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

QuadraphonicQuad

Help Support QuadraphonicQuad:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
There seems to be much confusion in this thread on how Atmos works and what you get on 5.1 systems. While AYanguas is correct. I feel like the implications haven't sunk in.

Let me address something basic he talked about. ALL Atmos recordings have either a 7.1 TrueHD base track (100% backwards compatible with older decoders because it IS a TrueHD track. NOTHING is different about it at all) or a 5.1 or 7.1 DD+ base track.

The mixing engineer is most certainly going to listen to the base track playback at some point or numerous points. They'd have to be an idiot not to. They can easily do this at any time because all Atmos setups are still based on 5.1/7.1 are layouts! I can easily switch to 7.1 or 5.1 playback on my Marantz receiver either by selecting the TrueHD base layer in surround mode (which I can then send to DTS Neural X to upmix instead if I prefer) or I could tell the receiver I have a basic 5.1 or 7.1 layout and it will handle it automatically.

The base 5.1 or 7.1 mix has everything placed where the mixing engineer wants it! He can place sounds in completely different locations if he desires it because the meta layer (that older AVRs can't see) contains all the sounds associated with objects. The way the Atmos mix works is rather incredible. It actually phase cancels out the same overlapping sound in the base soundtrack!

That's why you don't get duplicate sounds and the meta layer doesn't have to carry a complete separate soundtrack! It only cancels out objects. It can then render them anywhere in the room including rear surrounds (7.1) even when the base layer is limited to only 5.1!

There are DD+ 5.1 based Dolby Atmos demos that do precisely this while the exact same demo on one of their demo Blurays is TrueHD 7.1 based and they render exactly the same on an Atmos receiver (tested thoroughly here).

ATMOS is a room based format in the sense that sounds are supposed to come from the same place for all listeners. In practice, this may not happen with lower speaker count setups which is why Atmos at the theater has even more speakers employed (up to 64 speakers in the first generation units with support for 128 channels/objects internally).

This differs from a traditional 5.1 or 7.1 movie theater setup which might employ dozens of speakers, but the are arrayed copies of those same 5.1 or 7.1 channels. This makes the output user orientated (i.e. The "left side surround" in a 7.1 system will seem to come from the surround speaker closest to the listener, not one specific speaker in the room like Atmos can do.

The Atmos presentation therefore will be different for you when you sit in different places in the theater (i.e. A sound might be behind one person, but in front of another because it's stationary in the room). Thus, the notion of "one correct intended mix" is erroneous in the sense you think everyone is meant to hear the same thing. No. It's an adventure of sorts. The "intended" locations are specified and your specific system does the best it can with what you've got to represent it. That is by all means intentional by the mixing engineer or he wouldn't be using Atmos in the first place.

If you go see a concert live, doesn't the presentation change depending on what seats you get for the concert? Some seats might be considered better than others,but it's also subjective. Atmos attempts to capture the soundfield for the arena and you can play it back more or less accurately with your room/setup needs and where you sit in it. But that's by design. Similarly, the base 5.1 or 7.1 soundtrack attached is also b design. If it sucks, it's a bad mixing engineer's job/vision/intent not Atmos' fault.

Yes, you can include a separate 5.1 or 7.1 (or quad or whatever) mix like Steve Wilson did for his latest album where he did the 5.1 mix alone, but had assistance with the Atmos mix. So they sound different on a 5.1 system because they are different mixes.

You can even view the Atmos object placements on a high end Trinnov processor. It shows a rectangular room/grid with how many speakers you're using up to 32.1 and the objects appear as moving balls while the speakers light up when either the base layer or an object renders to it (including in part for pans).

Atmos is capable of rendering a mixing engineer's vision for whatever level his own mixing station can represent including the base 5.1 or 7.1 mix. Most don't have all 32.1 speakers to try out, buy then most of those speakers outside the 7.1.4 base rectangle exist for anchoring sounds for more rows of seats and off-axis listeners in general and thus more precision, not really "new locations" of sound as the all fit into divisions within that base 7.1.4 grid.

