Technical data

www.asti-usa.com ASTi DACS: Operation & Maintenance Manual (Version 4, Rev. B.1)
Copyright © 1999-2006 Advanced Simulation Technology inc. 35
Model Generation and Operation Overview
The following sections provide a general operational overview of the DACS.
Overview Of Hardware and Software Functions
The Model Builder software is both a development platform for developing models and what
actually runs the model. A model can be run as it is being entered, and changes can be seen (or
heard!) immediately. It is not necessary to compile models. Models are saved as les with an .mdl
extension.
There is one model running for each DSP card. The models run on the CPU, but they control the
DSP cards. (A DSP card is either a Waveform Synthesizer or TDM controller.) Each model directs
its DSP card to manipulate the sound it receives from the analog inputs, what sounds to generate,
and what sound to send to the analog outputs.
Sounds can be shared between DSP cards (and models) through a 40 wire ribbon cable which
connects the DSP cards together. Sharing sounds between DSPs requires that the models for those
DSPs be coordinated with each other.
Some sound sharing can be done nearly automatically. For example, intercom busses can be set up
across DSP cards to act as one bus. Simulated radio communications can also automatically send
sound from one DSP to the other. (For details on sharing sound across DSPs, see the section on
“Global Channels” in the Model Builder Reference Guide.)
Model Builder is designed with high level simulations of radios and intercoms. To the person
making the model, they will see a “Radio” object, and will specify the frequency, world position,
transmission power, etc. of the sender and the receiver. The Model Builder software will then
determine if the receiver can hear the sender, and the signal-to-noise ratio. It will then route the
audio from the sender, with the appropriate amount of noise mixed in, to the receiver. In this case,
the CPU determines where the sound goes, the DSP routes the sound and mixes in noise based on
the CPU’s instructions. The user, however, just sees a “Radio” object.