Tuesday, November 18, 2008

Old-school image processing, or Moon pictures remastered

Before anyone set foot on the moon, it was considered important to survey the place. To this end, NASA sent probes fitted with cameras and the ability to beam back pictures. There being no CCDs at the time, the probes actually used film, developed it and scanned the resulting prints for transmission. The transmissions were recorded on magnetic tape as a backup, thanks to the foresight and efforts of Charles J. Byrne; the preferred mode of storage was to print the pictures and store the prints. What most of us saw the first time around, including the famous Earthrise image, was actually third-generation material: reproductions of photos of those prints. [Or fourth-generation if you count the original film up in space.]

Twenty or so years later, Planetary Data System co-founder Nancy Evans, then at JPL, took the tapes into her care and started a project with Mark Nelson to find and refurbish tape drives that could read the old tapes. The project stalled for lack of funds. Evans retired from JPL to become a veterinarian and stored the tape drives in her garage.

Another twenty or so years later, Evans retired as a veterinarian and went looking for someone to take the drives off her hands and, hopefully, put them to their intended use. Dennis Wingo and Keith Cowing took on the job, moved the drives into a disused McDonalds in NASA Ames' research park and set to getting them working again. This involved a lot of cleaning, swapping of parts and working with circuits whose components were actually large enough to see and handle. It took them 99 days, but they got the thing working.

Even better, the results are now on the web, as is the more complete account I'm summarizing.

The web has acquired another significant chunk of history -- the digital images the probes would have sent back if they could have, and if there had been any place to put them.

Most definitely a neat hack.

1 comment:

David Hull said...

Note to self: double check "print" vs. "negative"