AF microadjustment tricks

BirdPhotographers.net

Help Support BirdPhotographers.net:

Arash,

Pausing Kaspersky Antivirus solved the problem. Odd that the error notice looked liked Windows, didn't mention Kaspersky, but that was it. Thanks again!
 
Hi Jim,

One other issue is that if AF sensor is slightly miss aligned with respect to the beam splitter it will give soft images no matter what you do, you can verify this by using live view AF at different distances to see if it can focus properly.

Best

Jeez this sounds like what I have been battling. I sent my rig in to be zeroed and when I got it back did some near and far testing for MA and everything looked gret at a setting of zero. However all shots were OOF. Only through testing in the field and adjusting everyday till I got a value of +9 has everything come into focus. But in servo on moving subjects I'm lucky to see one in 25 real sharp. But I have always been seeing different results for different distances.

Arash you've done us all a great service with posting this--and a nice thorough job of it. Where did you read up on this or is it your own concoction? I think if I knew more on the working intracacies of AF I'd be better able to diagnose things.

Paul
 
At first glance, this method looks OK. But, there is a BIG weakness. Suppose you are checking the focus of your 400mm f/5.6 lens. At f/5.6, there is a fair amount of depth of field at 33 feet distance from the target (the recommended distance for a 400mm lens). You could look at a flat target and see that your lens is in sharp focus. But what you cannot tell from a flat target is how much is in focus in front of the target and how much in back of the target. In other words, within the depth of field zone, where is your lens focusing?

I use the LensAlign system which incorporates a flat focusing target and an angled ruler. The camera focuses on the flat target and the ruler shows you very clearly where your depth of field focus "zone" lies. Having just done this with fellow Canon shooter, we saw his lens was in complete focus on the target. But most of the depth of field "zone" was in front of the target. Remember the rule of thumb, one third of the zone in front and two thirds behind. Using the LensAlign, we were able to calibrate precisely his 400mm lens on his 7D with one third in focus in front of the target and two thirds behind the target. Almost like focusing on an imaginary flying bird.:D

You would encounter this same issue on all big lenses - 300, 400, 500, 600, and 800. Don't you want to control the exact position of your depth of field? A flat focusing target will not show you this.

I do not work for LensAlign. I paid for my LensAlign just like everyone else. I like it because it works and I can repeat and reproduce the results accurately every time. It is not cheap. But neither are the wonderful big lenses you own!
 
At first glance, this method looks OK. But, there is a BIG weakness. Suppose you are checking the focus of your 400mm f/5.6 lens. At f/5.6, there is a fair amount of depth of field at 33 feet distance from the target (the recommended distance for a 400mm lens). You could look at a flat target and see that your lens is in sharp focus. But what you cannot tell from a flat target is how much is in focus in front of the target and how much in back of the target. In other words, within the depth of field zone, where is your lens focusing?

I use the LensAlign system which incorporates a flat focusing target and an angled ruler. The camera focuses on the flat target and the ruler shows you very clearly where your depth of field focus "zone" lies. Having just done this with fellow Canon shooter, we saw his lens was in complete focus on the target. But most of the depth of field "zone" was in front of the target. Remember the rule of thumb, one third of the zone in front and two thirds behind. Using the LensAlign, we were able to calibrate precisely his 400mm lens on his 7D with one third in focus in front of the target and two thirds behind the target. Almost like focusing on an imaginary flying bird.:D

You would encounter this same issue on all big lenses - 300, 400, 500, 600, and 800. Don't you want to control the exact position of your depth of field? A flat focusing target will not show you this.

I do not work for LensAlign. I paid for my LensAlign just like everyone else. I like it because it works and I can repeat and reproduce the results accurately every time. It is not cheap. But neither are the wonderful big lenses you own!


The purpose of MA is to align the sharpest focus plane with the precise location of subject, You can easily tell the location of focus plane within dof by using (>) or buttons (<), just click until you go oof each way and you can tell how many clicks you have in the front or back, more accurate than trying to read the cm scale on a ruler in a photograph. You can also then shift the focus plane within dof to your desired location while trying to maintain an acceptable sharpness if that is what you want to do although I don't see the point as long as the sharpest plane is on the intended subject. If the sharpest plane is too froward or too backward in the dof zone, MA cannot fix this issue and one of the optical elements in the lens is misaligned, although I have never seen a lens like this.

