Technical investigation into PS4 and XB1 audio solutions *spawn

So clearly MS sees sound as the next graphics?

I guess that the monitored for the X360 that sound takes quite a junk of the quite small CPU resources. To improve and evolve the X360 concept for the One, they included the dedicated sound processing hardware. Furthermore, they figured that at the same time they can use it for advanced Kinect voice processing.

I believe that this is the rational behind the sound processor, also seems to be stated in the recent X1 dev interview that way...and it seems to be a...sound strategy :)
 
... still perceived the SHAPE... as something that could alleviate the big performance gap between both next generation consoles.
Only fanboys are talking about a big performance gap at this point. I expect the PS4 to maintain a small performance lead in many usage scenarios, not all, and for the most part the gap will be small enough as to have negligible impact on on-screen graphics.



We need to get Cerny on the line, because I am pretty sure that the PS4 does have game centered audio hardware. Also I'm pretty sure that if he had 4 audio cores at his disposal, he would have used all of them for game audio, and none of them for voice recognition. I base this on the design mantra of both consoles.[/QUOTE]
 
I guess that the monitored for the X360 that sound takes quite a junk of the quite small CPU resources. To improve and evolve the X360 concept for the One, they included the dedicated sound processing hardware. Furthermore, they figured that at the same time they can use it for advanced Kinect voice processing.

I believe that this is the rational behind the sound processor, also seems to be stated in the recent X1 dev interview that way...and it seems to be a...sound strategy :)

well, this is what Bkilan said:

The Audio processor was originally devised to be able to offload Kinect Audio processing, and the chip designers came to the audio team and said "We have a bunch of extra transistors we can throw in for free, what would you like them to do?" or something close to that. The SHAPE block was the result of that conversation.
 
is SHAPE a part of a larger audio block or is "audio block" SHAPE? sometimes i see you guys talking about the audio block and other times i see you talking about shape.

so is parts of shape reserved for kinect/speech, or is parts of the audio block reserved for kinect/speech and shape is used for audio mixing/compression and all the other cool stuff you guys have been talking about?
 
The audio block is made up of Kinect & SHAPE.

Guess sometimes people will just say audio block because it's pretty obvious when your actually talking about Kinect or SHAPE, or both.
 
We need to get Cerny on the line, because I am pretty sure that the PS4 does have game centered audio hardware. Also I'm pretty sure that if he had 4 audio cores at his disposal, he would have used all of them for game audio, and none of them for voice recognition. I base this on the design mantra of both consoles.

Without the audio processing requirement for Kinect I guess there'd be no SHAPE for game audio.
 
how many cus would be needed to emulate shape? like say there was a game that made really good use of shape. in order for the ps4 port to equal they would have to use many cus right? wouldn't that effectively get rid of any rendering advantage?

The audio block is made up of Kinect & SHAPE.

Guess sometimes people will just say audio block because it's pretty obvious when your actually talking about Kinect or SHAPE, or both.

ah okay thanks.
 
how many cus would be needed to emulate shape? like say there was a game that made really good use of shape. in order for the ps4 port to equal they would have to use many cus right? wouldn't that effectively get rid of any rendering advantage?



ah okay thanks.

Some people have claimed it would cost 2CU's. That is assuming the figure is true and PS4 has no audio hardware at all.
And assuming that the hypothetical PS4 game would constantly require your voice input as well.
 
So the people who claimed 2CU's we're they talking about SHAPE or the audio block, because if 2CU's could do the whole audio block that's pretty interesting.

Releaving pressure of CPU whether it be by CU's or custom hardware can only be a good thing surely?

As for releaving pressure of CPU & GPU by having the audio block, I'm at a loss why people would think that's a bad thing, given MS vision of Kinect.

It's a different matter if you buy into their vision or not, and probably is not the right thread to explore that.

Sometimes it only takes giving people the tools or hardware for them to start to do things they may not have bothered with to much in the past with.
 
The audio block is more than the 8-core CPU could manage with the same quality effects. I recall this comparison being made.

SHAPE will save ~a CPU core from game audio?


Some people have claimed it would cost 2CU's. That is assuming the figure is true and PS4 has no audio hardware at all.
And assuming that the hypothetical PS4 game would constantly require your voice input as well.

No. PS4's dedicated audio hardware is principally for compression and decompression of audio streams according to Cerny, as with PS360. SHAPE is supposed to do pretty much everything else too.
 
No. PS4's dedicated audio hardware is principally for compression and decompression of audio streams according to Cerny
thats what shape appears to me like, and most of what ive read about shape seems to back that up

The audio block is more than the 8-core CPU could manage with the same quality effects.
where did you get that info from ?
 
thats what shape appears to me like, and most of what ive read about shape seems to back that up


where did you get that info from ?
From what I remember of bkilian's posts in this thread, and here: http://forum.beyond3d.com/showthread.php?p=1731291#post1731291
Interference has said a few times that another 8 CPU cores wasn't enough to effectively emulate the audio block in early devkits.

