Back Focus — for focus to sensor or just for aligning witness marks?

[Dana Neibert]

When you adjust the back focus on the camera (like on a shimmable PL mount), does that affect the focus of the image as it exits the lens and hits the sensor or is it just so your lens witness marks are on point? I am asking because I can't seem to figure out if I just have a lens resolution issue or if my mount needs to be shimmed as the image seems a little soft (8K on V-Raptor). I am pretty sure the set of lenses (Schneider FF) are spot on as they just came back from Duclos with a full CLA and when I put the camera three feet from the wall target (measuring from the film plane mark on the side of the camera) and focus and then look at the lens, it is dead on on the 3 feet mark. So is my softness a mount back focus issue (Kipperite PL Strata) or that's just the resolve of these lenses?

 

[Bill Ravens]

Yes, it affects where the focal plane lies. IE, in front of, or behind the sensor.

 

[Michael Hastings]

I think Phil attested that there are some image differences if back focus is not optimal given that the lens is designed for a specific flange distance, but I'm pretty sure these are extremely minor for minor back focus imperfection.

 

[Phil Holland]

Your Strata came with shims. There's a good chance you need to shim your mount for accurate back focus.

Xenons are based on a very old optical design, though with modern glass. That's why I call them sudo modern/vintage really. So not as sharp as say Sigma Cine or Tokina Cinema Vista. But you should be able to achieve critical focus. Schneider was good to send them out to spec, but back focus on the Strata PL plus the variable of the camera is what needs to be looked at likely. I shimmed all 4 of my Stratas and they were all off a decent bit.

 

[Patrick Tresch]

I think Phil attested that there are some image differences if back focus is not optimal given that the lens is designed for a specific flange distance, but I'm pretty sure these are extremely minor for minor back focus imperfection.

A lens can focus to infinity or further, when put closer to the sensor, or to macro when put further away from the sensor.
Back focus and lens shimming only affects the precision of the markings on the lens. There is no "image difference" of the lens if the focus is spot on even if it doesn't fit with the markings on the lens. If you set the focus to 2m and your object wich is at 2m is pin sharp then your backfocus (or the camera mount) and lens are shimmed perfectly.
You can affect the marking by shimming the lens or the mount (sensor position respective to the flange).
Best way is to callibrate first your sensor (or camera mount) with a Denz tool.
And then test if the lens markings are right and shimm them if necessary.

You can also check if your backfocus and lens are shimmed correctly for perfect focus if you use a wide angle (short lens) set to inf at the widest aperture possible.
Pat

 

[Phil Holland]

A lens can focus to infinity or further, when put closer to the sensor, or to macro when put further away from the sensor.
Back focus and lens shimming only affects the precision of the markings on the lens.
Pat

This is no longer (or never was) a correct mindset given the broad landscape of lenses. Though it works great with a host of stills lenses. It can impact the accuracy of scales and marks, which is also why those are adjustable on lenses as needed or dead to nuts on from the get go.

For instance, back focus being off at all on the mount and dead on the lens with anamorphic would lead to an image that will never achieve critical focus, sometimes nothing will ever get there if it's off enough.

Additional chaos comes in the form of some interesting modern lens designs ranging from the entire set of glass or specific focal lengths and how they project their light rays. This is why you see some weird scenarios where most of the lenses in a set achieve critical focus and then a wide angle or ultrawide being off.

Collimate the glass. Back focus the sensor. Every rental house should be able to provide this service if rental is the way. Owner/ops or production companies either pay for these services or have invested in the hardware to do these things themselves.

This most interestingly highlights the reason why a host of professional productions wouldn't typically rent online as piecing together equipment from various sources and not doing these checks in prep would lead to a severe situation if the desired standard is high.

Often things are never off enough to really screw a production, but when you get hit by it in the middle of a shoot, well I'll just say I've received many panic fueled calls regarding this over the years.

 

[Patrick Tresch]

This is no longer (or never was) a correct mindset given the broad landscape of lenses.

For instance, back focus being off at all on the mount and dead on the lens with anamorphic would lead to an image that will never achieve critical focus,

In my understanding the disitance between sensor and lens determine the focus of the object plane.
Be it shimmed or not you need to get to the standard done by the lens manufacturer (wich determines also the focal length) to get a picture in focus.
It can be 52mm as a standard but if it's shimmed for 51mm then you need to get the lens to 52mm to get Infinity focus. The marks won't add up on your lens though.
I don't see any loss of picture quality other than a harder job for the AC.

Could you explain what I don't get?
Thanks
Pat

 

[Phil Holland]

Could you explain what I don't get?

Convergence of optomechanically aimed lights rays designed to hit a focal plane at a very specific distance. In some cases having a detrimental impact on image quality, exaggeration of artifacts, impacting resolving power, and potentially across the field performance on spherical/rectilinear/planar optics. To the keen observer who has gone through this, some lenses have exaggerated vignette, very subtle. Or semi-bizarre/not-so-bizarre color casts to the vignette. Even a subtle impact on distortion in some cases can be measured and at worst seen.

Not surprisingly, performance on zooms can really take a hit if off.

