Building cameras

So, the Oct 30th RED announcement has been and gone.  RED are clearly working very hard but they have also clearly been extraordinarily ambitious.  Perhaps too ambitious?

I look forward to hiring one of those RED cameras.  But I now know for sure that I can't afford to buy the Scarlet S35 that I was so hoping I'd be able to buy.  The S35 brain price hasn't been announced but the module prices alone put a working S35 camera out of my budget.

This is a real shame.  But, interestingly, it leaves a gaping hole in the market.  vDSLRs have created a huge expectation for low cost, large-sensor video cameras.

So.  The gauntlet is down to make a camera which will take video enabled DSLRs to a point where they're usable.  Who will step in?  Canon?  Nikon?  Panasonic?  Perhaps Silicon Imaging will surprise us all?

Look.  I know nothing about making cameras.  But I like to day dream.  Here are some idle musings I had on the night bus whilst catching up on the RED announcments:

1) Keep it simple.   S35 sensor.  Compressed RAW (or 10-bit log RGB).  Dual XLR inputs (maybe S/PDIF if you're feeling fancy).  No options.  Just one model, with one clear purpose and one price.

2) Embrace the rapid development model.  Use as many off-the-shelf components as possible.  Sensor?  Off the shelf from Sony of Fuji or Kodak or someone. CPU?  Atom.  Compression ASIC?  Don't waste millions developing an ASIC, instead use a mobile GPU or two to do the number crunching.  The new nVidia Fermi GPUs are capable of over a trillion single-precision (32-bit) operations per second. Codec?  Cineform RAW (you don't even have to spend time developing any software!)  LCD?  Don't even bother making your own - just let people buy from the market.  OS?  Android?  Let 3rd party developers create modules for the OS (intervalometers, wireless triggers etc).  The camera will basically be a computer so provide a few EtherExpress ports for storage and add-on modules.  Build in WiFi.  No need to build your own remote; let 3rd party developers create iPhone / Android apps for smart phones which will talk to the camera body over WiFi.  Yes the camera will chew through batteries but who cares?  Completely open source the camera's OS (just look at the enthusiasm with which skilled folks have developed their own firmware for the 5D) and provide cash prizes for developers.

I can dream, can't I?

I certainly don't want to belittle RED's incredible achievements.  But, from my perspective of a cash-strapped filmmaker, what RED are doing is analogous to the NASA moon missions: awe inspiring and technically phenomenal.  But I can't afford a moon rocket.  I just want a fast train.  Crap analogy; but you get where I'm going with it, right?

Add new comment