The Steven Smith book was originally published in 1987, so some things in
it are out of date. The mathematical theory of DSP is still sound; so far as I know, there have been no revolutionary changes in the intervening years. There are probably some new computational shortcuts that have been developed since 1987. Using parallel processing is not addressed at all, as it was still a dream of most engineers at the time. (It was already a limited reality in supercomputing: mainframe CPUs with vector capabilities such as the Cray 1 and its successors, and experimental massively parallel systems like the Connection Machine.) The comments on the performance levels of DSP on personal computers are woefully out of date. An AVR microcontroller, such as found in a classic Arduino board or a QCX, is probably about on a par with an original IBM PC. (The AVR executes instructions significantly faster, but loses the advantage back by being only an 8 bit processor.) A high end microcontroller based on a Cortex-M7 (such as in the Teensy 4.0 and 4.1) is at least three orders of magnitude faster at doing DSP. (32 bits, hardware floating point, dual instruction issue, and higher clock speed.) The processors in your smartphone and computer are faster still. And if you REALLY need a bludgeon and your computing algorithm can exploit its massive parallel calculation abilities, a modern graphics card obliterates all of those, achieving gigaflops speeds. On Sat, Sep 26, 2020 at 10:57 AM Jerry Gaffke via groups.io <jgaffke= [email protected]> wrote: The Steven Smith DSP book is also featured on his personal website, along |