantial overlap between the Fronts and the Rears, and no benefit of heights. Instead of "Dome of Sound" it's like extended stereo. The DTS is much more discrete.

I'll do some more testing and see if I can find another example of the unintended results of playing an Atmos track on a non-Atmos system.

As I mentioned in an earlier post in this thread, when I play the ATMOS mix in 5.1 through my Oppo 205 directly from the disc, I hear the same mix as the 5.1 DTS MA stream...including the same degree of "discreteness." However, when I rip the ATMOS mix from Make MKV using DVD AudioExtractor, it shows up on the list as a 7.1 True HD stream. But I have been unable as of yet to get the Oppo to downmix the 7.1 rip properly into 5.1...It sounds like extended stereo as harync describes.

 

[himey]

Firstly, you are right that the final speaker output is created based on you AVRs speaker configuration. But it is based on the original mix. If a sound is authored to left back top and your AVR doesn’t have that speaker it’s routed to another nearest to the request position (decision is coded into the Atmos processor). So the priority is the mix position, actual output is determined on speaker configuration.

Those Atmos test files came from different Titles on the disc. Each Title has a different Atmos mix (and different video if you watch the video).

That isn't what I am asking. I understand that you ripped these from the disc and they are different titles.

 

[harync]

So, when a mixer artist produces a Dolby Atmos track, the final render production process already generates multichannel 5.1/7.1 track that contains all sound, apart from the metadata that contains the “objects”. The mixer engineer can assign location coordinates to an object or can assign directly a “bed channel” or speaker to that object. The terminology "bed channel" (or speaker) is a 'physical channel/speaker' either on floor or height. Some confusion exists thinking that bed channels are only floor channels as opposed to heights.

The R.E.M. track "Sweetness Follows" provides an a useful example. The Atmos BD Version has a cello isolated in the 4 Height speakers in my Atmos system (with it sounding loudest in the Left Front Height). When I play it in my non-Atmos system, the cello is in the Left Front, Left Surround, and Right Surround. So while the cello is indeed in the TrueHD track (no information is missing), the intention of the mixer to have the cello be isolated in the heights in the center of the room is lost and it is instead distributed across the room in 3 of the 5 channels.

TL,DR version: No musical information is lost playing an Atmos track without an Atmos decoder, just "Artistic Intent."

 

[marpow]

I am just a beginner to ATMOS and my experience is very little but decided I would start paying attention to this thread to learn some more. My goal as of this writing is to experience ATMOS in all the way it comes.
My first thing was just to do a manual test tone on all my 5.1 and 4 celing (height) speakers, relieved, everything works and takes a signal.
My first was INXS Kick, very immersive.
I am watching via Fire Cube Disney+ Mandolarian, ATMOS there but a little weak.
I bought 4 movies, based on top 5 You Tube reviews, need to watch them. Seems like war movies get a lot of votes.
Basic things I am learning. This is based on watching/listening to hardly anything.
First off, I am very excited, I think this is going to be great.
I have 3 amps, 2 channel for Front L & R main, 3 channel for Center rear R & L, and 4 channel for four height speakers. With 3 subs, so the floor speakers are all set to large.
The 3 amps have VU meters so I can watch the levels.
The Atmos levels are definately not as strong as the floor levels.
As with a lot of surround listening it is no suprise that listening to really get things active needs to be 55-60db+.
As I am new, I am not sure what experience I am supposed to have, so while listening to INXS KICK I turned off all the amps except the 4 channel height speaker amp and the music was coming out of all 4 speakers equally. I figured this was a complete immersive mix, which is great.
Thanks for letting me be a newcomer here, still waiting to hear rain, bullits, and helicopters, and of course weird noises coming out of music.
Going to try one of my movies and Tidal Atmos this week.

 

[Miker1102]

I remember reading about how a Dolby Atmos track is encoded. I will try to explain how it works:

• First, the track is a Dolby TrueHD (or Dolby Digital Plus) multichannel track. (5.1 or 7.1) (You could convert it to MCH FLAC, for instance).
  • With a receiver that decodes Dolby TrueHD, either because it has No Atmos functionality or because you select a Sound Option that “ignores” Dolby Atmos, YOU GET ALL CONTENT mixed in the multichannel track.
  • In some cases, even a core AC3 could be embedded into the track to be backwards compatible with old decoders.

• Then, for an encoded Dolby Atmos track, metadata is “added” to the 7.1 track, that describes the content and the location coordinates of each “object”.

• During decoding of the Atmos Track, the content of the “sound objects” is "subtracted" from the floor channels/speakers and “added” to the corresponding speakers, either floor speakers or height speakers, to generate the “sound object” from the location mixed by the mixer engineer.
  • The Atmos decoder decides in real time what speakers to use, depending on the AVR configured speakers, to generate the better possible image of the location of the sound object.
  • If the decoder does not support Atmos, simply the metadata is ignored and the full multichannel track is reproduced.

