Random Observation on Orientation Sensors

According to the US Patent Office a photographic camera provided with an orientation detector includes first and second elements that move relative to one another under the influence of gravity, and a signal device that responds to magnetic fields between the first and second elements to represent camera orientation. One of the elements is magnetized and the other is a Hall-effect sensor that responds to magnetic fields to produce an electrical output indicative of camera orientation. The signal device includes a recording mechanism responsive to the electrical signal for recording the camera orientation. ; Starting with the D2H in December of 2003- Nikon DSLR's have included an orientation sensors that record whether each image was captured in landscape, portrait left, or portrait right orientation by writing a flag into the EXIF header of the image (or RAW header). This allows the camera, and computer software to display the image in the correct orientation when it is displayed on screen.

I find it kind of interesting that the force on some of the fast paced thrill rides that Disney warn us 'include dramatic drops, turns, swoops, and lifts' can actually confuse the sensor in my camera. ; Which way is up? ; Or better yet- which way is down? ; My D700 obviously has no idea in the picture shown below. ; I think somebody needs to find something a lot more reliable than gravity to figure this all out for the next generation of cameras.

 

But what if I want to take a verticle picture of the Yeti?

 

There's no practical system beyond using gravity as a reference. ; If you had a specific application that involved unavoidable accelerations sufficient to throw off an acceleration sensor you COULD use something like a piezo gyro (probably what lens or camera stabilization systems use), but that would drift over time and you'd have to keep calibrating it to the horizon.

But really, what possible application could cause you to be regularly shooting while your camera is being accelerated that hard?


I guess cameras that have in body-stabilization might be halfway there already.. I have no idea what they use to detect motion, but if it's a gyro type of system then you'd just need to add one more, to give it three axis coverage, and then tie all that in to a system that can leave the sensors active and constantly calculate a reference to the horizon. ; But.. you'd have to start the system up and tell it what level the horizon is at at the start of the sensor active period. ; I can think of one other possible benefit of such a system, if you're prone to taking tilted shots like I am you could have the camera recording reasonably reliable horizon references so you'd know how much to rotate the image to level it.


Regarding the method the cameras measure gravity.. I suspect the system has changed. ; My first Dslr, the Digital Rebel, certainly seemed to use a system that used something like a lever that could flop down and indicate where "down" was, you could hear it clicking around inside if you tilted the camera. ; I actually saw a post on a forum by a user concerned that something was broken inside his camera because he could hear that rattle. ; But the newer cameras that I've used don't seem to have that sound. ; I'm guessing they've switched to something closer to what the Nintendo Wii controller uses, which I THINK operates by sensing the force being applied to a mass on the end of a tiny lever, probably one sensor for each axis. ; But just the force, I don't think it really moves much, it seems to be very solid, you can't hear any rattle in the Wii controller.

 

Great explanation of the way the sensors work Jeff. I like Roger's suggestion. ;

 

Hmm.. if you were using a square sensor I suppose you could have some sort of computer logic working to try to keep the brighter portion up...

It's kind of like the first rule of flying... ; keep the big blue thing pointed up.