Going, going..

Around 30 years ago, the digital signal processor burst upon the scece; the apparent answer to many an electronic engineer's prayers. Those early devices not only enabled new applications, they breathed new life into such exisitng systems as radar. Since those heady days, digital signal processing has revolutionised the world as we know it. Many of the most popular consumer products are based around digital signal processing; the technology is the foundation of the mobile communications network, to give just one example. But there is a key phrase here; it's digital signal processing, not the digital signal processor. Almost inexorably, the digital signal processor, as an entity, is disappearing. This, of course, is not a new theme. Many leading industry figures have debated this very point over the last few years. But it was interesting to see Gene Frantz, one of Texas Instruments' leading technologists, put on the record last year his belief that the digital signal processor as a deviceis dead. However, he believes fervently that digital signal processing as a technology is anything but. One of the technologies which has shown ther most potential in recent years for providing a platorm for digital signal processing is the fpga. Now, with Xilinx launching dedicated dsp development platforms, it looks as though the digital signal processor is taking another step towards retirement. Why, the reasong goes, would you want to use a standalone digital signal processor when you can perform the processing in an fpga fabric and get, if you believe the figures, up to 40 times the performance? Why indeed? Communications are digital, images are digital, music is digital, computing is digital: our lifestyles would collapse without digital signal processing, but not without digital signal processors.