The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

Why are sensors rectangular and not round?

GMB

Active member
I will probably be banned from this forum for asking such a stupid question :eek:

However, by having a rectangular sensor (3:2; 4:3; 1:1) the camera manufacturer imposes a crop of the round image produced by the lens on the photographer instead of letting him or her choose. I can see why film was rectangular, but why a digital sensor. The only reason that comes to my mind is costs--and that may already be sufficient.
 

Lars

Active member
hehe my first thought seeing the headline - stupid :D but it's a question that perhaps deserves to be asked.


I think if sensors were round then a lot of people would ask why sensors were not rectangular. Almost all presentation of information in almost all cultures builds on the rectangular shape.

It's very rare that the desired output format is not rectangular, so on a round sensor you would always have unused areas.

On top of that manufacturing cost for cameras would go up since they would have to be bigger, and sensor manufacturing would lose a lot in efficiency since each sensor would have to be bigger to cover the same rectangular image size, plus the lost space in between sensor circles.
 

dougpeterson

Workshop Member
Any given sensor you see is cut from a larger piece of material. If you cut circles there will be considerable wasted material in between each cut - lowering your yield and greatly increasing the price. Cutting into rectangles allows minimal waste between cuts.

Also each "row" of pixels on a CCD requires a readout. So a circle would mean that the top and bottom rows would have an entire row readout for a row of just a few pixels - very inefficient.

Of course if the demand had always been for circular images then I'm sure some entrepreneur would have found ways around the above issues somehow (profit is an amazing driver of innovation!) but as it is the easy technological/manufacturing/efficiency/cost way is rectangular and the demand is for rectangular images).
 

dougpeterson

Workshop Member
I guess any tessellating shape could be used with relative effeciency.

I'm going to petition Kodak and Dalsa to make tessellating-puppy-shaped sensors.

 

Bob

Administrator
Staff member
Sensors are manufactured on silicon wafers which are round.
Rectangular sensors are sawed from the wafer. Sawing round things might be done, but it would require some other technology such as high pressure water jet.
Say, for instance, that one might want a typical 4:3 sensor as a "nominal size" which would take up 12 "units" of the wafer. Should a round sensor be desired, it would take the equivalent square real estate of 19.6 units, or be more than 64% more. I say more than, since defects limit yeild, and the distribution of defects is random, so as sensors get larger, their probability of being useless due to a wafer defect increases, so not only does the sensor take up more silicon, its probability of being defective increases as well.
Perhaps the wafer area penalty might be reduced somewhat if a hexagonal closest packed structure were used, which is not compatible with current scribe and break methods for die separation, but assuming that it was, it would reduce the amount of wasted wafer area, but probably not enough to offset the increased cost of die separation.
-bob
 

Paratom

Well-known member
Because screens, frames (for the wall), magazines and posters are not round and the camera companies trust that photographers are able to frame when taking the image.
 

Bernard

Member
The obvious answer is to use the whole silicon wafer. An 8 inch wafer would fit in the back of an 8x10 view camera! Most 4x5 lenses will cover an 8 inch circle, which means you could do your camera movements in post (other than tilt, of course).

Given the resolution, the odd sensor defect shouldn't be a huge problem. Plenty of adjacent pixels to clone.
 

sizifo

New member
Good question!

One may also ask, why do sensors, and in fact film, have to lie in a plane?

Incidentally, I recently met a hungarian artist, Attila-Csorgo, who has designed cameras that take photos on curved films, with the film assembled as a mobius strip, and in various non-trivial geometric configurations. Here are some links:

http://www.translocal.org/images/new/Attila-Csorgo,-Orange-Sphere,-2004.jpg
http://www.c3.hu/events/2005/aegina/iq_csorgoattila_en.html

This is brilliant stuff, although I imagine it may not be obvious how cool it is from 2-dimensional webpages.
 
Last edited:
Top