Digital signal processors

1.

Digital signal processors

2.

Digital signal processor - specialized microprocessor designed for
digital signal processing

3.

Advantages of DSP
• Versatility
• Reproducibility
• Simplicity
DSP finds use in many practical applications
In each area of DSP processors are characterized
by common properties :
• They use a large amount of mathematical calculations
• They deal with signals from the real world
• signal research lasts a limited time

4.

Memory architectures
Typical DSP operations require many simple actions to be performed.
• make a selection of two operands
• perform addition or multiplication (usually both)
• save the result or hold it until repeated
There are two types of DSP processor architectures that allow implementing
multiple memory access mechanism :
• Harvard Architecture
• Modified von Neumann architecture

5.

Harvard Architecture
Important to note that it is often necessary to make a selection of three
components - instructions with two operands, which the Harvard architecture
itself is incapable of.

6.

Von Neumann architecture
This architecture has a number of positive features.
• It is cheaper,
• Requires fewer bus pins
• It is easier to use, since the programmer can place both commands and data
anywhere in free memory..

7.

I/O Interface
In practice, the DSP mainly deals with the real world. Although it is often forgotten, this
feature is one of the most significant differences between DSP processors and universal
microprocessors.
There are three data sources for DSP processors:
• input and output signals
• interaction with various system controllers
• interaction with similar DSP processors

8.

Analog signal conversion
When converting a signal from an analog form to a discrete part of the
information is lost due to:
• measurement errors
• inaccuracies in synchronization
•restrictions on the duration of measurements
Important to note that the signal is sampled only when the entire signal has been
previously stored. This means that slower ADCs can be used . But the circuit responsible
for pre-storing the signal must function fast enough so that the signal does not have time
to change significantly. After saving the signal, the ALU does not need high speed to
convert it into digital form.
ALU - Arithmetic logic unit

9.

Sampling errors
DSP systems are affected by three sources of errors:
• Limitation of the digital conversion of the signal of its accuracy by the finite length of
the machine word
• Limited accuracy of arithmetic calculations performed by the processor
• Limiting the accuracy of the signal by the length of the machine word when converting
it from a discrete form back to analog