Xbox One November SDK Leaked

[Scott_Arm]

Can i ask does the documentation say anything about the capabilities of shape ?

There's definitely some stuff on SHAPE in there. If I have time I can try to look through and see if there's anything interesting.

 

[Davros]

Thank you so much scott,
Theres a lot of confusion as to what shape does/doesnt do and afaik ms hasn't made any effort in publicizing shape to the general public other than "xb1 has an audio processor"

 

[Scott_Arm]

Thank you so much scott,
Theres a lot of confusion as to what shape does/doesnt do and afaik ms hasn't made any effort in publicizing shape to the general public other than "xb1 has an audio processor"

There's actually a lot of info in here without even reading through the APIs. Is there anything specific you're looking for?

Shape is fixed-function. It has DMA and an Audio Control Processor to control the fixed-function blocks. It has internal memory called Mix Buffers. A mix buffer stores one full frame of audio data. Typically reads from one or two mix buffers and writes to an output mix buffer. It does XMA format decompression of 512 voices in hardware. It supports PCM. It can do 512 channels of Sample Rate Conversion per audio frame, any mix of audio and stereo. SRC supports linear and polyphase interpolation of mono and stereo and a resampling range of 1:16 to 3.99:1. There's a long section on equalization, compression and filtering.

I don't see anything about direct access to the Audio Scalar Processor or the Audio Vector Processor. The ACP I guess is how you issue your commands to shape through a couple of queues that are handled differently.

 

[Davros]

I'm more after the high level info (aka info aimed at consumers, aka bullet points)
eg:
Does it wave trace (found out it doesnt)
does it take level material properties into account
Does it take into account level geometry (especially in levels with dynamic geometry)
does it handle diffusion
morphing/blending or is there a hard cutoff
does it calculate Doppler
does it calculate the effect of absorption (mainly for underwater ,sound traveling through walls ect)
does it calculate occlusion
regarding the convolution reverb : how are devs supposed to generate their impulse response recordings? does the sdk come with a range of them and devs are just supposed to tweak some parameters ? supply their own ? (how do they generate an impulse response of a non real life environment)

edit: why are the last 2 lines of my post in a different colour

 

[function]

edit: why are the last 2 lines of my post in a different colour

The text is a "hyperlink".

On the internet, text can be marked in html as a link, and it is common to highlight hyperlinked text with a colour different from "normal" text. Blue is a common colour for this.

If you click on the different coloured text it will take you to another location.

 

[Davros]

Strange I haven't added any hyperlinks

edit found the link when i selected edit and the use bb code editor
I just copied the words "impulse response" from wiki I didnt copy the page address

 

[iroboto]

I'm unsure as to where this quote from the polygon article goes. It's either here because it's addressing the rendering pipeline - or this post needs to go into the DX12 for Xbox One thread.

When I went back to Microsoft to get a better sense of that difference a company spokesperson told me that there are a handful of rendering pipeline features that are new in DX12 that we will be outlined at GDC. They added: "The power and frame rate wins we demonstrated come from improvements in CPU usage in the OS runtime and device drivers. And this was on DX11 devices."

If you missed it earlier, there was debates about Maxwell and X1 being DX12 fully compliant in which Phil responded to some tweets that 'We knew what DX12 was doing when we made Xbox One'

So I don't know if this is just unnecessarily turning our rumour mill or it's adding some strength back to the notion that the 2 GCP and 2 rendering pipelines are providing some sort of performance improvement over 1 and 1.

 

[bgroovy]

Nothing about your quote implies anything about additional command processors. In context the term "rendering pipeline features" is so imprecise it could refer to literally anything.

 

[iroboto]

True. It is too vague to link to anything.

 

[mosen]

True. It is too vague to link to anything.

We had this comment from 2013:

We also took the opportunity to go and highly customise the command processor on the GPU. Again concentrating on CPU performance... The command processor block's interface is a very key component in making the CPU overhead of graphics quite efficient. We know the AMD architecture pretty well - we had AMD graphics on the Xbox 360 and there were a number of features we used there. We had features like pre-compiled command buffers where developers would go and pre-build a lot of their states at the object level where they would [simply] say, "run this". We implemented it on Xbox 360 and had a whole lot of ideas on how to make that more efficient [and with] a cleaner API, so we took that opportunity with Xbox One and with our customised command processor we've created extensions on top of D3D which fit very nicely into the D3D model and this is something that we'd like to integrate back into mainline 3D on the PC too - this small, very low-level, very efficient object-orientated submission of your draw [and state] commands.

2nd GCP may not be related to DX12, but it seems that we will see some or all of XB1 command processor customizations on PC sooner or later.

 

[dcbronco]

Anandtech have an article up running early test on DX12.

http://www.anandtech.com/show/8962/the-directx-12-performance-preview-amd-nvidia-star-swarm/6

In an update to the section on Mantle versus DX12 Oxide says that because of the number of draw calls DX12 efficiencies allow, the command processor becomes the performance bottleneck. Same with Mantle. AMD chose a way to handle this with Mantle. Maybe the second command processor was Microsoft's way of handling it given Phil Spencer says they knew what DX12 was doing when they were designing Xbox One.

 

[iroboto]

Anandtech have an article up running early test on DX12.

http://www.anandtech.com/show/8962/the-directx-12-performance-preview-amd-nvidia-star-swarm/6

In an update to the section on Mantle versus DX12 Oxide says that because of the number of draw calls DX12 efficiencies allow, the command processor becomes the performance bottleneck. Same with Mantle. AMD chose a way to handle this with Mantle. Maybe the second command processor was Microsoft's way of handling it given Phil Spencer says they knew what DX12 was doing when they were designing Xbox One.

yes! good find! we are actually discussing this in the DX12 for Xbox One thread!