Yes, having less than 7.1.4 (e.g. 5.1.2, 5.1.4 and 7.1.2 are common) means some sounds will get moved, but it's inherently no different than moving rear surrounds on 7.1 system to general surrounds on 5.1 system. If you want more accuracy or for more seats ir larger rooms to sound as good, you may need to upgrade to get there.

I know this is long, but hopefully it clears some things up about Dolby Atmos.
 
The base 5.1 or 7.1 mix has everything placed where the mixing engineer wants it!

This is only partially true. Let's take the case of Atmos on a legacy 51 AVR (or maybe even 71 on 51). According to tests,

L & HL goes into L , BL & HBL goes into BL as expected.

But Side L & Height Side L goes 100% into BL. Not intuitive.

If you're having a lot of content in the sides, that will get moved rearwards. A Newer Atmos AVR will put the sides slightly more forwards, but still majority back-biased.

Yes, having less than 7.1.4 (e.g. 5.1.2, 5.1.4 and 7.1.2 are common) means some sounds will get moved

Also, if you have a 512/712 system & someone did a circular pan @ 100% height, it will get downmixed 100% into HSL & HSR, transforming a circular movement into a LR movement, and destroying artistic intent. The only way to avoid this is to pan at partial height so you get some height depth. Ironically, it will sound OK on 2D 51 & 71 systems.

The mixing engineer is most certainly going to listen to the base track playback at some point or numerous points. They'd have to be an idiot not to. ...I can easily switch to 7.1 or 5.1 playback on my Marantz receiver

Mixers can also switch different layouts in the Dolby renderer but it's not the same as rearranging your speakers & using a legacy decoder.

I don't think all mixers switch layout or audition on legacy system (or they think it sounds OK), otherwise there would not be mixes coming out which ONLY sound "OK" on 714+ systems . Like with lots of drums in the sides which gets downmixed to the back which will sound bad. I can PM an example.
 
Last edited:
This is only partially true. Let's take the case of Atmos on a legacy 51 AVR (or maybe even 71 on 51). According to tests,

L & HL goes into L , BL & HBL goes into BL as expected.

But Side L & Height Side L goes 100% into BL. Not intuitive.

That is caused by your surround speaker placement, not Dolby Atmos. Side surrounds are synonymous with "surrounds" in 5.1, not rear surrounds and that applies to all downmixing of 7.1 to 5.1, not just Dolby Atmos. Dolby offers a range of suggested placement for surround speakers and you cannot be sure every mixing studio has the same exact setup or placement either. But even the surround speakers recommendations have changed with Atmos (old school had them 1/3 or better 2/3 up the wall while at home they're expected to be at or just above ear level now). Yes, that can create some dichotomy between old school 5.1/6.1/7.1 soundtracks and Atmos systems because something like that jets in Top Gun might now appear at ear level (although at 1/3 in the old home system, they'd be even lower) instead of somewhat overhead. That wasn't fully fixed for Dolby modes until the Atmos version was released, but with DTS Neural X, it would move the jets overhead for you and even higher, really creating a better mix than the old Dolby system could possibly do so it's a mixed bag in some respects. But then 5.1 -> 7.1 is far more limited overall than Dolby Atmos or DTS:X.

Either way, what you're complaining about can be fixed by moving your speakers forward a bit or upgrading to 7.1 or greater. I don't see the issue here seeing as I have my side surrounds at about 105 degrees with a 11.1.6 system. There's no requirement for me to put them at 90 degrees and with front wides, I get smooth panning through the area regardless and then 5.1 sounds "correct" as you say. I suppose by your reckoning, Atmos would sound "wrong" here, but that's not my impression. I wouldn't presume to think I know where some instrument should be located to the "side" anyway. It sounds more like you like drums in a certain location and if they move back a bit, it sounds "weird" to you. But if I move to my 2nd row, the drums are going to be in front of me instead of behind me in my setup. As I said, Atmos is room-centric, not listener-centric. My rear speakers at about 11 feet behind my MLP in the front row! That's not typical for a smaller room or setup. But then some people have their couches against the back wall and side surrounds at 90 degrees. Unless you're going to assume a perfect quad system (center of "X"), which is really only good or one or two people at most anyway, you can't really make precise assumptions about 5.1 layouts, IMO.

