D800 & LensAlign

Started Apr 28, 2012 | Discussions
Christian Wagner Senior Member • Posts: 1,547
Re: D800 & LensAlign

Makes sense, and thanks for the explanation. It still seems an easy fix, I just need to place a cereal box or some target perpendicular to the 0 (or whatever the center point is on the pyramid) and I would accomplish the same thing as lensalign for substantially cheaper. Is that correct or am I missing something else as well? All I need to do is put a ruler perpendicular to the center line on the pyramid and put the cereal box flush with the ruler and then focus on the cereal box. Back focus the number behind the center point will be in focus, front focus the opposite, adjust until correct?

Eli S wrote:

But if the point is a little above the center, doesn't that mean that the af point is off and if i adjust it forward it will be right on?

I think you are confusing vertical and horizontal focus mark accuracy with front-and-back focus plane accuracy. Consider a slanted target:

A B C
1[2]3
X Y Z

in this case, what you are thinking, is if the mark is centered on "2" but actually focuses on "B" there is a back focus problem - which is not the case - what you may actually have is a focus mark that is off-center but still accurate in focus. In order to check exclusively for back or front focus, you need the entire target equidistant from the camera.

Which is why having a perpendicular target (Lens Align) is better than a slanted one (Focus Pyramid) because it precludes any focus mark shift and measures just the distance to target, and not any error in the focus sensor marking vs. actual position of the focus sensor.

-- hide signature --

Chris

Christian Wagner Senior Member • Posts: 1,547
Re: D800 & LensAlign

Thanks, I will use the cereal box as I already ordered. Given I'm spending 5K on lenses and a body, whats another $20, I guess the same thing could have been said about the $80 for lensalign!

Thanks again!

-- hide signature --

Chris

studio460
studio460 Contributing Member • Posts: 712
Re: D800 & LensAlign

The "CA" you're noticing is due to overexposure. The highlights are "blooming," which closely resembles the appearance of CA, which it is most likely, not (I know this because I've made the same erroneous conclusions with my own lens tests).

Oscarroos Senior Member • Posts: 2,914
Re: D800 & LensAlign

You used the lens align correctly but if I were in your shoes, I would take the 70-200 outside today nice and sunny now, and take some photos wide open of brick walls, whatever to see how the photos look. Based on your low shutter speed could be motion blur. Testing outside with high shutter speed should answer some of the doubt.

Where in Texas are you?

Oscar DFW area north Texas

Stangs55 wrote:

Robin Casady wrote:

The 70-200 @ f/2.8 looks like it has serious motion blur. Look at "Michael Tapes Design" and notice how the blur is all vertical. Self timer isn't sufficient. You need to use mirror lock up with at least a 3 second delay. Better to do it manually and wait longer after mirror up. You have to use a remote release cord or wireless.

What was the shutter speed for both exposures? If the shutter speed is long enough, mirror flap will be diminished because it is a smaller % of the exposure.

What tripod and head are you using?

2.8 at 1/10 sec
8 at 0.5 sec

If it was motion blur...why is it that when I stop it down to f/8 taking my shutter speed from 1/10 sec to 0.5 sec, does it get very sharp? If it was motion blur, the problem should have been exaggerated since all I did was rotate the command dial and take another shot...I just can't imagine mirror lock up improving the shot when going from 1/10 to 5/10 sec...

I'm using a Manfrotto tripod (not sure which one, but it's a heavy monster I've had for over a decade) and a Manfrotta 498RC2 ballhead.

 Oscarroos's gear list:Oscarroos's gear list
Nikon Coolpix 5700 Canon PowerShot S95 Nikon D200 Nikon D700 Nikon D800 +13 more
JediLight Forum Member • Posts: 98
Re: D800 & LensAlign

Deegee wrote:

As I delve deeper into this thread I get the over-riding impression that there are lot of people, (including potentially me for sure – I mean no absolutely offense, I am in the same boat!) who are in danger of making fundamental adjustments to gear that they have probably spent many tens of thousands of pounds on, based on pretty crude measurements that they have no experience of taking and obvious great difficulty in interpreting.

