NextGen Audio: Tempest Engine, Project Acoustics, Windows Sonic, Dolby Atmos, DTS X

@ToTTenTranz So, I think his point was the receiver is able to render the sound objects into the correct positions in the channels because the receiver can be aware of the physical locations of the speakers relative to the listener. For example, you'd record the physical distance of each speaker to the listener and provide those to the receiver so it can balance the mix correctly based on your exact speaker layout. I don't mess with home theatre, but my understanding is these room calibrations are common now, and a console that was mixing the audio into channels before sending it to the receiver would lose the benefit of receiver mixing the sound objects for your particular setup. Of course, the console could support the same type of room calibration.

Automatic room calibration systems don't make the receiver aware of the physical location of each speaker. I've installed a few home theater surround setups and I haven't found nor heard of any consumer AV Receiver that does it.
Room correction procedures only use one omnidirectional microphone. All they can gather is distance, speaker performance (frequency response, volume) and some room-specific responses like echoes and reverbs. In order to know speaker positions they'd need to use microphone arrays, but what they use is a mono microphone like this:

aotAWT8.jpg



Any Audyssey / MCACC / YPAO room correction setup wouldn't return an error if you physically put the right speaker on top of the rear left one.

They can do wonders if you have e.g. surround speakers that need to be too far away in your room, compared to the main front speakers, essentially by injecting some latency into the nearest ones and doing a pre-equalization to the far reaching ones to make them sound like they nearer.
They don't compensate for incorrect angles, since they don't even detect those.
 
I am not an audio expert but I think we lack Atmos specific information here, from my vague reading on it I think the key point is it is not a codec it's 3d audio object metadata. DDPlus and DD Master are the codecs, audio streams and yes you could argue close to LPCM whoever even thoes contain basic metadata about intent that the final audio device should be aware of. Sending LPCM removes that.

Atmos is the object meta data on top of the 7.1 or whatever actual data.

LPCM is not the same as the mix is set, fire and forget. In ideal setups that's probably OK but the other formats are designed to get as close to the authors intent regardless.

https://www.dolby.com/us/en/technologies/dolby-metadata.html
 
I am not an audio expert but I think we lack Atmos specific information here, from my vague reading on it I think the key point is it is not a codec it's 3d audio object metadata. DDPlus and DD Master are the codecs, audio streams and yes you could argue close to LPCM whoever even thoes contain basic metadata about intent that the final audio device should be aware of. Sending LPCM removes that.

Atmos is the object meta data on top of the 7.1 or whatever actual data.

LPCM is not the same as the mix is set, fire and forget. In ideal setups that's probably OK but the other formats are designed to get as close to the authors intent regardless.

https://www.dolby.com/us/en/technologies/dolby-metadata.html

Yes, Atmos uses 3d object metadata. It works as if it was an audio game engine where the distribution of the objects throughout the speaker channels is processed on the AV Receiver (and here's probably where the 32 voice limit comes in, since that might be how much the DSPs in current Atmos receivers can widthstand).
But if you have an audio game engine running in your game then there's no need to process that on the receiver, you can process that on the console.

Which is what's already been happening on the PC for decades. The Creative EMU10K1 soundcards from 1998 would already distribute positional object data (i.e. output from videogame directional audio engines), and in games like Unreal or Thief 2 it would distribute it towards 2 front + 2 rear speakers, using the popular FourPointSurround set of that time.
Then in the next year the Live 5.1 was doing that on analog 5.1 sets, then a couple of years later the Audigy 2 upped that to 7.1. In the meanwhile nvidia with SoundStorm in 2001 on the nForce chipsets did that while encoding the result for Dolby Digital, followed by other soundcard manufacturers using Realtek and C-Media codecs (by now Dolby was licensing the encoding algorithm and called it Dolby Digital Live).
In the console space the Playstation 2 did multispeaker surround on a few games through Dolby Digital 5.1, then the first XBox did it on all its games (using nvidia's SoundStorm), and even the GameCube had directional sound for surround speakers in its games by using Prologic II through stereo analog outputs.

The only difference AFAIK is that Atmos uses height information, whereas most positional audio engines in videogames use only a horizontal plane.


So to summarize, in the end, it doesn't really matter if you process each speaker's output in the A/V Receiver (using Dolby Atmos or DTS: X) or in the PS5 (using LPCM) in a videogame. The room correction features from an AV Receiver will work whether you process Atmos on the receiver or just forward a LPCM stream (they'll even work if you use analog inputs for each speaker which is certainly an option for many PC setups -> this is a question I personally posed to a number of reps from IIRC Pioneer and Onkyo a while ago).