Also you can use this method with any target not just a flat target, it doesn't really matter what target you use... it does matter for target be parallel to sensor though. The method is not about the target but by how you choose what value to dial in. The scale on a ruler does not correspond to the MA values in the camera menu, so you have to dial in values randomly and find the best by trial and error. Also the 200% magnified live video feed from sensor makes it a lot easier to judge edge sharpness than using camera's LCD, and the remote control operation allows for very fine stepping of the servo, not possible if you try to manually focus using the focus ring on the lens.

Any ways, of course you should use whatever method you like and feel more comfortable with that works for you.
 
Last edited:
Jeez this sounds like what I have been battling. I sent my rig in to be zeroed and when I got it back did some near and far testing for MA and everything looked gret at a setting of zero. However all shots were OOF. Only through testing in the field and adjusting everyday till I got a value of +9 has everything come into focus. But in servo on moving subjects I'm lucky to see one in 25 real sharp. But I have always been seeing different results for different distances.

Arash you've done us all a great service with posting this--and a nice thorough job of it. Where did you read up on this or is it your own concoction? I think if I knew more on the working intracacies of AF I'd be better able to diagnose things.

Paul

Hi Paul,

I just came up with this when I was playing around with EOS utility, found that works well! As I mentioned MA will not help with random AF misses or soft images during tracking in AI servo mode, you should always do adjustment and testing for stationary subjects on tripod to diagnose any problems.

Good luck :)
 
Arash, thank you very much for this thread! When focusing in live mode on my 7D I have observed that the center auto focus point is not focusing on the exact same location as when looking through the view finder. The difference is about half the width of the focus point. It is close, but not spot on. Not sure if this is normal and wondering if others have found the same thing. Ken
 
Arash, thank you very much for this thread! When focusing in live mode on my 7D I have observed that the center auto focus point is not focusing on the exact same location as when looking through the view finder. The difference is about half the width of the focus point. It is close, but not spot on. Not sure if this is normal and wondering if others have found the same thing. Ken

Not sure I understand what you're saying, the active sensor is indicated with a red frame in the LV screen, within the red frame sensor will pick any point that has the highest contrast, this is normal.

Also make sure 7D is not in spot AF mode when performing MA.
 
Thanks Arash, I did not realize that I should not be in spot AF mode. What I was getting at with the active sensor is when the camera is locked down with a tripod and I am looking at the red sensor in the view finder and compare it to the red sensor in the live view screen, they are covering a slightly different location. Ken
 
Thanks Arash, I did not realize that I should not be in spot AF mode. What I was getting at with the active sensor is when the camera is locked down with a tripod and I am looking at the red sensor in the view finder and compare it to the red sensor in the live view screen, they are covering a slightly different location. Ken

I see what you mean, how much of a difference do you see? The focusing screen may not be aligned then, which is strange given that 7D is calmied to have 100% finder coverage. I will chek with mine later.
 
Arash, I assumed the difference probably had something to do with the registration of the transmissive lcd screen in the viewfinder. The difference between the focus point in the viewfinder and the focusing point on the LV screen is about 2 inches with my 300mm lens at 75 feet. I am probably nit-picking but I was curious if others see the same thing. This could definitely be problematic if it was any more. Ken
 
Arash, I assumed the difference probably had something to do with the registration of the transmissive lcd screen in the viewfinder. The difference between the focus point in the viewfinder and the focusing point on the LV screen is about 2 inches with my 300mm lens at 75 feet. I am probably nit-picking but I was curious if others see the same thing. This could definitely be problematic if it was any more. Ken

yes Ken, it's the transmissive LCD screen that is right behind focusing screen. 2" can cause issues, I will check with mine tonight and let you know. Good catch!!!
 