I totally agree with the above. I have just come across a software based solution - Reikan FoCal, that I downloaded and ran today and, for me, it has overcome my previous issues I have had with testing. Essentially, the software guides you through the process, collects and analyses a load of data and, for me therefore has given me proper confidence in the results.

I wrote up more of my observations in this thread here along with a screen shot of the convincing data the software is producing.
http://forums.dpreview.com/forums/readflat.asp?forum=1021&thread=41356064

 JediLight's gear list:JediLight's gear list
Nikon D80 Nikon D7000 Nikon D800 Nikon AF-S DX Nikkor 35mm F1.8G Nikon AF-S Nikkor 16-35mm F4G ED VR +2 more
PHXAZCRAIG
PHXAZCRAIG Forum Pro • Posts: 13,445
Re: D800 & LensAlign

A year or so ago I got Leonard to actually point me to an AF target he considered acceptable. Perhaps the world's only acceptable target? I suggested at the time that he simply link to that URL in his signature, but he never took me up on that.

So, with one exception that I know, Leonard persists in slamming every autofocus target anyone has posted, while neither giving an alternative nor acknowledging that AF systems might actually be accurately using some of these targets. I have the LensAlign, and my results with it have been repeatable and reliable, for at least the focus distance used when testing.

The 45-degree target indeed gave me lots of grief and inconsistent results when I got my D300 and tried to test lenses with it.

-- hide signature --
 PHXAZCRAIG's gear list:PHXAZCRAIG's gear list
Nikon D80 Nikon D200 Nikon D300 Nikon D700 Nikon 1 V1 +35 more
PHXAZCRAIG
PHXAZCRAIG Forum Pro • Posts: 13,445
Re: D800 & LensAlign

If you are bothered by the price of the LensAlign, and you think you'll only use it one to check all your lenses, just rent one from lensrentals.com and save money.

-- hide signature --
 PHXAZCRAIG's gear list:PHXAZCRAIG's gear list
Nikon D80 Nikon D200 Nikon D300 Nikon D700 Nikon 1 V1 +35 more
PHXAZCRAIG
PHXAZCRAIG Forum Pro • Posts: 13,445
Re: D800 & LensAlign

I'm suspicious of your 70-200 having an issue.

However, maybe that is normal for the version 1 of that lens? A friend of mine brought over his 70-200vr, which he bought used, and it didn't wow either one of us when we tested it. Certainly not wide open at 200mm. But we didn't know if that was supposed to be normal or not.

I never posted any of the test shots, but if you want, I'll try to dig them out and stick them up on a web page.

-- hide signature --
 PHXAZCRAIG's gear list:PHXAZCRAIG's gear list
Nikon D80 Nikon D200 Nikon D300 Nikon D700 Nikon 1 V1 +35 more
Eli S Forum Member • Posts: 62
Re: D800 & LensAlign

Yes, that's all you would need to determine back/front focus. Make sure the cereal box is exactly perpendicular to the camera, and perfectly aligned with your 0 mark.

BTW, do have an actual focus problem, or are you just making sure there isn't one?

Christian Wagner wrote:

Makes sense, and thanks for the explanation. It still seems an easy fix, I just need to place a cereal box or some target perpendicular to the 0 (or whatever the center point is on the pyramid) and I would accomplish the same thing as lensalign for substantially cheaper. Is that correct or am I missing something else as well? All I need to do is put a ruler perpendicular to the center line on the pyramid and put the cereal box flush with the ruler and then focus on the cereal box. Back focus the number behind the center point will be in focus, front focus the opposite, adjust until correct?

Eli S wrote:

But if the point is a little above the center, doesn't that mean that the af point is off and if i adjust it forward it will be right on?