In summary, NO CONTENT is never Lost from a Dolby Atmos track when played by a decoder that does NOT support Atmos.

Someone says that then, the Non-Atmos decoder “downmix” to the available speakers.

But as far as I understand it is the opposite. You have a “core” multichannel 2.0/5.1/7.1 as the basis of the track with ALL content in it. It is the Atmos decoder who “UPMIX” with the metadata the object sounds to the available speaker configuration.

So, when a mixer artist produces a Dolby Atmos track, the final render production process already generates multichannel 5.1/7.1 track that contains all sound, apart from the metadata that contains the “objects”. The mixer engineer can assign location coordinates to an object or can assign directly a “bed channel” or speaker to that object. The terminology "bed channel" (or speaker) is a 'physical channel/speaker' either on floor or height. Some confusion exists thinking that bed channels are only floor channels as opposed to heights.

Great explanatio that makes sense.

 

[junh1024]

Hopefully we’ll get a definitive answer to what happens when users playback Atmos and TrueHD 7.1 mixes on non Atmos/7.1 surround systems.

I recommend you to link to Aya's informative comment Atmos and TrueHD 7.1 playback on 5.1 systems - Tests, Results, questions, experiences in your initial comment.

...wholesale down mixes to a traditional 5.1 system, you lose the intent and control the mix engineer had in mind for that music, for not only the Atmos mix, but what the 5.1 mix would have been also. This is also mainly addressing the fact (hopefully not a trend) of the Goats Head Soup release, not providing a “dedicated” 5.1 DTS-HD Master Audio mix.

Playing back Atmos on a 51 system will obviously sound different. And providing a "dedicated" 51 mix isn't going to solve the problem cuz let's be realistic, studios aren't going to make another mix at their expen$$$e (unless there was one done say, 10y ago), and so any "dedicated" 51 mix which is beside a Atmos mix is likely just an additional render. Downmixing Atmos 71x to 51 sounds a little different on a new vs old AVR. so providing a "dedicated" 51 mix is OK, all it does is close the gap.

If I'm understanding Pup correctly, though, his basic point is that mixing for Atmos is different from mixing for 5.1. So the 5.1 TrueHD "core" .... may sound perfectly fine... but not "intentionally."

Most studios wont make a different 51 mix (as demonstrated by Beatles - here comes the sun 51 vs Atmos in previous comments), but It's possible to preserve artistic intent while downmixing to 2D for a Atmos mix, by taking into account 51 downmix behavior while you mix. Ideally, you'll also monitor on a legacy 51 AVR. You now have a Atmos mix that still sounds great in 51. But most Atmos mixes are not mixed like this.

Not true. Avid Pro Tools | Ultimate will encode/decode Dolby Atmos, provided you have an interface with sufficient I/O. I'm sure there are other professional DAWs/NLEs that will do the same. Otherwise, how would Atmos even be mixed in the first place?

PT, like a few other DAWs can mix/encode/decode in Atmos, but ONLY the professional formats. They can't encode or decode consumer deliveables like THD & EC3. YOu need to purchase more stuff.

For example, this one shows Mac only:
https://www.avid.com/plugins/dolby-atmos-production-suite

Many Atmos stuffs are on mac prolly due to mac's better audio architecture.

However, Tidal's hardware requirement for an Atmos receiver does suggest Tidal is trying to make sure the overall experience is somewhat consistent and they're trying to avoid unintended consquences of non-Atmos playback.

Yes, but it's not entirely preventing playback on 51, as you can insert an intermediate media box that fakes Atmos ability, and budget 51 AVRs also to Atmos nowdays. Not really sure if it's to do with money too. Maybe both.

. The Atmos BD Version has a cello isolated in the 4 Height speakers in my Atmos system (with it sounding loudest in the Left Front Height).

I'd say this is a questionable decision.

 

[HomerJAU]

I recommend you to link to Aya's informative comment Atmos and TrueHD 7.1 playback on 5.1 systems - Tests, Results, questions, experiences in your initial comment.

Good idea. Now done.

 

[harync]

Yes, but it's not entirely preventing playback on 51, as you can insert an intermediate media box that fakes Atmos ability, and budget 51 AVRs also to Atmos nowdays. Not really sure if it's to do with money too. Maybe both.

I attached my Firestick to my Atmos AVR and signed in to Tidal, and then plugged it into my non-Atmos AVR and it worked for a while, so Tidal only periodically checks. My guess it was a Dolby requirement, partly to drive hardware sales and partly to ensure mixes were being heard as intended.

[Regarding isolating sounds in the ceiling in REM's Automatic for the people]
I'd say this is a questionable decision.

This was Steve Genewick's first-ever Atmos mix. Since @Steve G is a member of this forum, it would interesting to hear from him if he would change anything if he had to do it over. Also it would be nice to know in general how he approaches Atmos mixes knowing that non-Atmos AVRs will certainly be playing back those mixes.