Similarly some fisheye lenses can be very, very soft if off.

Anamorphic can be in a place where it doesn't perform anywhere near expected.

I would say in my opinion the optical hits are far, far more an issue than scales being off or on.

A well designed lens has scales tied to ideal optical performance or adjustable scales to cater to some additional variables. If the scales are dead on, once a lens is properly collimated and shimmed everything should like up nicely. If it doesn't, adjustable scales are needed.

Over the decades some lenses have been detuned by adjusting the air gaps/spacing between elements within the lens design. This will have typically an impact on focus roll-off, but very often produces different lens flares, vignette, etc. Exploring some of these concepts can take a high performance lens and create a subjectively more pretty image, which is the whole point. More complicated hyper corrected optics that use aspherical elements add some complexity to this depending on the design as well.

I could go deeper into this topic, but I'll just mention an interesting reality. 5 sets of Summilux-Cs that weren't all performing equally, though literally consecutive sets from the factory with the same "glass". To get them all to play correctly, collimating, and shimming the lens as well as tending to a consistent metric with the camera's back focus is what was required of all the tools to hit the same performance across cameras and with the different sets.

Rental houses deal with the back focus on cameras, most often they set that consistently across bodies and then collimate to spec with the glass, which will then work as expected across all cameras in inventory. In some cases in the past, Panavision integrated an optical slug to deal with some of the dance between digital optical paths versus film (OLPF and other glass) by offsetting that variance via optics rather than adjusting lenses back and forth.

It's amazing what 0.001" can do even.

If I was king of the world I would have a mechanism similar to DSMC2 cameras for back focus as well as integrated back focus adjustment on each lens. Shims are fine, but when you deal with a large variety of tools, it leaves a lot to be desired. Some companies get this, which is why you see integrated back focus on $35-$50K zooms and such. Though I would bet it would be a hard expense to endure on sub-$10K glass which is why shims are an affordable and decent option.

 

[Phil Holland]

I should add, if you take a look at extension tubes, which typically only work for a good quantity of spherical prime lenses effectively, it's a fun way to identify some of the issues that pertain to this and when/how you can get away with it. With the additional question mark of what happens when things go the other way at all, and that's a big part of the rub.

For instance I have some zooms that achieve decent resolving power with a tube, and allows for silly things like focusing on the front lens element, but throughout the range of that zoom it's a hornets next of what it can and can't do from there. A tube is maximizing one distance a lens could be off, but that magnifies (correct term) what is actually going on. Using that same strength tube with say anamorphic of fisheye lenses results in no usable image at all.

Things get weird when you go the other way as well.

 

[Les Hillis]

I've dabbled a bit with the whole back-focus and lens shimming 'thing' while messing around with some odd lens combinations and find it interesting.

Glad the DSMC1 & 2 camera's had adjustable sensors, even though I can see why it might have caused more problems than it solved.

Probably would have got one of those self-illuminated collimators RED used to make too if they hadn't been PL mount only.

Interesting that the most tweaking and adjusting I've done was in fact to an anamorphic lens set up, where all sorts of minor variables had to come together to get a decent image at all.


When it comes to adjusting back-focus and its effect on a lens, the way I see it, if a lens is designed to look a certain way, then strictly speaking, any change to the way it looks could be called a 'loss of image quality'.

I find it easier to understand though if I think of it as a 'change in image quality' rather than a 'loss of image quality'.

It also helps to think of the degree of changes taking place. For example, the changes could be on the level of an image not being able to focus at all on a subject at infinity or some other distance, or on the level of the difference between a lens designed to converge light rays at different emulsion layer distances on actual film, and one designed to converge the rays of light all at the same distance on a flat digital sensor.

Within the range of possible changes to the image resulting from inaccurate collimation or back-focus, some would be easily observable and generally considered a loss of image quality, while some would be harder to see at all or could even be considered a desirable gain for a certain look.

Another way to put these kinds of image quality concerns in perspective, is to consider the different optical qualities of a lens when it's focused at it's minimum focus distance and at infinity, as a result of the physical movement of the elements from one position to another other within the lens as focus is changed.

If I take two pictures where the subject stays at the same distance, but I change the back-focus, so that I then also have to change the physical placement of the lens elements inside the lens by adjusting the focus ring in order to keep the subject in focus, then I will end up with two different looking shots.

The extent to which they're subjectively and objectively different will vary, but they will be different, simply as a result of the different physical paths the light rays have taken through the lens to get to the sensor. At one extreme you go beyond the limits of the lens design itself, at the other, the changes from the 'ideal' settings make no real discernible or practical difference to the resulting image.

Most people using most cameras and lenses don't know or care about back-focus, or need to, as things are made to a tight enough specification to fit the end-use and just work as they're designed to. You either like the way the thing looks and works or you don't.

Beyond that comes increasingly more attention to detail and adjustability and precision, to the point where you're not just trying to get something in focus but are trying to elimininate (or introduce) or adjust other aspects of the image at the same time, which leads straight into lens designing itself (real and imagined).

It's an interesting area, with either none or a whole bunch of perfect solutions depending how you look at it.