Technical investigation into PS4 and XB1 audio solutions *spawn

Where you getting that information from ??

 

Davros, bkilian worked directly with the Xbox One SHAPE hardware. He's getting that from hands-on experience.

 

oooh,
/me needs to think up some questions

 

That is impressive and now looking forward to what fancy things will be done with audio that will be unique to the system. I'm very happy to hear that sound processing is actually not an after thought on MS's part and that it is an area of focus they have spent time, and a few transistors it seems, to advance it. If you're ever able to spare more details about the SHAPE processor I'm sure we would all be enthused to read about them. But I would like to thank you for revealing just this information here as it does give us a better idea of what the thing can do. If I may ask, what are the other 4 components of the chip for?

 

It's starting to sound like SHAPE is the most interesting\powerful part of the whole XB!

 

its certainly interesting, the cpu is nothing really new, same for the gpu
lots of unknowns about shape
does it support hrtf (Head-related transfer function) how good is it especially in the vertical plane
does it wavetrace, if so how many voices and to what depth
does it support occlusion, obstruction, refection, diffusion, absorption, detuning, pitch shifting
Low-frequency oscillation and delay for chorus

We wont know this untill ms release the capabiliies or a programming guide to the shape api

 

Yup, it is the only hardware feature (other than possibly Kinect 2.0 capabilities, but that's just a refinement of Kinect 1.0) of next gen consoles that I'm actually hugely interested in as well as hoping at least a few developer's attempt to exploit its capabilities.

Nothing else on next gen consoles excites me in any way. Well, other than the fact that unlike PC, all software will be native 64 bit.

Regards,
SB

 

it is interesting, and powerful. But it is also very limited in scope. It has the potential to decrease CPU requirements for audio significantly, but it's never going to be sexy.

 

I dunno, I've found audio hardware to be quite sexy ever since the Amiga broke the mould with PAULA. And then the alternative reconfigurable DSP replacement for the EMU10K in the old live cards. And there are so many other options available if you start to look at how to represent and manipulate data other than sound with the same hardware.

 

hrtf: it's not a native function of SHAPE, no, but could easily be run on one of the AVPs. Wavetrace: no.
All the rest: yes. All of those things can be achieved by a combination of the SHAPE blocks. Some of them just a single block. You might want to read the vgleaks doc linked earlier in this thread.

The stuff you're talking about is good in an audio engine, but are implemented by using one of the standard audio building blocks of filters, compressors, and src. SHAPE provides those building blocks in high quality. The audio engine (like, say Wwise) is responsible for stringing those blocks together into something the game would use, like occlusion, and others.

I'm surprised you didn't ask about the really expensive stuff, like reverb.

 

First of all, SHAPE is my favourite hardware module in the Xbox One. I don't know... I just admire bkilian.

Besides that, finally we have dedicated audio again on consoles! The thing I disliked the most about the PS3 and X360 was that they didn't have dedicated audio chips, and it certainly showed. The sound was not that good most of the time.

I was used to the original Xbox, which had a proper audio chip, and the lack of in on the X360 was a huge turn off.

My my, almost everything about the SHAPE sounds amazing. What is Wavetrace? I've searched on the internet and found nothing of real interest.

Also, it just dawned on me that the reverb can be expensive, but most of my experience with reverb is software reverb -unless Sound Blaster cards had hardware reverb, which I think is true, but they seemed to be very limited-.

Is SHAPE capable of equalising a top of the line, high quality hardware reverb? Could it be considered a fine piece of hardware reverb?

What do you mean when you define it as great but not sexy yet? -it's just a four letter word, after all-. Is it 'cos of hardware budget? (is is Microsoft's doing asking you to create the chip but not going overboard with features?)

 

Shape can't do reverb. The AVPs can, but it is unknown how much access to them developers have. The other missing expensive bit would be hardware Fourier transform.

By not sexy, I mean the average gamer doesn't give a rip how many channels you can mix. All they care about is pixels.

 

Trump. Thanks as usual! I gotta take a look at what the AVPs can do. I thought they were part of SHAPE...am I missing anything?

Well, you are probably right there but as you can see there are a few exceptions here in this forum. I am used to mix channels with simple but effective reverbs and other effects and I have to be very careful because whenever I venture down the experimentation path it brings my PC to its knees.

I should be sleeping already -gotta work tomorrow-, but I just wanted to ask. If the Sound Blaster XFi can do 10000 MIPS.... Then the SHAPE can do 10 raised to the power of 10 MIPS? 10^10 -I mean...because of the an order of magnitude term-

 

This is my favorite thread on B3D. Awesome is powerful gaming audio hardware!

 

What else in the system would need reverb besides games? If its just games, why wouldn't devs have acces to it?

 

Order of magnitude is x10, and really anything from 10x and less than 100x would be classed as an order of magnitude. 100x is two orders of magnitude.

And SHAPE is just one component in the audio chip, which is called something else.

 

The AVP is a general purpose DSP-like core. It is used for Kinect audio pipeline. According to the VGLeaks leak.

 

Are you saying Xbox had better audio than PS3?

 

The dedicated audio processing is certainly sounding interesting. I don't give a rip about how much it frees up resources so that xbone might be less behind ps4... I'm interested in what cool effects we might hear that aren't otherwise possible, and excited about quality audio in general being made a priority.

Any idea how the dedicated audio processor in xbone compares to the hardware found in capable consumer audio hardware, such as a good multichannel prepro that has advanced ambiance extraction/simulation, say something the likes of a lexicon processor (though their top of the line is long in the tooth and it doesn't look like a new one is coming soon)? Or perhaps compared to a device capable of multichannel FIR speaker/room correction? By compare, I mean both processing power and programming flexibility. Would devs even has access in a generalized way to allow such processing of incoming multichannel audio via HDMI? I ask, because $500 is nothing compared to many dedicated audio devices, and while the xbone obviously lacks the connectivity necessary to fulfill any switching functions, that isn't always a requisite for a high end audio processor. Wondering what possibilities there are for unusual applications of this "media box" MS has created looking years down the road

 

"Given A3D's similarity to OpenGL the plane data (i.e. vertex data and material type) imported earlier for the graphics engine can be used by the sound engine as well. Each plane surface will be of a specific material type given as a number key. This key (say 302 for wood) can be used to index into an array of textures and an array of sound absorption and transmission coefficients simultaneously. For wood the application can find the six absorption coefficients that represent a wood surface for the standard range of frequency bands. These six absorption coefficients can be converted to high and low frequency components used as plane material data for the A3D wave tracing engine.
http://www.acoustics.salford.ac.uk/res/drumm/index.php?content=virtual_reality