Technical explanation of Atmos and DTS:X?

[Reverend Slim]

How is it not? It sounds like it’s turning YPAO and other room correction algorithms on it’s head, in a way removing a processing layer.

Because it has nothing to do with room correction. It is just an object-based format for steering audio rather than the channel-based analog panning of formats before. Absolutely nothing related to processing sound or correcting your room's acoustics. You would still ideally use room correction to get better cross-channel placement and panning with Atmos.

 

[FriedChicken]

Because it has nothing to do with room correction. It is just an object-based format for steering audio rather than the channel-based analog panning of formats before. Absolutely nothing related to processing sound or correcting your room's acoustics. You would still ideally use room correction to get better cross-channel placement and panning with Atmos.

Room correction steers audio based on location of speakers.

The old paradigm is to match the speaker location to the ideal, and then deliver the ideal. It sounds like object sounds does away with the ideal and steers the audio.

 

[FriedChicken]

Please can you provide a link if you have found such?

Unfortunately This is from a long past search... a quick search should make it available though? Maybe try Dolby Labs or whoever makes DTS

 

[Reverend Slim]

Room correction steers audio based on location of speakers.

The old paradigm is to match the speaker location to the ideal, and then deliver the ideal. It sounds like object sounds does away with the ideal and steers the audio.

No, room correction doesn't steer anything. Room correction corrects the measured frequency response of a speaker in your room to make it align to a defined target curve. It has nothing whatsoever to do with steering of audio, just equalization.

And you still have to match the speaker location to the ideal, because Atmos has predefined speaker locations based on their expectation in the renderer. What you're conflating room correction with is speaker remapping, which is only done on higher end processors (though some of that may be making its way into more consumer gear soon).

 

[FriedChicken]

And you still have to match the speaker location to the ideal, because Atmos has predefined speaker locations based on their expectation in the renderer. What you're conflating room correction with is speaker remapping, which is only done on higher end processors (though some of that may be making its way into more consumer gear soon).

Oh I gotcha. Yes, speaker remapping.

It sounds to me like Atmos is doing that, or something like that.

 

[Reverend Slim]

Oh I gotcha. Yes, speaker remapping.

It sounds to me like Atmos is doing that, or something like that.

Nope. Atmos doesn't do any speaker remapping; it places sounds based on cartesian coordinates, with certain predefined locations for some objects to represent channels within an object-based context. Here's the logic:
1. The forwardmost speakers represent the 0.0 longitudinal range, and the rearmost speakers represent the 1.0 longitudinal range.
2. The leftmost speakers represent the 0.0 lateral range, and the rightmost speakers represent the 1.0 lateral range.
3. The ear-level speakers represent the 0.0 vertical range, and the height speakers represent the 1.0 vertical range.
4. If only 2 height channels are available, they represent the entire range 0.0-1.0 longitudinally. If you have 4, the forwardmost are 0.0 and the rearmost are 1.0. If you have 6: 0.0, 0.5, 1.0. And so on.

So, for example, if you have 5.1.x, the surrounds actually represent the back of the room (which is why surrounds are placed further behind the listener in this layout). If you have 7.1.x, the rear surrounds represent the back of the room and the side surrounds represent the midpoint for rendering purposes. So you still have to ideally put the speakers where Dolby has dictated they should go for any given layout.

What speaker remapping does in this context is try to correct for when speakers can't be placed at those ideal positions. But again, this isn't a thing for the overwhelming majority of Atmos systems, which simply rely on the above logic.

 

[FriedChicken]

Nope. Atmos doesn't do any speaker remapping; it places sounds based on cartesian coordinates, with certain predefined locations for some objects to represent channels within an object-based context. Here's the logic:
1. The forwardmost speakers represent the 0.0 longitudinal range, and the rearmost speakers represent the 1.0 longitudinal range.
2. The leftmost speakers represent the 0.0 lateral range, and the rightmost speakers represent the 1.0 lateral range.
3. The ear-level speakers represent the 0.0 vertical range, and the height speakers represent the 1.0 vertical range.
4. If only 2 height channels are available, they represent the entire range 0.0-1.0 longitudinally. If you have 4, the forwardmost are 0.0 and the rearmost are 1.0. If you have 6: 0.0, 0.5, 1.0. And so on.