I think you are confusing vertical and horizontal focus mark accuracy with front-and-back focus plane accuracy. Consider a slanted target:

A B C
1[2]3
X Y Z

in this case, what you are thinking, is if the mark is centered on "2" but actually focuses on "B" there is a back focus problem - which is not the case - what you may actually have is a focus mark that is off-center but still accurate in focus. In order to check exclusively for back or front focus, you need the entire target equidistant from the camera.

Which is why having a perpendicular target (Lens Align) is better than a slanted one (Focus Pyramid) because it precludes any focus mark shift and measures just the distance to target, and not any error in the focus sensor marking vs. actual position of the focus sensor.

-- hide signature --

Chris

SteveL54 Senior Member • Posts: 2,696
Looks like mirror slap

Stangs55 wrote:

Robin Casady wrote:

The 70-200 @ f/2.8 looks like it has serious motion blur. Look at "Michael Tapes Design" and notice how the blur is all vertical. Self timer isn't sufficient. You need to use mirror lock up with at least a 3 second delay. Better to do it manually and wait longer after mirror up. You have to use a remote release cord or wireless.

What was the shutter speed for both exposures? If the shutter speed is long enough, mirror flap will be diminished because it is a smaller % of the exposure.

What tripod and head are you using?

2.8 at 1/10 sec
8 at 0.5 sec

If it was motion blur...why is it that when I stop it down to f/8 taking my shutter speed from 1/10 sec to 0.5 sec, does it get very sharp?

I've found 1/10s with the 70-200mm VR is about the worst shutter speed to use on a tripod. When you use longer exposures most of the exposure takes place after the rig's vibration has settled out.

Get a lot more light on your subject and use higher ISO to get the shutter up to 1/100s and use mirror up with remote release or use exposure delay.

Steve

MichaelT Veteran Member • Posts: 3,780
Allow me to add to the conversation

Hi,

I am the developer of LensAlign, and this thread was just brought to my attention. Let me see if I can clarify some of the issues.

Not in any order

  • It is undeniable that AF Adjustment can in most cases improve your sharpness hit rate . I would say that 50% of lenses match up well with a given body (no adjustment needed) and the other half can be improved in terms of getting a higher yield of sharper images. Also some lenses (a small percentage) are found to be out to lunch and require factory calibration.

  • The more resolution the camera can capture, the more you can see focus errors and the more you need to adjust the AFA adjustment.

  • I do not agree with those who say that rather than Adjust a lens to +12 (for example) they would rather get a foc repair . However the lens is not broken, and it is likely that it will come back the same way it was sent in. The object of AF Adjustment is to tweak the A performance on a lens to lens basis to a point that is superior to the out of the box performance. As noted by Canon " For those who need to make fine adjustments to the focus position [AF Microadjustment] is available. " They do not note this as an abnormal circumstance.

  • More from Canon " Because of the increase in resolution of camera sensors, any slight focus mis-alignment is more visible when reviewing images. Although the cameras and lenses are made to extremely high tolerances there is a tolerance range and, in some cases, the camera could be at one end of the range and the lens at the other. In this instance, you would notice the focus point is either in front, or behind where you thought it should be. By using AF Microadjustment you can bring the two together. "

  • Leveling - The LensAlign does not have to be leveled. The patented True Parallel Alignment (centering the rear dot within the front dot), assures that the camera sensor and the LensAlign Focus Target are parallel. The front to back leveling is fully accommodated in the alignment process. The left to right leveling is simply cosmetic in that the picture will not look crooked (need rotation). That has no effect on the parallel requirement.

  • Mr. Sherman's opinion that focus targets do not meet Nikon's guidelines is something that Mr. Sherman has posted 100s of times on the web with no documentation. I have tested his ideal target against LensAlign and find no difference. Mr. Sherman is entitled to his opinion.

