• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

Is 3D audio truly a next-gen exclusive (PS5) feature?

We already know that the PS4 supports 3D audio:

https://en.wikipedia.org/wiki/PlayStation_4_technical_specifications#Hardware_modules
https://en.wikipedia.org/wiki/AMD_TrueAudio

We also have current-gen (non-VR) games supporting 3D audio:

https://www.polygon.com/2017/8/9/16120082/hellblade-binaural-audio-psychosis


https://www.dualshockers.com/resident-evil-2-reception-tech/

So why does Cerny downplay PS4 audio capabilities? It definitely seems to be an upgrade compared to the PS3, which didn't even have an audio DSP and it had to use Cell for audio processing.

Was he trying to oversell PS5 or something? I mean, it's probably OK to oversell stuff like the Zen CPU and the SSD (huge next-gen upgrades!), but an audio DSP (or even GPGPU audio ray-casting) doesn't bring anything new to the table compared to what we already have.

If they're gonna support 3D audio in every PS5 game as a baseline (TRC mandated) feature, that's really cool and I welcome it, but it still doesn't explain why so few PS4 games support it.

I'd argue the same about gyroscopic controls, but that's a subject for another topic.

Please let's keep it civil, thanks! :)
 
Hellblade's audio is amazing, but it's just binaural recordings. It doesn't move with you. The 3D audio we all want is something that only VR games have started doing recently, with very rare exceptions of non-VR games like CS:GO. 3D audio is extremely processor intensive, which is why developers don't really chase after it.

PS5's part here is to provide a hardware accelerated audio solution so that less resources are used, making developers actually want to utilize 3D audio in their games. This will lead to a lot more HRTF based audio and likely even audio propagation for certain games. Raytraced audio was briefly mentioned as one method to enable propagation.
 

Shin

Banned
3D audio supported by ray-tracing, which the PS4 doesn't support.
I didn't bother looking it all up as it's boring but RT+audio = better positioning/audio is what I got out of all that.

Interesting, I didn't think of that.
Assuming that it all works (again not something I'm really motivated to read about for hours and hours) then it would make one heck of an immersive experience I reckon.
Not only visually but audible (is that even a real word in this regard?).
The 3D audio we all want is something that only VR games have started doing recently
 
Last edited:

Mikey Jr.

Member
Sound is cool, but for me, its usually the last thing on my list. Music, VO, sound effects, sure.
But how it actually sounds? Not something I'm really hyped for. Maybe just one of those things I have to hear to believe I guess.
 
Hellblade's audio is amazing, but it's just binaural recordings. It doesn't move with you. The 3D audio we all want is something that only VR games have started doing recently, with very rare exceptions of non-VR games like CS:GO.
I thought Hellblade supported 3D audio? Isn't binaural audio/HRTF the same thing? What's the difference?

3D audio is extremely processor intensive, which is why developers don't really chase after it.
Are you sure about that?

Because Aureal did it 20+ years ago: https://en.wikipedia.org/wiki/A3D

Modern processors are way faster. AFAIK, it's more of a legal than tech issue. We could have had it 20 years ago, if it wasn't for the stupid Creative vs Aureal lawsuit (long story).

It's amazing how legal hurdles can halt technological progress for such a long time. :(

PS5's part here is to provide a hardware accelerated audio solution so that less resources are used, making developers actually want to utilize 3D audio in their games. This will lead to a lot more HRTF based audio and likely even audio propagation for certain games. Raytraced audio was briefly mentioned as one method to enable propagation.
I still don't understand if PS5 will utilize a dedicated audio DSP (like OG PS4 does) or a more flexible GPU compute solution (seems to be the case for Polaris and onwards).

3D audio supported by ray-tracing, which the PS4 doesn't support.
OG PS4 supports fine-grain GPU compute features, which is probably what some games are using (via custom coding):

https://www.dualshockers.com/resident-evil-2-reception-tech/

"In terms of audio, we used Capcom’s dynamic mixing stage, a newly built 3D audio booth in our R&D building with 7.1.4 channel Dolby Atmos. This enables changes in sound in real-time using binaural technology, enabling realistic audio that makes players feel as if they are actually there. Zombie moans can be heard to the right and left, in front and behind, affecting not only the sense of sight but of sound to make users experience a greater feeling of horror."