So, for example, if you have 5.1.x, the surrounds actually represent the back of the room (which is why surrounds are placed further behind the listener in this layout). If you have 7.1.x, the rear surrounds represent the back of the room and the side surrounds represent the midpoint for rendering purposes. So you still have to ideally put the speakers where Dolby has dictated they should go for any given layout.

What speaker remapping does in this context is try to correct for when speakers can't be placed at those ideal positions. But again, this isn't a thing for the overwhelming majority of Atmos systems, which simply rely on the above logic.

How is what Atmos is doing any different from speaker remapping does, except on across a broader range?

 

[Reverend Slim]

How is what Atmos is doing any different from speaker remapping does, except on across a broader range?

Because it isn't remapping anything. The signal goes where it is supposed to based on the predefined expectations. Those predefined speaker locations are what they are. Remapping is to adapt misplaced speakers to the Atmos renderer's expectations.

 

[FriedChicken]

Because it isn't remapping anything. The signal goes where it is supposed to based on the predefined expectations. Those predefined speaker locations are what they are. Remapping is to adapt misplaced speakers to the Atmos renderer's expectations.

Is the speaker placement as critical with atmos though? Isn't that part of what atmos is saying: speaker placement isn't as critical, or did I mis-remember some of the marketing?

 

[Reverend Slim]

Is the speaker placement as critical with atmos though? Isn't that part of what atmos is saying: speaker placement isn't as critical, or did I mis-remember some of the marketing?

Dolby's marketing wants everyone to jump in, so even their own home setup guides are kinda' nebulous and unclear to accommodate people adopting the format. They want people to think that speaker placement isn't that critical because they want more people to do it. The reality, if you've done several rooms with Atmos in several different ways, is that there will always be an ideal placement method that will sound better if followed critically. For my money, the logic in their mix room guidelines gives you the best practice for placement and the logic to maintain separation between the layers. I think some Atmos is usually better than no Atmos at all (even with upfirers, if you do it right), but well-placed Atmos will always be better than poorly placed Atmos. But again, I've had the benefit of having heard just about every Atmos layout one can do.

 

[FriedChicken]

Then where does the improvement over traditional surround sound come from? I've seen people... on much more skeptical forums than this one, rush in to claim atmos is "the real deal". "Object based sound is the future". From my perspective, the same goals can be achieved (have been achieved!) with "traditional" surround sound. Maybe it makes the process easier in the mixing studio?

Does Atmos, by necessity, require existing room measurements to properly function? Does it give higher granularity to sound placement over a pre-mixed track?

 

[BJL]

Then where does the improvement over traditional surround sound come from? I've seen people... on much more skeptical forums than this one, rush in to claim atmos is "the real deal". "Object based sound is the future". From my perspective, the same goals can be achieved (have been achieved!) with "traditional" surround sound. Maybe it makes the process easier in the mixing studio?

Does Atmos, by necessity, require existing room measurements to properly function? Does it give higher granularity to sound placement over a pre-mixed track?

Audio that is purely channel based will be limited as to speaker arrangement by the number and arrangement of channels. An Atmos mix will (or should, if done right) accommodate the number of speakers in your system, assuming that the speakers are set up per Dolby's specification (which are extensively documented on Dolby's website).

For example, if you have a 7.1.2 set up, the Atmos renderer should provide a good approximation of the spatial characteristics of the mix. Improve your set up to 7.1.4, and the renderer will provide a more precise spatial mix. Go to a 7.1.6 set up (adding height middle speakers), and the renderer will take advantage of those speakers giving a more refined spatial presentation. Add front wides, and the atmos renderer will take advantage of those speakers as well. I believe that the Atmos renderer is downward compatible so that it will work even if there are no height speakers (although obviously you will not get the height spatial aspect).

With a pure channel based mix, you are stuck with the number of speakers equal to the number of channels in the original mix (putting aside upmixing post processing). So a 4.0 (quad) mix will provide four channels for four speakers, 5.1 surround will provide five main channels of audio, 7.1, seven main channels of audio.

Edit: Just for completeness, an Atmos mix usually includes both objects and channel based audio. The channel based audio is called the "bed". The link that follows has a series of short videos on how at Atmos mix is done using Logic Pro X, they are interesting to view for the listener, as well as an aspiring mix engineer, at least I found them interesting, they made me a more educated listener:

https://professionalsupport.dolby.com/s/article/Dolby-Atmos-Music-Getting-Started-with-Logic-Pro?language=en_US