  • Axial Chromatic Aberration is common when fast lenses are shot wide open, especially on hi contrast subjects, like a focus target. Our The eye can me fooled by the colors in terms of determining the actual focus plane. Our recommendation is to shoot Large Fine JPEG with the settings of Monochrome, Sharpening Off and contrast high.

  • It is not appropriate for me to comment on other products here, but let me say this. Here are the requirements for proper AF testing, that will ensure the consistent and accurate results:

  • Focus target (subject) must be flat and parallel to the imaging sensor. Slanted targets are not acceptable.

  • DOF display device (ruler) must be perpendicular to the focus target and extend behind and in front of the desired focus acquisition plane (focus target)

  • The AF testing must be done at a minimum distance of 25x the focal length of the lens. testing closer can lead to inaccurate AF Settings. The best test distance is a close to your normal shooting distance as possible. (if you are going to shoot some tight head shots that are closer to the subject than the 25xFL...then test/adjust at that distance for that session. The setting might not be proper for shooting at longer distances.

  • Any procedure that uses a flat target only, require sequences of photographs in order to evaluate focus error and type of error. methods that show a Depth of Field display are all encompassing. In one picture, you can asses the amount and direction of the error. This is especially useful to test a body/lens that has already been calibrated and is just being tests. A flat target system cannot determine anything with one picture. A slanted display system can confirm proper AF performance with one picture.

To summarize.....

  • AF Adjustment will improve yield of sharp pictures in about half the time (could be more or less).

  • If one shoots stopped down to f/8 and at long distances all of the time the advantage might be swamped by the wide DOF. So AF Adjustment is not for all shooters. It is? for all D800/E shooters :> )

  • There are many methods that will allow the test of AF performance (the first part of the Adjustment process). You can use batteries, cereal boxes, VHS tapes, etc. And on any day they might be accurate. What you are paying for with LensALign and other proper devices is the accuracy and repeatability of the test environment which is crucial for proper AF Adjustment.

I have tried not to make this an educational piece and not a marketing piece. If I have not succeeded, I am happy to remove the post. I posted because there were many questions and some misinformation about the AF Adjust process and LensAlign and was asked to respond by a poster. I am happy to respond furtner here if someone makes me aware of new posts. Also we are available for inquiries at our help desk which is updated at least daily.

-- hide signature --

Regards,

Michael Tapes
Owner: RawWorkflow.com
Creator: WhiBal, LensAlign

reginalddwight Senior Member • Posts: 2,003
Thank you for the post

Many thanks for the post. Useful information that may help clarify misunderstanding of LensAlign for some DPR forum members. BTW, I think the dude's name is Shepherd.

I don't own LensAlign, but tested the accuracy of my D800's outer AF points with other subjects, including the D800 camera box and, believe it or not, real people.

MichaelT Veteran Member • Posts: 3,780
Glad it was helpful (nt)
n/t
u007 Senior Member • Posts: 1,681
Re: D800 & LensAlign

The 70-200 shot has definite motion blur.

You can see it very clearly in the writing. It's an up-down motion blur. Were these shot with the mirror up?

Also, as a control you should shoot a perfectly manually focused (with magnified live view) shot of the target. Then you know the sharpness of the lens when you take autofocus out of the equation.

Then adjust the AF until you get a result as close to the manual focused shot as possible.

Either way, you definitely need mirror up for those.

-- hide signature --

My travel photography blog - http://www.frescoglobe.com

lanef Forum Pro • Posts: 10,317
Re: Thank you MichaelT

To adjust a primes lens would be quite easy but what about zoom lenses?
Do you have to adjust both ends and the middle FL or the middle FL only?

Stangs55
OP Stangs55 Regular Member • Posts: 479
Re: Allow me to add to the conversation

MichaelT wrote:

Hi,

I am the developer of LensAlign, and this thread was just brought to my attention. Let me see if I can clarify some of the issues.

Not in any order
.......
--
Regards,

Michael Tapes
Owner: RawWorkflow.com
Creator: WhiBal, LensAlign

Thanks for stopping by Michael.

