Using System Objects in complex programs (DSP)

I'm looking for advice on best practice in using system objects in complex programs.

The examples in the documentation I've spotted so far are all oriented around either:

• toy problems (read a file, filter it, write it), or
• creating system objects from scratch

I'm looking at how to structure a more complex program combining many system objects. Specifically, I want to modularise the code rather than have a massive main program calling dozens of system objects. Any ideas where I should look?

Here is an example. I'm planning a program that processes signals embedded in a long timeseries. Just one of the tasks is to locate the times at which synchronisation pulses occur. The obvious approach is to apply a matched filter to the timeseries and then an adaptive threshold to the output of the matched filter, perhaps using the system objects:

• dsp.FIRFilter
• dsp.MovingAverage
• dsp.MovingStandardDeviation

I'm considering creating my own system object that calls the system objects listed above. But I haven't seen any examples that do that. Is that because there is a better way?

Any advice greatefully received.

Andrew

Answer 1

[Honglei Chen]

Yes you can do it, there are many shipping System objects that do this. This is essentially done the same as a regular class.

HTH