In the case of Sony, the decision to go with LPCM (instead of encoding voice coordinates into an Atmos stream) seems to have been a conscious one not only to save licensing costs and processing time to encode and compress the audio stream but also to overcome some limitations like Atmos' 32 voices that most receivers can process.
I don't think Sony would invest all this money on developing Tempest, an updated audio engine, new HRTF technologies etc.and then decide to skip Atmos just to save on royalties, if they had anything to gain in adopting Atmos.


On some newer and higher end Yamha AVR's YPAO has angle measurement and calibration, but we should not turn this into a Hifi thread:D
Are you sure? If you're talking about YPAO RSC then it doesn't measure angle, and that's the tech they're using in their top of the line receivers (at least the ones I can see in the website right now).
Regardless, I think AVRs using a small microphone array to detect and correct angle placements is bound to happen. I just don't know of any consumer product that does it yet.
 
Are you sure? If you're talking about YPAO RSC then it doesn't measure angle, and that's the tech they're using in their top of the line receivers (at least the ones I can see in the website right now).
Regardless, I think AVRs using a small microphone array to detect and correct angle placements is bound to happen. I just don't know of any consumer product that does it yet.

"YPAO-R.S.C Sound Optimisation with High Precision 64-bit EQ Calculation and Speaker Angle Measurement
YPAO analyses room acoustics and performs speaker angle measurements, then calibrates audio parameters to achieve the best sound at any of several listening positions. The height of the presence speakers is also measured, to optimise the 3D sound field. R.S.C. (Reflected Sound Control) corrects early reflections for studio-quality sound. YPAO also uses 64-bit EQ Calculation for realising the most natural room acoustics possible.YPAO 3D provides automatic sound parameter optimisation that maximises the 3D sound field effectiveness of CINEMA DSPHD3 as well as Dolby Atmos."

https://europe.yamaha.com/en/products/audio_visual/av_receivers_amps/cx-a5100_g/features.html
 
"YPAO-R.S.C Sound Optimisation with High Precision 64-bit EQ Calculation and Speaker Angle Measurement
YPAO analyses room acoustics and performs speaker angle measurements, then calibrates audio parameters to achieve the best sound at any of several listening positions. The height of the presence speakers is also measured, to optimise the 3D sound field. R.S.C. (Reflected Sound Control) corrects early reflections for studio-quality sound. YPAO also uses 64-bit EQ Calculation for realising the most natural room acoustics possible.YPAO 3D provides automatic sound parameter optimisation that maximises the 3D sound field effectiveness of CINEMA DSPHD3 as well as Dolby Atmos."

https://europe.yamaha.com/en/products/audio_visual/av_receivers_amps/cx-a5100_g/features.html
Thanks!

Hum.. it's still just using a mono microphone, but it seems during setup they ask you to move the microphone around, which is how they get angle measurements.
Regardless, considering it's a person moving the microphone then I wonder how accurate the angle measurements can ever be.
 
@ToTTenTranz Something is missing here. What was the point of Atmos sending object data to the receiver? If the receiver can't add any benefit, why not just mix the sounds into the channels in the studio before encoding?
 
Are there any statistics on what audio setups console gamers use? Would be interesting to know if 2speaker/headphone/soundbar setup is the dominant one or has multichannel or even more than 5.1 channel become typical?
 
@ToTTenTranz Something is missing here. What was the point of Atmos sending object data to the receiver? If the receiver can't add any benefit, why not just mix the sounds into the channels in the studio before encoding?
Mixing the sounds into the channels before encoding is exactly what every encoding/compression technique from Dolby and DTS did, before Atmos and DTS: X.
The benefit of Atmos in series and movies is that the precision of sound positioning (or its perception) with scale up as you add more speakers.
- A 7.1 surround system can't do much for the rear left+right surround speakers when receiving a Dolby Digital 5.1 source, other than replicate the side surround speakers.
- A 5.1.2 system can't do anything for the height speakers when receiving a 7.1 Dolby TrueHD source. Redirecting the rear surround to the height speakers will just worsen the situation.

With Atmos, the system will just calculate each and any speaker's output from the object data. With an Atmos source you (theoretically) always stand to gain by increasing the number of speakers. Go from 2.0 to 5.1 and with Atmos you gain positioning. Go from 5.1 to 5.1.4 and you gain positioning. Go from 5.1.4 o 7.1.4 and you still gain positioning.