I actually made a post about this originally in the LensAlign "forums"...but, wow...those may be the only forums that are more archaic than dpreview... I made a post, but then went to reply to my post and it replaced the original post with my reply....I then couldn't reply to that post, make a new post, edit a post, post pictures...I really couldn't do anything after I had made one post...

Either way...

Could you specifically comment on what you see in my second set of images here: http://forums.dpreview.com/forums/readflat.asp?forum=1021&message=41369140

There seems to be the consensus that I have motion blur in my 70/200, but I'm having a hard time believing that. I'm going to retest outside today where I can get faster shutter speeds....but I'm very curious to hear your thoughts.

Thanks.

MichaelT Veteran Member • Posts: 3,780
Prime vs Zoom

Since the cameras only have one slot for the AF adjustment to be stored, that complicates the calibrating of zoom lenses (note exceptions below). The method that we have come to in our lab and with our customer base is to align the long end of the zoom first. Then keeping the adjustment in place, test the lens at the short end, changing the LensAlign to the appropriate distance for the shorter Focal length. If the focus is pretty close, then do this test. Turn off the AF Adjust ( disable for Canon, off for Nikon). Then you can see if you are better off with or without the adjustment.

  • If it is better then you can leave that adjustment in place. If you are obsessive you can also test for the correct setting for the short end of the zoom and then make a judgement as to whether you want to compromise the overall setting, maybe biased toward one end or the other. Generally it is thought that since the short end of the zoom will always have more DOF, it is more forgiving. But also...at the same aperture and relative distance (like 25 x Focal length) the DOF will be identical. So consider your shooting style and use your judgement.

  • If the short end is worse with the setting then you definitely have to test for the correct setting, and then make a decision as to what to do. Since LensAlign always shows you the DOF in all tests shots (not true with any non-sloped display method) you always have the full picture (so to speak), so you can make a decision.

  • In both cases when you are done setting the long and short ends, you should test in the middle.

Exceptions to the zoom conundrum

  • The newest canon DSLRs (1Dx and 5D3) now allow the setting of both the long and short end of zoom lenses. You test and align the long, and then after changing distance you test and align for the short end. Then the camera will use the ends to calculate the proper AF adjustment setting for intermediate zoom positions.

  • These same Canon cameras also allow you to store adjustment settings for more than 1 lens of a specific model (they allow you to store the serial # with each lens).

  • Olympus DSLRs have always allowed the setting of both ends of the zoom. Additionally they allow you to test and adjust for each focus point individually.

-- hide signature --

Regards,

Michael Tapes
Owner: RawWorkflow.com
Creator: WhiBal, LensAlign

MichaelT Veteran Member • Posts: 3,780
Regarding your 2nd set of images
  • Test 1 24-70 @ 2.8 - moderate BF (more than a touch). In addition to checking at the focus plane...also check out the extremes. The top 36 is much sharper than the bottom 36. That is how you can quickly tell what is going on. For subtle adjustments, that is not best. But when a lens has a reasonable amount of AF error, just scan the ends of the ruler for a quick look. Then you can guess at an adjustment and go from there. (BTW you should specify what Fl this is shot at...I assume 70mm.

  • test 2 - Large BF

  • 105mm f/2.8 Nikon - Also BF. Just look at the large 16 top and bottom.

  • 70-200 @ f/2.8 - yes, motion blur. But not enough to obscure the fact that this lens is also back-focusing

  • @f/8 - same BF

In all of the above cases, your pictures will be sharper with an adjustment. Sometimes it is academic (if it is a slow lens)...but in each case here, you can see how the zero line is less sharp then towards the rear.

In this case the find edges or emboss are not needed, since the focus errors are quite visible on the naked display ruler.

