For the past few weeks I’ve been taking a lot of test shots with my 1DMkIII, which was just sent back from Canon after having the sub-mirror fix done. Mine wasn’t exhibiting a whole lot of focusing problems before the fix, but now the focusing seems to be just a little more accurate and about as good as I think it’s going to get. Canon only calibrated mine – they didn’t replace any parts according to the repair ticket. I’m just glad it was checked out and given a clean bill of health. Everything was working fine until… dun dun duuuunnnn: This weekend’s…Cold weather. Now, “technically” according to the specs for the Canon 1D mkIII the camera should only be operational down to the freezing mark – so I really shouldn’t have anything to complain about:
“Operating temperature range: 32 degrees to 113 degrees F (0 degrees to 45 degrees C); 85% or lower relative humidity.”
I had lurched out of bed well before dawn on Saturday to track down some hawks, buffleheads and hooded mergansers. It was admittedly chilly, and slightly below the quoted “Operating Temperature” range – but nothing tremendously frigid. At 7am it was roughly 22°, and by 11am it had warmed up to almost 30°. While chimping some of my shots, I started to notice something pretty weird going on. I was shocked to see my images showing up on the LCD looking like this:
![]()
7:40 am – an icy Goldeneye. The cold problems begin…
At first I thought it was some kind of fluke – maybe the cold weather was affecting my CF card or the LCD screen was going goofy because of the temperature. So I switched cards and held my palm over the LCD for a few minutes to try and warm it up… same results. “Meh”, I shrugged. “Nothing I can do about it now…” So I kept shooting with the 1D – alternating with the 5d here and there. On the LCD, most shots didn’t look too bad. At one point I thought it had only affected a few, and maybe I could just clone out some of the crap that was happening to them. To my utter dismay, when I returned home and started to process my images, *every single image* from the 1d looked like the one above – or worse. The images from the 5d however, were perfectly fine. No problems whatsoever. I was pissed. There were some great shots that were ruined, but I could only blame myself for continuing to shoot with it.
My anger quickly turned to curiosity, and I started testing things to try and figure out what the hell was going on. After letting the camera warm up for a while in the house, everything was back to normal again. Images were coming out just fine. Still puzzled, I took the camera out to the back porch and let it sit for about 15 minutes, then brought it inside and took some test shots. Same screwy results. Those weird purple, white and green “digital grid lines” were everywhere. I tried different lenses, different CF/SD cards, different batteries – nothing helped. I noticed that as the camera body warmed up the lines would only start showing up towards the top of the images. Ok, I figured… today was really cold. Too much cold is bad. Tomorrow is supposed to be warmer, around 40°, so I’ll test it out in the car with the heat on and see what happens. Should be fine, right? Well, not really.
Sunday morning – (9am~35°, 11am~39° Humidity: 60%)
I was shooting from inside the car with the passenger side window completely rolled down and the heat intermittently on high to warm me up. There was a slight breeze. The 1d never felt more than slightly cool to the touch, and nothing out of the ordinary was showing up on the LCD when I checked some of the shots. I thought everything was going to be fine, but when I got home and started looking at the images again there were still some weird digital green lines on a few shots, just like the ones below. WTF! It had to have been at least 50-60° inside the car!
Back at home, out on the porch… more tests with the same results.
This afternoon: A toasty 41°. Partly cloudy with light wind from the WNW. Relative humidity: 50% On the porch for 10 minutes. Results? You tell me. What the hell is this?
In the fridge you go! I know it’s not lower than 32° in there!
To make a long story short, it’s going back to Canon ASAP to see what the hell is wrong with it this time. Hopefully they’ll fix it. What do I think it is? Aliens? I honestly don’t know, but I think it’s got something to do with the sensor. My best guess is that the sensor (or sensor connector) is contracting ever-so-slightly in cold temperatures *just* enough to separate itself from whatever type of conductors/connectors it has, resulting in the data getting scrambled and generating the weird digital interference on the images.
You can definitely see in the sample images and on a lot of the other ones I shot that it isn’t some random pattern. They’re digital lines, like lines on a grid, spaced about 8 pixels across when you zoom in to 400%. It almost reminds me of wiggling the video cable on a computer monitor – same effect, although this is undoubtedly temperature related and digital. Anything below 45°, and things start to get messy. With some brief research, I was able to find a good number of people on several forums claiming to have used their MkIII’s and other SLR’s in sub-freezing temps without a problem, and uh, I would assume that it would have to be below freezing for folks shooting skiing like this. My trusty 5d was performing flawlessly alongside the MarkIII when this started happening, so I’m hoping this really is just some random image sensor fluke that they’ll be able to fix quickly and send me on my way.
One thing is for damn sure: I didn’t drop $4500 bucks on a camera that can only be used for half of the year it’s above 50° around here. And as I said to my extremely understanding wife tonight – “That camera could be absolutely fantastic… it fits like a glove in your hand, the frame rate is astounding, it has impeccable noise control at high isos, and the battery just won’t die – I love using it… if only it could do the one thing it’s supposed to… produce a decent goddamn #^@&)!* image!” Frustrating, to say the least.
Update with the consensus and fix from Canon USA.
