The nForce Audio Processing Unit :

Unless you've been locked in a room working on spread sheets for the last five years you can't have failed to notice the incredible improvements in audio quality on the PC. From the early days of bland, characterless FM synthesized MIDI the PC has evolved into a genuine home audio solution with professional quality MIDI, multiple 3D positional audio streams, powerful DSPs and much improved signal clarity.

So what could NVIDIA do to improve on the already impressive standalone sound cards available right off the shelves of your local PC store? Enter the nForce2 APU!

nForce2's new APU features five discrete digital signal processors (DSPs) three of which are fixed function while the remaining two are programmable.

The heart of the APU is a processor audio rendering engine.

Setup Engine
This unit is responsible for performing all data and parameter setup for the other processors. All memory management, mapping and DMA resources are controlled in this unit.
Voice Processor
This unit contains several fixed function digital signal processing (DSP) units responsible for processing voices and mixing the results in the mixer buffers.
Global Processor
This unit is built around a programmable DSP. The DSP is responsible for adding varied effects to the data in the mixer buffers and producing the final output stream to the OS.
Dolby Interactive Content Encoder
This unit is built around a programmable DSP, which is responsible for encoding Dolby Digital (AC-3) data that’ll be sent over the SPDIF to an external consumer decoder. This allows 5.1 speakers (left front, right front, center, right rear, left rear, sub-woofer) to be transmitted over a digital interface.

Setup Engine
The vast majority of the more complex calculations are carried out by the setup engine. The functions handled here include ~:
- DMA channel management and prioritization.
- Management and updating of the voice structures in system memory.
- Processing of the voice execution lists¯2D and 3D.
- Parameter setup for the voice processor.
- Data de-interleaving for >2 channel voices.
- Data formatting - all data is signed 24-bit going into the voice processor.
- Data gathering - the voice processor is presented the sample data it needs from the sample buffers.
- Taking care of alignment and loop conditions, if needed.
- Down sampling averaging, if needed.
- Gathering output data, interleaving in the correct format, and then DMAing to system memory.
Voice Processor
The Voice Processor (VP) renders all the 2D and 3D voices. Running alongside the 2D audio processing is the Head-Related Transfer Function (HRTF) which is responsible for factoring in the listening cues for each ear including things such as environmental factors and shoulder reflections which are necessary for accurate positional effects. All samples are mixed into one of 32 mixer bins. The input data and parameters are ping-ponged on a voice basis.
Global Processor
As one of the two programmable DSPs the Global Processor is responsible for performing the following functions on each frame:
- Global Effects (reverb, chorus, flanger, etc.)
- Equalization
- 3D Cross-talk cancellation
- I3DL2 Reverb, Occlusion & Obstruction
- Post mix to Setup Engine
Dolby Interactive Content Encoder
The nForce APU integrates the Dolby Interactive Content Encoder into a programmable DSP with a fix-to-float format engine. It is used to take the output of the Global Processor and encode it into a Dolby Digital (AC-3) stream. Now users are able to experience true theater-quality, multi-channel surround sound rendered in real-time, from their Dolby Digital-equipped PCs and home-theater set-ups.

 

With support for the full audio feature lineup of DirectX 8 (DirectSound® 3D) nForce2 is angling to be the most complete and fully featured audio solution for the modern PC. With up to 256 hardware-processed voices (audio streams) or 64 hardware voices in 3D and support for multi-speaker 3D audio for up to six speakers NVIDIA have raised the bar for integrated audio solutions. Furthermore, the APU can also process and output a Dolby Digital audio stream directly to a home theater system via a SPDIF connection.

APU Features List 256 Total Voices Input Format Support 1-18 samples per block De-interleaving 8-, 16- and 32-bit containers Output Format Support 1, 2, 4, 6 samples per block 16- or 32-bit containers DirectX 8 Capable Sub-mix bins Second pass processing 64 3D Voices HRTF with cross-talk for speakers I3DL2 reverb Occlusion and obstruction Near field effects Full cross-fade per voice DLS2 Acceleration Two envelope generators per voice Two LFO generators per voice Loop and release segments DLS filtering Pan, pitch, vibrato, tremolo Reverb and chorus send 32 Bin Mixer 8 per voice volumes are mapped to each bin DX8 loop back processing 7-Band Graphic Equalizer I3DL2 Reverb and Occlusion Hardware/Software Interface Complete system memory based structure and control Fire-and-forget interface – minimum software control

 

Of course not all of the nForce2's advanced APU features will be integrated by all motherboard manufacturers so to simplify matters NVIDIA have introduced the "SoundStorm" standard. PCs with the NVIDIA SoundStorm solution have implemented the most complete digital audio feature set on their boards delivering stunning audio and fantastic sound effects, all powered by the nForce APU. SoundStorm exposes a vast array of inputs and outputs, including a digital SPDIF connection for access to the Dolby Digital 5.1 real-time encoder, connections for Microphone, CD, TV, and Line-in, Headphones, Front Left and Right speakers, Rear Left and Right speakers, a Center channel and a Subwoofer connection. To be certain you're getting the very best that the nForce2 APU has to offer look for "SoundStorm" support in the specs.

One thing that's noticeable by its absence is support for any of the main proprietary 3D audio standards such as Sensaura and EAX, at least it's not mentioned in any of the specs I have to hand. The improvements Microsoft have made to their DirectSound® 3D may make this less of an issue in the future but for some people I'm sure this alone will make them wary of relying on nForce for older titles. With no evaluation boards available to us at this stage it's hard to judge what kind of impact the lack of these proprietary audio technologies will have on gameplay but clearly NVIDIA believe the future of positional 3D audio rests with DirectX. DirectX 8 introduces a whole new set of audio features that the CPU must process if there is no dedicated audio hardware to support it. The APU was designed as a DX8 audio processor and has the ability to process advanced features in hardware such as:

• Head Reference Transfer Functions (HRTF) algorithms with cross talk
  cancellation
• Near field effects
• Macro effects
• 7-band graphic equalizer