Lenrentals does the mother of all flange distance tests

Having spent a considerable amount of time the last two months on testing I can now certainly appreciate the work that went into testing the flange distance across 2,500 cinema cameras.  Lensrentals goes into why this is important for cinema, and the reasons are very important.

1. The cinema lens distance scales are very precise unlike the ones on regular still camera lenses (one of the reasons for the cost increase)
2. Cinema cameras have a hard stop at infinity so if your flange to sensor distance is too long, you may not be able to achieve focus at infinity.

The first problem I think is more of a problem, while the second is more of a perfectionist.  Perhaps with 8K and the future of 16K this is more problematic, but on 4K and lower, your calculated hyperfocal distance would be so large that the lens set to "near infinity" would still achieve infinity focus unless it was way off.

That all being said, this is a measure of how good the mechanical engineering of the cameras are and how well these cameras are quality controlled from the factory.

With the costs of cinema cameras usually much higher than that of mirrorless or DSLR stills cameras, you'd expect a greater degree of accuracy from the camera.  Especially for an industry that does use manual focus and those distance scales quite frequently.

So how well did Canon do? Quite well actually.  With a sampling of a few hundred Canon Cinema cameras, lensrentals got a very well controlled graph.  Once they removed the 3 outliers of which 2 were because they were sent in for service, and another in which they assume the mount was damaged, they ended up with a very normal looking graph;

The chart axis is a bit confusing.  if the number is positive, it's the amount of shimming you need to add to get the camera and lens to be correct.  So a positive number is an offset closer to the sensor.  A negative number here is bad.  because you can't add a negative shim. 

In other words, it appears as if Canon errs on the side of caution when it comes to the calibration of the cinema cameras, perhaps allowing for thermal expansion.

There is one element to this that isn't tested - which I would consider as big of a problem, if not more.  I would suspect that flange to sensor tilt would be a greater concern.  For instance, this is when one side is closer to the sensor than the other side.  So while in the center (which these measurements are taken at) may be perfect, the edges may not be.  This would of course change the focal plane, as the focal plane is parallel with the sensor. We of course assume that the flange is also parallel to the sensor.

I'm certainly looking forward to the part 2 of this blog article which shows how our cameras fair.

Read lensrentals article here