PS4 audio chip detailed

While the slide might be "new", that information was already known for months.
 
I think they just posted it because there was starting to be lots of fud/intentional confusion about it around the internet.

But yeah, Cerny laid out the details on this himself.
 
hey i just post this s***, and half the time im not even looking at what i post. But 2012 could be Dec...Hell, it could have been made in 2012, and not distributed till this year.

We appreciate you posting it, just thought it was odd that this was being posted now, instead of months ago if this slide is that old, seems to me I have seen it before too...a few months ago lol.
 
hey i just post this s***, and half the time im not even looking at what i post. But 2012 could be Dec...Hell, it could have been made in 2012, and not distributed till this year.

It was distributed to VGLeaks last summer. They had an article that included this info way back then except without the official slide. Nothing seems to have changed though based on Cerny's comments in the past. It's interesting that VGLeaks posted a direct slide with Sony's copyright on it though.
 
So what's the difference between this and the Xbox1 chip people seem to praise ?
Astro can go into details if he wishes, but in a nutshell the Sony chip is a very small part of what the X1 chip does. You can find detailed info on Shape (the X1 chip) by searching tech sites like VGLeaks, Anandtech or Beyond3D.
 
So what's the difference between this and the Xbox1 chip people seem to praise ?

This chips needs processes running on the CPU to feed it data and perform effect processing, like adjusting volume, reverb echo, and compression. That introduces latency. So if you wanted to use something like a voice changer, the feedback signal will be a fraction of a second after you talk. Which is hugely distracting. If the developer wants to run an effect stream that is even minimally complex(Like the Audio in Forza for example), the whole ensemble will have to be run and mixed on the CPU. In which case it is not a trivial amount of CPU processing. (this last sentence is my concession to astro)

It is also not able to run software that processes audio in to a stream of say... FFTs (Fast Fourier Transforms) which can then be fed into a voice recognition system(iirc, Shape goes beyond this emitting proper words, rather than just FFT phenomes). The FFT stream would have to be generated by the CPU. Again, the problem is latency. Voice recognition systems work best when they happen pretty much instantly. When you say "Grenade!" you want the grenade to throw that moment.

For next gen control schemes you need low latency. This is not a low latency audio system. This is last gen. You will have to run for the TV volume control to adjust volume when you change games, because everybody mixes game volume differently.
 
This chips needs processes running on the CPU to feed it data and perform effect processing, like adjusting volume, reverb echo, and compression. That introduces latency. So if you wanted to use something like a voice changer, the feedback signal will be a fraction of a second after you talk. Which is hugely distracting. If the developer wants to run an effect stream that is even minimally complex(Like the Audio in Forza for example), the whole ensemble will have to be run and mixed on the CPU. In which case it is not a trivial amount of CPU processing. (this last sentence is my concession to astro)

It is also not able to run software that processes audio in to a stream of say... FFTs (Fast Fourier Transforms) which can then be fed into a voice recognition system(iirc, Shape goes beyond this emitting proper words, rather than just FFT phenomes). The FFT stream would have to be generated by the CPU. Again, the problem is latency. Voice recognition systems work best when they happen pretty much instantly. When you say "Grenade!" you want the grenade to throw that moment.

For next gen control schemes you need low latency. This is not a low latency audio system. This is last gen. You will have to run for the TV volume control to adjust volume when you change games, because everybody mixes game volume differently.

Can't wait for next gen voice changing :txbcool:
 
This is last gen. You will have to run for the TV volume control to adjust volume when you change games, because everybody mixes game volume differently.
My control is always right by me. Where the heck do you keep yours down the street.lol
 
My control is always right by me. Where the heck do you keep yours down the street.lol

The more elegant solution is to not have to adjust the volume because the console is running system wide volume compression, and the peaks and quiets are the same when I switch from live TV to my game...
 
  • Like
Reactions: Ceger