Hi Ken,
I tested my 7D, the AF indicator box on the transmissive LCD screen seems to be perfectly aligned with the AF sensor, at least I can't measure any error, the target box is 5mm X 6mm approximately, I centered the AF box in the finder and then superimposed the sensor in DPP.

Are you sure your were looking straight through the finder? (you should look completely perpendicular to the eye piece, if you look at angle you will see a different area under the AF box. This may not be easy if you wear glasses). I suggest trying again, more carefully this time, if you see misalignment I would send the camera to Canon for adjustment.

Best
 

Attachments

  • 7D-AFsensor.jpg
    7D-AFsensor.jpg
    88.9 KB
Last edited:
Don,

I used this method in Combo with the LensAlign. Basically making sure the LensAlign target we dead on first and then checking the ruler for front/back focus.

Be careful not to assume a 1/3 and 2/3 front and back. Use the the LensAlign calculator to show DOF as with a 300 at the recommended test distance it is nearly 50/50

Lou
 
Hi Ken,
I tested my 7D, the AF indicator box on the transmissive LCD screen seems to be perfectly aligned with the AF sensor, at least I can't measure any error, the target box is 5mm X 6mm approximately, I centered the AF box in the finder and then superimposed the sensor in DPP.

Are you sure your were looking straight through the finder? (you should look completely perpendicular to the eye piece, if you look at angle you will see a different area under the AF box. This may not be easy if you wear glasses). I suggest trying again, more carefully this time, if you see misalignment I would send the camera to Canon for adjustment.

Best
Thanks Arash, I checked my camera again to confirm the alignment. I agree that you must have your eye centered in the viewfinder as there is a significant amout of parrallax if you are off to one side. My test confirmed that there is a discrepency in the target location of the focus point when comparing the viewfinder with LV. The best that I could determine is that the difference between the two is 1.25 inches with my 300mm lens at 50 ft. For my test I used only the certer focus point. Again, I may be nit picking but I could see how this might be an issue when attemting to auto-focus on something such as a birds eye.
 
Some notes on AF alignment:

There are 3 positions, of which 2 have been discussed:

1) The true position of the AF sensors
2) The red rectangle in the viewfinder, and
3) The rectangle superimposed on the image (by software).

Only # 2 and 3 have been addressed in this thread so far and neither are necessarily the the true location of the AF area. When light from the lens enters the camera, it first encounters the reflex mirror. That mirror is really a beam splitter reflecting some light up and into the pentaprism and to our eye, and some light is transmitted through the mirror. Behind the reflex mirror is another mirror which reflects the light down to the bottom of the camera where the AF sensors are located. The red rectangles we see in the viewfinder are in the pentaprism section, far from the actual AF sensors.

The true location of the AF sensors is controlled by the small mirror behind the main reflex mirror and the placement of the AF module on the floor of the camera box. The red squares we see through the viewfinder are only an approximation of the true position, as is the software generated rectangle. There is no guarantee that any or all of these will line up perfectly.

The only way to test AF alignment is to use a high contrast target surrounded by very low contrast area. An isolated light bulb (e.g. a street light at night), the Moon, or some white paper with a small square grid. Move the lens just slightly out of focus and place the AF point (one AF point at a time) near the red rectancle and note where the camera will focus. Start well away from the AF point and try focusing. Move closer and closer to you test target until the AF system locks on. Note that position. Approach the AF point from all four directions marking where the camera will lock on to your target. The true AF area will likely be a little larger than the red rectangle.

Roger
 
Some notes on AF alignment:

There are 3 positions, of which 2 have been discussed:

1) The true position of the AF sensors
2) The red rectangle in the viewfinder, and
3) The rectangle superimposed on the image (by software).

Only # 2 and 3 have been addressed in this thread so far and neither are necessarily the the true location of the AF area. When light from the lens enters the camera, it first encounters the reflex mirror. That mirror is really a beam splitter reflecting some light up and into the pentaprism and to our eye, and some light is transmitted through the mirror. Behind the reflex mirror is another mirror which reflects the light down to the bottom of the camera where the AF sensors are located. The red rectangles we see in the viewfinder are in the pentaprism section, far from the actual AF sensors.