Hope this helps and sorry you had trouble at the help desk. Many people submit picture for eval if they are not sure. it is like an optical illusion. Once you see it you will know what you are looking for and it will be quite simple. BTW...your camer alignment is perfect.
--
Regards,

Michael Tapes
Owner: RawWorkflow.com
Creator: WhiBal, LensAlign

lanef Forum Pro • Posts: 10,317
Re: Prime vs Zoom

MichaelT wrote:

Since the cameras only have one slot for the AF adjustment to be stored, that complicates the calibrating of zoom lenses (note exceptions below). The method that we have come to in our lab and with our customer base is to align the long end of the zoom first. Then keeping the adjustment in place, test the lens at the short end, changing the LensAlign to the appropriate distance for the shorter Focal length. If the focus is pretty close, then do this test. Turn off the AF Adjust ( disable for Canon, off for Nikon). Then you can see if you are better off with or without the adjustment.

  • If it is better then you can leave that adjustment in place. If you are obsessive you can also test for the correct setting for the short end of the zoom and then make a judgement as to whether you want to compromise the overall setting, maybe biased toward one end or the other. Generally it is thought that since the short end of the zoom will always have more DOF, it is more forgiving. But also...at the same aperture and relative distance (like 25 x Focal length) the DOF will be identical. So consider your shooting style and use your judgement.

  • If the short end is worse with the setting then you definitely have to test for the correct setting, and then make a decision as to what to do. Since LensAlign always shows you the DOF in all tests shots (not true with any non-sloped display method) you always have the full picture (so to speak), so you can make a decision.

  • In both cases when you are done setting the long and short ends, you should test in the middle.

Exceptions to the zoom conundrum

  • The newest canon DSLRs (1Dx and 5D3) now allow the setting of both the long and short end of zoom lenses. You test and align the long, and then after changing distance you test and align for the short end. Then the camera will use the ends to calculate the proper AF adjustment setting for intermediate zoom positions.

  • These same Canon cameras also allow you to store adjustment settings for more than 1 lens of a specific model (they allow you to store the serial # with each lens).

  • Olympus DSLRs have always allowed the setting of both ends of the zoom. Additionally they allow you to test and adjust for each focus point individually.

Thanks for taking your time to throw some light on that complicated issue for DPR members, cheers.

Stangs55
OP Stangs55 Regular Member • Posts: 479
Re: D800 & LensAlign

Robin Casady wrote:

The 70-200 @ f/2.8 looks like it has serious motion blur. Look at "Michael Tapes Design" and notice how the blur is all vertical. Self timer isn't sufficient. You need to use mirror lock up with at least a 3 second delay. Better to do it manually and wait longer after mirror up. You have to use a remote release cord or wireless.

-- hide signature --

Robin Casady
http://www.robincasady.com/Photo/index.html

SteveL54 wrote:

I've found 1/10s with the 70-200mm VR is about the worst shutter speed to use on a tripod. When you use longer exposures most of the exposure takes place after the rig's vibration has settled out.

Get a lot more light on your subject and use higher ISO to get the shutter up to 1/100s and use mirror up with remote release or use exposure delay.

Steve

Oscarroos wrote:

You used the lens align correctly but if I were in your shoes, I would take the 70-200 outside today nice and sunny now, and take some photos wide open of brick walls, whatever to see how the photos look. Based on your low shutter speed could be motion blur. Testing outside with high shutter speed should answer some of the doubt.

Where in Texas are you?

Oscar DFW area north Texas

So I just retested outside on a hot, sultry Oklahoma day using MichaelT's suggested in-camera settings (including the in-camera monochrome settings) and this is what I'm seeing at 200mm.

Tripod mounted
Mirror lock up
Remote cord

But if DOF typically extends 2/3 behind and 1/3 ahead...this eyeballing method still seems rather difficult to try and determine where my true focal point should be...because I shouldn't just be looking for an equally sharp #4 on the front and back as they should put my DoF 1/2 in front and 1/2 in back...or am I missing something there...

ISO 100
1/2000 at 2.8

ISO 100
1/160 at 2.8

Keyboard shortcuts:
FForum MMy threads