Bkilian suggest that "the performance described in that supposedly leaked doc could not be replicated on the 360, even if all 3 cores were used purely for the audio."

I'm not sure how the 360 matches up with the PS4/X1 CPU for audio.
 
From what I remember of bkilian's posts in this thread, and here: http://forum.beyond3d.com/showthread.php?p=1731291#post1731291
Interference has said a few times that another 8 CPU cores wasn't enough to effectively emulate the audio block in early devkits.

Bkilian suggest that "the performance described in that supposedly leaked doc could not be replicated on the 360, even if all 3 cores were used purely for the audio."

I'm not sure how the 360 matches up with the PS4/X1 CPU for audio.

but isnt that for the entire audio block and not SHAPE? shape is only 15gflops itself.
 
but isnt that for the entire audio block and not SHAPE? shape is only 15gflops itself.

15.4GFLOPs is the vector cores. SHAPE is fixed function audio equivalent to a few CUs or >100GFs of audio processing, according to interference.

Here's the post I was thinking of earlier with that SHAPE/CPU comparison:
SHAPE, if it were utilized 100% at all times, would be hard to equal in CPU, yes. But they don't need to equal it. For one, it's using better, but more expensive algorithms, for which the cheaper versions work fine, and have been used for the last generation without complaint. And second, It's highly doubtful that it will be utilized 100% for most games. I'd be surprised if developers used even 50% of it's capabilities for most titles.
By the time developers are looking to push the capabilities of the audio block, I suspect using GPU compute audio will be well understood and a reasonable solution.
http://beyond3d.com/showpost.php?p=1766656&postcount=317
 
Last edited by a moderator:
thats what shape appears to me like, and most of what ive read about shape seems to back that up
Nope. The only part of Shape that does decompression of audio streams is the XMA decompression block, and that is a souped up copy of the XMA hardware in the 360. The codec DSP also does decompression/compression, but it's not part of Shape. Shape is the fixed function block. The full audio block, which includes the 4 tensilica cores and Shape, has a different name.
With "nothing to write home about" I meant that if half of it is Kinect and the other half is.. 50% scheduling? It's possible that they just put a relatively standard component as an extra core. And by standard I mean: something like an integrated audio chip like you have on a 60 dollar motherboard these days: even the cheapest ones can do HRTF 3D Positional Audio, which would probably be received as a very big thing over here.

We need to get Cerny on the line, because I am pretty sure that the PS4 does have game centered audio hardware. Also I'm pretty sure that if he had 4 audio cores at his disposal, he would have used all of them for game audio, and none of them for voice recognition. I base this on the design mantra of both consoles.
I can't speak to what the PS4 has, but the X1 audio block would put the best sound card you can buy to shame. And that's _before_ you add in _any_ of the DSP cores. And the DSP core for scheduling removes a huge burden from the CPU requirements for audio processing. If all the chip did was offload effects scheduling and mixing it would easily half the CPU requirements for audio compared to the 360. It does a lot more than that.

But as I've said before, it helps with CPU processing and will not perform any magical GPU upgrade. It just means games on the X1 will have more CPU headroom to either use for reducing the amount of time it takes to get the game to a happy place CPU wise, or increase the amount of CPU tasks being done. The realist in me suspects it'll be the former. When given the choice of "We can have a lower development cost" versus "We can fit in a few more AI tasks", any game company that is not a first party is going to choose lower costs.
 
I am a bit excited that X1 will be an audio beast since I am an amateur audiophile! Looking forward to better sound from next-gen games and not just graphics increases.
 
Nope. The only part of Shape that does decompression of audio streams is the XMA decompression block, and that is a souped up copy of the XMA hardware in the 360. The codec DSP also does decompression/compression, but it's not part of Shape. Shape is the fixed function block. The full audio block, which includes the 4 tensilica cores and Shape, has a different name.
I can't speak to what the PS4 has, but the X1 audio block would put the best sound card you can buy to shame. And that's _before_ you add in _any_ of the DSP cores. And the DSP core for scheduling removes a huge burden from the CPU requirements for audio processing. If all the chip did was offload effects scheduling and mixing it would easily half the CPU requirements for audio compared to the 360. It does a lot more than that.

But as I've said before, it helps with CPU processing and will not perform any magical GPU upgrade. It just means games on the X1 will have more CPU headroom to either use for reducing the amount of time it takes to get the game to a happy place CPU wise, or increase the amount of CPU tasks being done. The realist in me suspects it'll be the former. When given the choice of "We can have a lower development cost" versus "We can fit in a few more AI tasks", any game company that is not a first party is going to choose lower costs.

Okay you convinced me. One day I might have an Xbox One, and if that day ever comes I look forward to hearing the hopefully amazing 1st party titles.
 
Back
Top