To me it seems like that PS4 games don't have access to an OS API for 3D audio, they have to do it on their own (it reminds me of Remote Play on PS3 vs PS4). Maybe that's why so few games implement it?

XB1 seems to support Dolby Atmos features on an OS level and there's a chance Sony is doing the same on the PS5.

On the PS4 you can also buy a (pretty expensive) Sony Platinum headset that only supports 3D audio in certain games (like Uncharted 4 and HZD).

I'm genuinely confused, because I have the suspicion that this headset is not necessary at all (judging by Hellblade/RE2) and it will be redundant next-gen, if not already.

Sound is cool, but for me, its usually the last thing on my list. Music, VO, sound effects, sure.
But how it actually sounds? Not something I'm really hyped for. Maybe just one of those things I have to hear to believe I guess.
I think it's a real shame that so few people care about audio. It's all about 4K resolutions etc. Vanilla stereo audio is so archaic.

You need a pair of headphones to experience 3D/binaural audio. It can work through DualShock 4 (and possibly 5 as well).
 
Last edited:

Shin

Banned
Seeing how RT works, wouldn't 3D audio that's supported by RT give more depth / accuracy and "just" 3D audio (TrueAudio)?
That's what I made of their mention at least but IIRC there is something about headphones in the Wired article unless I remember wrong.
 
Seeing how RT works, wouldn't 3D audio that's supported by RT give more depth / accuracy and "just" 3D audio (TrueAudio)?
That's what I made of their mention at least but IIRC there is something about headphones in the Wired article unless I remember wrong.
OG PS4 GPU supports 3D audio ray-casting (among other things) according to Cerny:

https://www.dualshockers.com/ps4s-asynchronous-fine-grain-compute-gpu-what-the-heck-can-it-do/

Few people will remember this though. :)
 
All i care about is dolby atmos
DTS:X is excellent as well. It's a shame that quality audio gets the back burner these days. Playing a game like Gears 4 in true Dolby Atmos on a good Atmos HT is incredibly immersive, which is why I'm very happy that MS stays on the front lines supporting these newer formats. Oh, how I remember the days of trying to cop the latest and greatest audio cards on PC. Long live Soundblaster Wave. That S.O.B. was half the size of a motherboard. Lol.
 

Shin

Banned
Definitely the HD rumble of the PS5, it will be neat feature but most people won't care about it.
They should, late at night, girlfriend/child sleeping, headphones are just that much more better.
Sure if you got a 7.1 or whatever it is system these days it would sound a lot better and make u feel like you're living in the world (through sound alone), but I think headphones make it all sound better for some reason.
Perhaps I've had to use them too many times and/or played too much CSGO.
 
no. you can already do it on ps4. the problem is that it eats a good chunk of performance from the GPU.

maybe next gen it wont matter much
How much?

Audio is a lot less computationally intensive than graphics. You don't need Teraflops for audio.

It was a huge selling point in a presentation:

PS4FeatureSet-670x376.jpg


Sony has also said that the PSVR unit has a 3D audio chip (redundant since the GPU already does it?) and it doesn't consume a lot of watts. That's 2016 tech.

That's the thing though, is it the same?
Because now you got me wondering why they would mention that, might also want to look into WWise / AudioKinetic that they bought and what that was good for beside middleware.
Hmmm, that acquisition might be the culprit. I guess they had to amp PR a bit.

It reminds me of the PS4 Remote Play & Gaikai acquisition, even though the PS3 also supported PSP/Vita streaming (via Cell).
 

DunDunDunpachi

Patient MembeR
Yes, it was one of the standout features of the PS5 to me.

I don't know if it will truly be "exclusive" because Microsoft could still offer a similar featureset. In either case, it piqued my interest.
 

DunDunDunpachi

Patient MembeR
MS will also use a Navi GPU, so that's a given.

XB1 also supports 3D audio though, so it doesn't seem next-gen exclusive to me. :)
Right, I'm not sure why the title frames it as an "exclusive", though it sounds like Sony plans to standardize it and therefore every game will use it (or something?)

Hard to say until we see the hardware and can compare it ourselves.
 
Right, I'm not sure why the title frames it as an "exclusive", though it sounds like Sony plans to standardize it and therefore every game will use it (or something?)

