Fujifilm Rumored To Be Incorporating In-Body-Image-Stabilization (IBIS) In Future Cameras; Can They Do It?

On July 3, 2017, Patrick, the admin and principle writer for the web site FujiRumors.com reported that he is 100% certain Fujifilm will incorporate in-body-image-stabilization (IBIS) into future cameras.  I was very surprised to read that.  To me, this is great news. I've been reading this site for the past few years and Patrick, in my opinion, has an excellent track record for accuracy.  Because of this, I think there is substance to this rumor.

I'm a big fan of having image stabilization built into the camera body.  IBIS becomes even more useful and much more effective when it is coupled with image stabilization built into compatible lenses as well.  With my Olympus E-M1 Mark II and Olympus 12-100mm f/4 PRO lens, Olympus says that 6.5 stops of stabilization is possible.  That is an amazing claim!  

Personally, I've made several images handheld, sharp and repeatedly at more than 1 second.  I haven't needed to push it more than that yet.  Still, an amazing technology that I would welcome for my Fuji gear.

Thinking about this revelation, I started to wonder just how Fuji will accomplish this.  In a past interview a Fuji engineer indicated that IBIS was not possible in X-series cameras as the lens mount flange and circle of image projection by their XF lenses was insufficient to accommodate IBIS.  How could then then accomplish this?

There are at least 5 questions that immediately I thought of that need to be answered:

Can a sensor shift mechanism be implemented in Fuji X-series cameras with current the electronic and mechanical architecture?
How far does a micro 4/3 sensor need to shift for best implementation?
In what directions does a sensor need to shift to handle all forms of movement?
Are the image projection circles of the Fuji XF lenses sufficient to accommodate the optimum amount of shift?
Is the diameter of the lens flange sufficient to accommodate optimum sensor shift?

Obviously, I don't have the answers to these questions but I thought I would do a little figuring just to see what may be possible.

I calculated the diagonal of the Fuji 24mp sensor is 28.28mm.  This was figure was derived by calculating the square root of the sum of the squares of the length and width of the sensor as published by Fuji.  What?  Here is how I arrived at that measurement.

(Here are some figures I found on the Fuji website.)

Fuji 24mp APS-C sized sensor measure
15.6mm x 23.6mm

A — 23.6mm²= 556.96mm

B — 15.6mm²= 243.36mm

C — 243.36 + 556.96 = 800.32 square mm

The square root of 800.32 = 28.28mm, 
which it the length of the diagonal of the sensor (C).

To cover then entire sensor with an image, specifically the corners, the minimum Fuji lens' circle of projection has to be just a tiny bit more than 28.28mm.

What is the image projection circle of Fuji lenses?

From their site and the MTF graphs, it appears to be 28.4mm (14.2mm X 2)

That gives a projection circle of 0.12mm more than the minimum required to cover the entire sensor, again, especially at the corners.  You must divide that figure in half to center the sensor within the projection circle.  That then provides 0.06mm on any corner to accommodate sensor shift.

I have no idea of how much shifting "sensor shift" on a micro 4/3 format camera requires, but if it is more than 0.06mm, then the current crop of XF lenses as well as how the cameras and sensors and resulting image size are currently engineered, it won't work. Also, remember that with IBIS, we are talking about 5-axis image stabilization in which the sensor shifts straight up/down, straight left/right, tilting forward/backward, tilting left/right, as well as rotational movements. To accommodate all of those directions, the sensor may require even more shifting.  I just don't know.

There is another potential issue, the Fuji X-Trans sensors have a much different RGB pixel layout than Bayer Array sensors.  That different layout may require more of a shift than a Bayer sensor since the distribution of RGB pixels is different.  It may require less.  I just don't know that as well.

Then think about what the Fuji engineer said about flange diameter.  The flange diameter may be just enough to cover the current design configuration of camera flange diameter, distance from flange to sensor and sensor size.  Any sensor shift may cut off the image by the flange itself blocking the focused rays of light.

Depending upon the specific issues, there may be three ways of incorporating IBIS in Fuji cameras.  First, if IBIS is mechanically possible, Fuji could develop some sort of computer algorithmic wizardry to handle light falloff in the corners or second, Fuji could slightly increase the size of the sensor to accommodate sensor shift and third, if mechanical limitations are a factor, Fuji could reduce the dimensions of the final image so the final image is slightly less than what we have now on their full APS-C sensor size.  In other words, the final image would be smaller than you would currently enjoy.  Of course, that increases the lenses "crop factor," or field of view, which I think better describes it. Instead of a 1.5x field of view, it may be necessary to increase it to a 1.6X  or 1.7X field of view.  Canon's APS-C sized sensor cameras are already 1.6X.

I believe some camera manufacturers currently crop final output size of their very high quality video when IBIS is employed in their cameras.  It seems to be an option for some. Of course, APS-C sized sensors are already on the smallish side so I would hate to lose any surface area at all.

Then, of course, there is magic!  LOL  Anyone have a spare magic wand?

Without knowing specifically the technical challenges involved, I'm hoping there is a way for Fuji to accomplish incorporating IBIS into their future APS-C sensored cameras without compromise.  IBIS is the only reason I can now think of that would cause me to leave Fuji sometime in the future.  IBIS may not be an issue for you, but I'm in my mid-60s and as time passes, I find the steadiness of my hands less reliable than when I was younger.

Where have I gone wrong?  I'm sure I've forgotten something or miscalculated in some fashion or left something important out.  If you find my logic or math faulty, please correct me.  It will be welcomed.

Thanks for looking. Enjoy! 

Dennis A. Mook 

All content on this blog is © 2013-2017 Dennis A. Mook. All Rights Reserved. Feel free to point to this blog from your website with full attribution. Permission may be granted for commercial use. Please contact Mr. Mook to discuss permission to reproduce the blog posts and/or images.