I see you're also still using backwards cause/effect language as well.

L & HL don't actually go into ANYTHING! They're moved out of the base layer, not into it. There is no actual "downmixing" in that sense. The base track is what it is whether you like the placements or not. Height layers, meanwhile are moved "out of" the base layer by phase cancelling out the sounds there and re-rendering them in a different location based on the meta data. That meta data layer only contains sounds for the objects in the mix. True base layer sounds that stay there aren't in the meta data layer as they aren't touched. The renderer "re-renders" the new location after it's canceled out the old one. Thus, heights are created out of the base track, not moved into it. Whatever the mixing engineer decided to put in the base track is what's in it. It's the height layer and other rendering positions that are created in Atmos from the base track, not the other way around. So nothing is moved to the "back surround" (save 7.1 down into 5.1, which is not an Atmos function, but what all AVRs do with 7.1 soundtracks which is to "fold" them into the 5.1 soundtrack into the surrounds). It's already there. Things are moved out of the base track and placed elsewhere.

If you're having a lot of content in the sides, that will get moved rearwards. A Newer Atmos AVR will put the sides slightly more forwards, but still majority back-biased.

YOU place the side speakers, not the AVR. This "back biased" notion isn't true either. Most people complain that Atmos soundtracks barely use the rear speakers in most soundtracks. I was pleasantly surprised to hear Booka Shade and Yello make heavy use of the rear surrounds as most Atmos soundtracks either just duplicate side surround effects (like arrays of old) or put very little discrete in the rear surrounds (a few movie exceptions). But that's the mixing engineer's fault, not the system. Even with my side surrounds at 105 degrees, I find most music surround albums in 5.1 only have most of the sounds to the sides or just behind me when played in a direct mode (letting it upmix to 7.1 brings some things further back, but that can be turned off if it's not desired).

Also, if you have a 512/712 system & someone did a circular pan @ 100% height, it will get downmixed 100% into HSL & HSR, transforming a circular movement into a LR movement, and destroying artistic intent. The only way to avoid this is to pan at partial height so you get some height depth. Ironically, it will sound OK on 2D 51 & 71 systems.

I'm not sure what your point here is other than if you want the "best possible" Atmos rendition, you better buy a Trinnov and a 32.1 speaker layout.

I've already said 7.1.4 is the "base rectangle" for an Atmos system that doesn't get any compromised rendered locations. The gist is, don't buy anything less than 7.1.4 if you're worried about missing something placed where it's meant to be (many people are very happy with 5.1.2 or 5.1.4 or 7.1.2). The mixing engineer probably mixed on a 7.1.4 or perhaps a 9.1.6 setup at most. It is the Atmos renderer that places the sounds to those extra speakers, not directly by the mixing engineer. He places OBJECTS. The renderer places the sounds to the nearest available speakers with exceptions for 100% height layer (which stay in the height layer).

Those demos you refer to are 100% ceiling sounds. If you had a sound that was part-way between the ceiling and bed layer, it would not stay in the height layer. These are mixing decisions by the mixing engineer (see "snap to" above). No single demo should be taken as gospel for how things are moved.

Mixers can also switch different layouts in the Dolby renderer but it's not the same as rearranging your speakers & using a legacy decoder.

How do you know the exact studio mixing layout of every surround album you own? You don't. It sounds to me like you're used to hearing various albums a certain way and if you play them on a different system, it changes somewhat and you're not happy about it. Since Atmos is room-centric rather than listener-centric, where you sit affects the presentation just like where you sit in a live concert affects it. That's a strength, IMO, not a weakness. I can listen to Atmos surround albums in different rows here and get a very different perspective. It's not wrong. It's different.