The true location of the AF sensors is controlled by the small mirror behind the main reflex mirror and the placement of the AF module on the floor of the camera box. The red squares we see through the viewfinder are only an approximation of the true position, as is the software generated rectangle. There is no guarantee that any or all of these will line up perfectly.

The only way to test AF alignment is to use a high contrast target surrounded by very low contrast area. An isolated light bulb (e.g. a street light at night), the Moon, or some white paper with a small square grid. Move the lens just slightly out of focus and place the AF point (one AF point at a time) near the red rectancle and note where the camera will focus. Start well away from the AF point and try focusing. Move closer and closer to you test target until the AF system locks on. Note that position. Approach the AF point from all four directions marking where the camera will lock on to your target. The true AF area will likely be a little larger than the red rectangle.

Roger
Thank you very much for clarifying this issue. I now have a much better understanding of what I am seeing. Your post makes me realize how little I know about my cameras. Ken
 
Thanks Arash, I checked my camera again to confirm the alignment. I agree that you must have your eye centered in the viewfinder as there is a significant amout of parrallax if you are off to one side. My test confirmed that there is a discrepency in the target location of the focus point when comparing the viewfinder with LV. The best that I could determine is that the difference between the two is 1.25 inches with my 300mm lens at 50 ft. For my test I used only the certer focus point. Again, I may be nit picking but I could see how this might be an issue when attemting to auto-focus on something such as a birds eye.


Hi Ken,

I checked the AF indicator positions again, this time using a laser beam in the lab for each AF sensor, they all seems to be perfectly aligned with the respective frame in the finder which is not surprising to me at all, given that camera is a precision device. If I were you I would send the camera to Canon for calibration. I am sure they will notice this. BTW, It's not nit picking at all, 7D has spot AF mode for a reason, sure will cause issues in certain cases!


Good luck,
 
Last edited:
The only way to test AF alignment is to use a high contrast target surrounded by very low contrast area. An isolated light bulb (e.g. a street light at night), the Moon, or some white paper with a small square grid. Move the lens just slightly out of focus and place the AF point (one AF point at a time) near the red rectancle and note where the camera will focus. Start well away from the AF point and try focusing. Move closer and closer to you test target until the AF system locks on. Note that position. Approach the AF point from all four directions marking where the camera will lock on to your target. The true AF area will likely be a little larger than the red rectangle.

I checked several cameras using a LED light in a dark room. Each camera would get a good focus lock about half the rectangle width beyond the red rectangle in all directions. Also, a hunting focus at one width beyond the red rectangle would sometimes catch focus. I observed similar behavior on every camera I've tested, including 20D, 40D, 1D Mark II, 5D Mark II. In each case, the red rectangle seemed centered on the AF region, but the AF region is about double the size of the red rectangle.

Roger
 
Thanks Arash,

This is a very useful thread, I just adjusted all of my lenses with 7D, I only wish that my 1D MKIIN also had micro adjust capability :( BTW, I was wondering if this should be done is spot mode since I use this mode for perched birds in my backyard setup, I usually put the sensor on the eye.

Thanks again

Dan
 
Thanks Arash,

This is a very useful thread, I just adjusted all of my lenses with 7D, I only wish that my 1D MKIIN also had micro adjust capability :( BTW, I was wondering if this should be done is spot mode since I use this mode for perched birds in my backyard setup, I usually put the sensor on the eye.

Thanks again

Dan


Hi Dan,

Glad you found it useful! In spot AF mode 7D samples a smaller area of the AF sensor indicated by the smaller inner square in the finder, ideally you would want to adjust for the reading that is from the entire AF sensor area. The problem with spot mode is that it does bot perform well in AI-servo mode when tracking birds in flight, and is not recommended by Canon for moving subjects. It works better espeically in one shot AF and for perched birds, so if that is what you shoot most of the time it is OK to do MA in spot mode. for AI-servo best is to use and MA for full size sensor.

Best,
Arash
 
Last edited:

Latest posts

Back
Top