Last month, a friend of mine invited me around his house to see how his latest restoration project was progressing. In the corner of his living room was a guitar amplifier. Being an avid fan of old technology, I asked him to remove the back panel of the unit in which sat four old tubes (or valves, as they are called in England).
Unfortunately, the unit needed much work. One of the tubes had expired and it was not hard to discern that the heating element had burnt out. My friend was not discouraged having ordered the relevant replacement from an online store that specialized in replacement tubes that were made in Russia.
Without this tube, the amplifier was still operational and he demonstrated how, playing a Gibson Les Paul, it’s warm, natural tone sounded better that any transistor amplifier on the market today. I was impressed. While transistor amplifiers may be of lighter weight and less expensive that their elderly cousins, the sound that they produce is not even comparable. Certainly, while newer technology has its benefits, the cognoscenti question these developments.
In the machine vision industry, the announcement that Sony has been rumored to discontinue its line of CCD sensors has caused a similar furor. While many have compared this announcement to the replacement of the CRT by solid-state LCD flat panel TVs as the inevitable progress of technology, nothing, in fact, could be further from the truth.
Since its emergence, the machine vision industry has leveraged developments made by those designing products for consumers. Whether this was the broadcast standards developed in the 1930s for television, those to interface peripherals to personal computers or the operating systems and development tools used in many machine vision systems, the story has remained much the same. Without such developments, the technology used in high-speed inspection systems in use today would not be possible.
Unfortunately, relying on these developments can have its drawbacks. If Sony does stop producing its line of CCD imagers, it will be because the company wishes to concentrate on the billion dollar consumer market where over 80% of CMOS devices are used. In machine vision applications such as low-light level, near infrared and UV applications, however, the performance of CCDs may be better suited than CMOS-based imagers. Luckily, numerous manufacturers of CCD devices and cameras are still committed to supporting CCD technology.
Just as the emergence of the transistor heralded the decline of tube-based designs, the emergence of high-performance CMOS imagers will certainly impact the use of CCDs in machine vision systems. However, there will always be a place – perhaps even a nostalgic one - for the invention that won Willard Boyle and George Smith a Nobel prize for physics.