I don't think all mixers switch layout or audition on legacy system (or they think it sounds OK), otherwise there would not be mixes coming out which ONLY sound "OK" on 714+ systems . Like with lots of drums in the sides which gets downmixed to the back which will sound bad. I can PM an example.

Maybe you could move your speakers if you don't like how it renders. I can move the "phantom image" of the channel using my active mixer mod (between arrayed front wide and side and ss#2 and rear; they're all fully adjustable without moving the actual speaker). In other words, any problem can be fixed with a bit of will. There are speaker selectors out there that will let you have more than one setup in the room using one receiver (e.g. Auro-3D and Atmos or Atmos and old school 5.1 with the surrounds 2/3 up the side wall).
 
I wonder if some discussions here, with different opinions (or doubts) could be solved if we knew how the Atmos encoding really works.

I am speculating: I can imagine two different scenarios:
  1. The mixer engineer mixes first the different sounds into a 5.1, 7.1 base layer. Thus, the sound locations on a 5.1, 7.1 playing are already determined by the mixer. Then in a second phase, he develops the sound objects and its coordinates, and then generates the encoded final Atmos with the metadata. Sound objects comes from sounds already existing on the base 5.1, 7.1 core layer.
  2. The mixer engineer goes directly by defining Atmos sound objects and assigning them its coordinates. Then the Atmos encoding generates an Atmos full track that include all sounds and objects metadata, and the encoding process also generates the base core 5.1, 7.1 by assigning locations based on the proximity of the 3D coordinates going into 2D coordinates of fixed 5.1,7.1 channels.
I do not know the Atmos production internals, but I guess the scenario (2) would be the correct one.

So, the base layer 5.1, 7.1 embedded in an Atmos track for retro compatibility playback would be automatically generated by the Atmos encoding process and based of best location proximity.

If this automatic approach is not to the liking of the mixer artist, then he could mix another 5.1 apart from the Atmos mix, with his 2D surround artistic intent. That could be the reason we find those two different mixes in albums like the Steven Wilson TFB.

Perhaps, the tools available for the Atmos production are sophisticated enough to give many options to the mixer engineer that could tune many settings. and they could allow the both two scenarios I describe above.

Anyone with a good knowledge of Atmos production tools is welcome…
 
I wonder if some discussions here, with different opinions (or doubts) could be solved if we knew how the Atmos encoding really works.

I am speculating: I can imagine two different scenarios:
  1. The mixer engineer mixes first the different sounds into a 5.1, 7.1 base layer. Thus, the sound locations on a 5.1, 7.1 playing are already determined by the mixer. Then in a second phase, he develops the sound objects and its coordinates, and then generates the encoded final Atmos with the metadata. Sound objects comes from sounds already existing on the base 5.1, 7.1 core layer.
  2. The mixer engineer goes directly by defining Atmos sound objects and assigning them its coordinates. Then the Atmos encoding generates an Atmos full track that include all sounds and objects metadata, and the encoding process also generates the base core 5.1, 7.1 by assigning locations based on the proximity of the 3D coordinates going into 2D coordinates of fixed 5.1,7.1 channels.
I do not know the Atmos production internals, but I guess the scenario (2) would be the correct one.

So, the base layer 5.1, 7.1 embedded in an Atmos track for retro compatibility playback would be automatically generated by the Atmos encoding process and based of best location proximity.

If this automatic approach is not to the liking of the mixer artist, then he could mix another 5.1 apart from the Atmos mix, with his 2D surround artistic intent. That could be the reason we find those two different mixes in albums like the Steven Wilson TFB.

Perhaps, the tools available for the Atmos production are sophisticated enough to give many options to the mixer engineer that could tune many settings. and they could allow the both two scenarios I describe above.