Hard to say until we see the hardware and can compare it ourselves.
Next-gen exclusive = a feature available exclusively on next-gen consoles (PS5 and XBOX)

Sorry for the confusion. MS hasn't talked about it (yet) and I didn't expect Sony to tout it that much. I guess they have their reasons.
 

DunDunDunpachi

Patient MembeR
Next-gen exclusive = a feature available exclusively on next-gen consoles (PS5 and XBOX)

Sorry for the confusion. MS hasn't talked about it (yet) and I didn't expect Sony to tout it that much. I guess they have their reasons.
I think the fault of the confusion is mine. No problem though. I think we can agree that the use of sound in videogames is about to get a bit of an upgrade next-gen, and I'm thrilled about that.
 
How much?

Audio is a lot less computationally intensive than graphics. You don't need Teraflops for audio.

It was a huge selling point in a presentation:

PS4FeatureSet-670x376.jpg


Sony has also said that the PSVR unit has a 3D audio chip (redundant since the GPU already does it?) and it doesn't consume a lot of watts. That's 2016 tech.


Hmmm, that acquisition might be the culprit. I guess they had to amp PR a bit.

It reminds me of the PS4 Remote Play & Gaikai acquisition, even though the PS3 also supported PSP/Vita streaming (via Cell).
no. this is 3d pathtraced audio. it eats 1 compute unit out of the GPU. which is significant.
 
no. this is 3d pathtraced audio. it eats 1 compute unit out of the GPU. which is significant.
Source?

Because I gave one: https://www.dualshockers.com/ps4s-asynchronous-fine-grain-compute-gpu-what-the-heck-can-it-do/

Compute tasks don't necessarily have to eat CUs in a fixed manner (remember the 14+4 rumors? that was false according to Cerny).

OG PS4 GPU has 18 CUs that can also run tasks asynchronously (filling the gaps with compute tasks).

When people say that GCN GPUs are less efficient in graphics compared to nVidia (70% vs 90% utilization), this means that they have an extra 30% to fill the gaps with (it's the equivalent of SMT/HT for GPUs):

large-video-hyper-q-2-en.jpg


That's where audio, physics etc. can fit in nicely.
 
Last edited:

Arkage

Banned
XB had atmos support. I have a 5.1 system and don't really notice much difference between normal surround and 3d surround. Meh.
 
Source?

Because I gave one: https://www.dualshockers.com/ps4s-asynchronous-fine-grain-compute-gpu-what-the-heck-can-it-do/

Compute tasks don't necessarily have to eat CUs in a fixed manner (remember the 14+4 rumors? that was false according to Cerny).

OG PS4 GPU has 18 CUs that can also run tasks asynchronously (filling the gaps with compute tasks).

When people say that GCN GPUs are less efficient in graphics compared to nVidia (70% vs 90% utilization), this means that they have an extra 30% to fill the gaps with (it's the equivalent of SMT/HT for GPUs):

large-video-hyper-q-2-en.jpg


That's where audio, physics etc. can fit in nicely.

thats not how anything works. 3D audio eats 1 compute unit

https://en.wikipedia.org/wiki/AMD_TrueAudio

as you can see wikipedia it says you can reserve a compute unit . as you can see it says it has to utilize the GPU.
 

Jigsaah

Gold Member
3D Audio? I remember hearing something Cerny was saying like...raytracing and 3d audio kinda work hand in hand because sound is a wave and light is a wave so they basically behave the same way (I assume he obviously knows the difference in speed).

My brain started to bleed a bit trying to process all of that so....why is this a big deal again?
 
thats not how anything works. 3D audio eats 1 compute unit

https://en.wikipedia.org/wiki/AMD_TrueAudio

as you can see wikipedia it says you can reserve a compute unit . as you can see it says it has to utilize the GPU.
It is not compulsory to reserve CUs for compute tasks (physics, audio etc.)

What's so hard to understand? Have you done some research about asynchronous GPU tasks?

Btw, you earlier said that audio is computationally intensive. If that's the case, then 1 CU is not enough. That's only 102 Gigaflops.

For reference, TrueAudio DSP is around 15 Gigaflops and it can handle 3D audio as well.

Why is old tech suddenly such a talking/selling point?
Because Cerny revived it from its grave for some unknown reason?
 
