Hi Everybody,

I was reading Sonia's post about using a DSP and it got me to thinking. Over the years I've developed somewhat of an intuitive notion of the difference between a DSP and a normal µ-processor, just like I've developed a rather intuitive notion for the difference between a µ-processor and a µ-controller. But it occurs to me that I've never actually heard or developed a definitive difference.

For example, I get that DSPs are designed with processing analog signals in either time or frequency domain. Nonetheless, strictly speaking any UART is a digital signal processor since it processes a time domain signal into digital data. And I think most everyone here has thought about where to draw the line between µ-controller and µ-processor.

So what I am proposing for discussion is this. What are the definitive differences between these various types of processors? Are there definitive differences, or are the labels more intuitive distinctions?

Obviously if someone labels a UART as a DSP they'll be run out of business. But if I design a selection of hardware filters and incorporated them into a board with an 8052, could I call the resultant SBC a DSP? If not, why?

Joe