Anyone with a good knowledge of Atmos production tools is welcome…

I'm guessing the 2nd also, but somehow I imagine it can be adjusted afterwards as well, quite possibly only one the base 5.1 or 7.1 level. There's no technical reason why it couldn't be adjusted for the base layer separately since the meta data moves it out of that layer to technically anywhere in the room. The mixing engineer should certainly be at least happy with the base layer results before passing it on. The point is you can't gauge the capability of the system by a few speaker test Atmos demos.
 
I can listen to Atmos surround albums in different rows here and get a very different perspective. It's not wrong. It's different.

I guess I could move myself or the speakers, but a bit tricky with a small space.

I do not know the Atmos production internals, but I guess the scenario (2) would be the correct one.

Dolby would prefer you to do a Atmos-first approach, but the 2 approaches can sound very similar in practice.

I had a look at some docs, and a few more things:

So you can make a separate 51 render from Atmos. This will sound a bit better than a AVR downmix. You can also:

3- bind the render to the THD track so that it automatically plays on a 51 (legacy?) system. This feature is available dating back to 2000s THD days, but I've never seen it used.
4- Have a separate 51 render track on the BD, so that consumers are happy that a "separate" mix is provided.

There's also a few 51 downmix modes available.

5- Alternatively, there's "zone masking" in PT which can subset speakers while mixing. An advanced concept, which I don't think people are using.


There's no technical reason why it couldn't be adjusted for the base layer separately since the meta data moves it out of that layer to technically anywhere in the room.

Well, technically it could, but it would be a management headache for humans managing 2 different mixes using 1 infrastructure (and maybe a technical headache too). Might as well use the render (or make a different mix) & bind .

The tools described above could make a logically derived & acceptable 51 downmix, rather than a completely different mix. Hopefully with "zone masking", or a dedicated 51 render can make it sound OK on 51 systems. Or just mix taking into downmix law as I said in my post previously.
 
Last edited:
I haven't had a chance to read everyone's post in this thread, but I have a couple of questions that I'm hoping someone can answer.

I have a Samsung HW-Q950T sound bar system that is a 9.1.4 system which I want to test levels and a bit of EQ'ing on each channel separately (particularly the Atmos channels, which up to now I haven't been able to isolate).

1. Which is the best format to use - mkv or mp4?

2. How do you test 9.1.4? None of the files are specifically labelled as 9.1.4.

My DVD player will be bitstreaming and the decoding will be done in the sound bar.
 
I have a Samsung HW-Q950T sound bar system that is a 9.1.4 system which I want to test

1. Which is the best format to use - mkv or mp4?

2. How do you test 9.1.4? None of the files are specifically labelled as 9.1.4.

My DVD player will be bitstreaming and the decoding will be done in the sound bar.

Try both mp4 & mkv 916 tests, both should be the same. As long as your DVDP supports the files & can bitstream it should be fine. BDPs should fare better on these formats.
 
Have a problem.
My Blu-ray DVD player accepts the mkv files, but says that it can't use the audio.
And the mp4 files are not recognised at all.
Not sure what to do now as I thought these files would give me exactly what I wanted.
BTW, the video part of the mkv files work, but with no sound. :(
 
BTW, the video part of the mkv files work, but with no sound

How is your BD player connected to your system?

You’ll need an AVR with an Atmos decoder, connected via HDMI and configured for HDMI pass-through (Bitstream).

Your new Panasonic player does not mention Atmos or TrueHD capability in its specs, which is strange, but does mention 7.1 analog outputs and FLAC, WAV, DSD etc. (but no mention of Atmos). Maybe your manual will clarify
 
How is your BD player connected to your system?

You’ll need an AVR with an Atmos decoder, connected via HDMI and configured for HDMI pass-through (Bitstream).

Your new Panasonic player does not mention Atmos or TrueHD capability in its specs, which is strange, but does mention 7.1 analog outputs and FLAC, WAV, DSD etc. (but no mention of Atmos). Maybe your manual will clarify
The Panasonic is set to Bitstream and is connected to the Samsung HW-Q950T soundbar with a new high speed (18gbps) HDMI cable.
The soundbar is (supposedly) capable of decoding and playing Atmos tracks.