I thought Hellblade supported 3D audio? Isn't binaural audio/HRTF the same thing? What's the difference?
This is binaural audio:



It has to be recorded using microphones that mimic the human ear. HRTF audio can use standard microphone recordings, as it's all done in software. Binaural recordings pick sound up at a fixed distance. If you pop a balloon next to one of these microphones, the binaural effect occurs because it was a set distance away in a particular direction. Therefore, you can't modify this in software, because the recording is fixed.

HRTFs on the other hand are sound sources that play in 3D space relative to the player and attempt to mimic binaural effects by providing a generalized head-transfer function. To put it simply, they use an algorithm that figures out how the sound waves should enter your eardrum (which is positioned normally where your camera is) by making the sound pipe through a general ear structure. It's imperfect, but very convincing. The key here is that it can be modified in real-time and works at any distance. You can't use binaural recordings for a monster drooling as it circles around you, because that monster isn't pre-determined; it has algorithms driving it's dynamic AI behaviour and movement. Using HRTF audio, the drooling sound would follow the monster and therefore change in distance and direction as it circles around.

For the perfect solution, we need a way to scan a real person's ear, use that as the HRTF model and propagate sound around a scene using a raytraced method. PS5 will likely do all of this aside from scanning your ear. Oculus are working on this though, so expect to see truly perfect 3D audio by the 2nd or 3rd generation of VR.

Are you sure about that?

Because Aureal did it 20+ years ago: https://en.wikipedia.org/wiki/A3D

Modern processors are way faster. AFAIK, it's more of a legal than tech issue. We could have had it 20 years ago, if it wasn't for the stupid Creative vs Aureal lawsuit (long story).

It's amazing how legal hurdles can halt technological progress for such a long time. :(
Oh I'm totally aware it's been done decades ago. This was as much as a frustration for Mark Cerny as it was for me. Since no one pushed it, optimization barely improved. It worked in the 90s because they were doing it via tech demos in very simple scenes. We want it in full games, and now companies like Oculus offer their spatial audio SDK with lots of huge optimization tricks (still ongoing) to make it more feasible.
 
Last edited:
Seeing how RT works, wouldn't 3D audio that's supported by RT give more depth / accuracy and "just" 3D audio (TrueAudio)?
That's what I made of their mention at least but IIRC there is something about headphones in the Wired article unless I remember wrong.
Raytraced audio is a really useful method for propagating sound around a scene. Basically think of it as an enabler for propagation. We've never seen real dynamic propagation in gaming, so it should be a big deal for immersion.

Sample video:

 
Last edited:
Definitely the HD rumble of the PS5, it will be neat feature but most people won't care about it.
Every VR user will thank Sony for it. I'm sure if developers support it well, a lot of people will care even without VR. It's a definite step forward when done right.
 

Rodolink

Member
Hellblade's audio is amazing, but it's just binaural recordings. It doesn't move with you. The 3D audio we all want is something that only VR games have started doing recently, with very rare exceptions of non-VR games like CS:GO. 3D audio is extremely processor intensive, which is why developers don't really chase after it.

PS5's part here is to provide a hardware accelerated audio solution so that less resources are used, making developers actually want to utilize 3D audio in their games. This will lead to a lot more HRTF based audio and likely even audio propagation for certain games. Raytraced audio was briefly mentioned as one method to enable propagation.
"Its just binaural recordings" sounds like is no big deal, but to do it right is.
(I dont know what Cerny is actually referring to with their 3d audio solution, have to research about it though)
 
"Its just binaural recordings" sounds like is no big deal, but to do it right is.
(I dont know what Cerny is actually referring to with their 3d audio solution, have to research about it though)
Hellblade is really great, but one day we won't use binaural recordings for the same effect. We'll use HRTFs and it will sound just as good, if not better. This will let you whisper into people's ears in multiplayer VR in a way that feels completely real.
 

Rodolink

Member
Hellblade is really great, but one day we won't use binaural recordings for the same effect. We'll use HRTFs and it will sound just as good, if not better. This will let you whisper into people's ears in multiplayer VR in a way that feels completely real.
That's what I'm hoping for this ps5 would be total game changer.
 

chinoXL

Member
this is good to see the 5 getting dedicated hardware to work on the audio side...i love my surround sound and headphone setups, so this is a good selling point for me. the 5 is lookin quite sexy right now
 
Top Bottom