But this is only new for video (movies and series). At least on horizontal plane sound sources, videogames have had this "feature" for decades.
Which is why the PS5 (nor the SeriesX) doesn't need it.
 
Mixing the sounds into the channels before encoding is exactly what every encoding/compression technique from Dolby and DTS did, before Atmos and DTS: X.
The benefit of Atmos in series and movies is that the precision of sound positioning (or its perception) with scale up as you add more speakers.
- A 7.1 surround system can't do much for the rear left+right surround speakers when receiving a Dolby Digital 5.1 source, other than replicate the side surround speakers.
- A 5.1.2 system can't do anything for the height speakers when receiving a 7.1 Dolby TrueHD source. Redirecting the rear surround to the height speakers will just worsen the situation.

With Atmos, the system will just calculate each and any speaker's output from the object data. With an Atmos source you (theoretically) always stand to gain by increasing the number of speakers. Go from 2.0 to 5.1 and with Atmos you gain positioning. Go from 5.1 to 5.1.4 and you gain positioning. Go from 5.1.4 o 7.1.4 and you still gain positioning.

But this is only new for video (movies and series). At least on horizontal plane sound sources, videogames have had this "feature" for decades.
Which is why the PS5 (nor the SeriesX) doesn't need it.

I still don't really see the point of atmos. So essentially atmos exists to save disc space for movies, so they don't have to encode the wide variety of speaker configurations that are available. Atmos will mix real-time to whatever configuration you have, up to 7.1.4.

In the game situation, the game would have to be aware of your speaker configuration to know how to produce the mix. Like maybe at the system level in your console you specify whether you have height channels or not etc.
 
Complexity of different speaker setups/room acoustics is probably the big reason why sony is starting with headphones and then 2 speaker setups. I suspect another big reason for that is the equipment majority of users have/use. Statistics on console user audio setups and their use would be interesting,...

If sony wanted to do something really fancy for multichannel they could run some room/speaker calibration setup by asking user to place the controller into different positions in room and then playing calibration signal from different speakers. Read back the calibration signal back from microphone in controller and finetune audio output. In essence sony can do all the traditional receiver audio processing/calibration in ps5 if there is will to do a lot of implementation. There is no mandatory need to do that processing in receiver if ps5 has sufficient implementation and power to run similar or even better algorithm.
 
I still don't really see the point of atmos. So essentially atmos exists to save disc space for movies, so they don't have to encode the wide variety of speaker configurations that are available. Atmos will mix real-time to whatever configuration you have, up to 7.1.4.
Atmos is speaker setup-agnostic, and that's as good as movie audio will ever get (apart from the 32 source limit).
I don't know if the 7.1.4 is a hard limit for Atmos. I think in theory you could have a 24.2.8 speaker configuration and Atmos would know what to do with it.


In the game situation, the game would have to be aware of your speaker configuration to know how to produce the mix. Like maybe at the system level in your console you specify whether you have height channels or not etc.
It already happens in Windows.

go38xol.jpg


It shouldn't be hard to do this on a console.
 
Automatic room calibration systems don't make the receiver aware of the physical location of each speaker. I've installed a few home theater surround setups and I haven't found nor heard of any consumer AV Receiver that does it.
Room correction procedures only use one omnidirectional microphone. All they can gather is distance, speaker performance (frequency response, volume) and some room-specific responses like echoes and reverbs. In order to know speaker positions they'd need to use microphone arrays, but what they use is a mono microphone like this:

aotAWT8.jpg



Any Audyssey / MCACC / YPAO room correction setup wouldn't return an error if you physically put the right speaker on top of the rear left one.

They can do wonders if you have e.g. surround speakers that need to be too far away in your room, compared to the main front speakers, essentially by injecting some latency into the nearest ones and doing a pre-equalization to the far reaching ones to make them sound like they nearer.
They don't compensate for incorrect angles, since they don't even detect those.

Don't be so sure!

I have sony Str-dn1080 (2-3 years old model with atmos support) and it uses stereo mic for calibration. I'm not sure if it detects wrong placement, but I have hazy memory that it will if wires are on wrong places. And it is not even that expensive, I paid 399 € around 2018 (it started at 899€ on launch) new on sale when I bought it, and back then it got plenty of 5/5 reviews. So maybe you have installed poor quality receivers or other manufacturers wont see this as necessary thing (to have stereo mic)

71T6fo0o8SL._AC_SL1500_.jpg
 
Back
Top