However, I am still coming to terms with the soundbar.

Using REW and ASIO drivers, I am trying to level and EQ the various channels (the soundbar has level adjustments and basic equalisers on each channel).
But testing has proved that the equaliser settings have absolutely NO effect on the SPL graphs. Not even the Treble and Bass adjustments have any effect.
Have talked to someone at Samsung Australia who had no answers.
Have latest firmware installed.
This is rather disappointing has I am moving from a 7.1 dedicated home theatre to a small room and a soundbar. :cry:
 
I have tested the 5.1.2 and 7.1.4 files on my 5.0.2 system, where the 2 height speakers are set up as Top Middle on my Denon receiver.

5.1.2 on a 5.0.2 system.
I can hear all channels.
The LFE is redirected to the front speakers which are full range.

7.1.4 on a 5.0.2 system.
I can hear all channels.
The LFE is redirected to the front speakers which are full range.
Side speakers are split between front and rear speakers. Side left therefore plays front left and rear left creating a phantom side left channel.
Height speakers are downmixed to 'stereo', if I can use that expression. Therefore, left height front and rear are played in the left top middle speaker, while right height front and rear are played in the right top middle speaker.

Just for completeness, these are the dimensions and details of my system (I've included the Dolby specs, the Denon Manual and a sketched plan/section of the room):

1647329077495.png
 
Thanks Alberto.

From your findings I’d think users going for X.1.2 systems should install the 2 heights in the Top Middle position if going with only 2 heights. Quad heights is still my recommend system if it can be done (5.1.4 or 7.1.4).
 
Thanks Alberto.

From your findings I’d think users going for X.1.2 systems should install the 2 heights in the Top Middle position if going with only 2 heights. Quad heights is still my recommend system if it can be done (5.1.4 or 7.1.4).
Yes, that would be the best option if there is no space behind the listening position. The dolby specs and the AVR manuals should provide enough info about the angle you are supposed to use to install the speakers as top middle. I'm happy with how my system sounds in Atmos and it is an improvement over 5.1 for sure.
 
Cancellation:

With any type of downmixing (2.0 > 1.0 thru 7.1.4 > 5.1 for example), some sound cancellation is probably inevitable, are there any cases where there's an obvious loss in an Atmos to 5.1 downmix?


Kirk Bayne
 
Cancellation:

With any type of downmixing (2.0 > 1.0 thru 7.1.4 > 5.1 for example), some sound cancellation is probably inevitable, are there any cases where there's an obvious loss in an Atmos to 5.1 downmix?


Kirk Bayne
As far as I understand how Atmos is coded into Dolby TrueHD or DD+, there is a 5.1 "layer" that will play just as 5.1 mix. The Atmos decoder will use the embedded object metada to "substract" the object sound from the base channels and "add" it to the immersive Atmos channel, either "floor" or "height".

So there is no really "downmixing" from Atmos to 5.1, but playing the underlaying 5.1 mix track.
 
In recognition that many many people with surround systems only have 5.1 capability, wouldn’t the mixers who are doing an Atmos-only mix take this into account when they do a mix? Kind of like when there were stereo and mono mixes they would fold down the stereo to make sure it sounded ok on mono systems?
 
I may be wrong, but I understood Atmos to be on a 7.1 "bed", not 5.1. Your point, regardless, remains.
Yes.
Atmos in TrueHD is 7.1 (8 channels).
But it use to have also an inner layer in 5.1 (6 channels). You can 'see' it when you load a TrueHD Atmos track on MKVToolNix and get two audio tracks 7.1 (8 channels) and 5.1 (6 channels).
When played on 5.1 system, the AVR could play just that inner 5.1 layer or mix/downmix the sides and rears if the 5.1 layer is not implemented in the file.

Atmos in DD+ is 5.1
 
Back
Top