Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Nikon Focus Chart Download
70 views
Skip to first unread message
Remedios Fava
Dec 29, 2023, 11:19:36 PM12/29/23
to
Below is the easiest and quickest way to test if your DSLR has an autofocus issue, along with a recommendation on what to do if there is a problem. This test can be used to detect front or back focus issues with a particular lens or a camera body. I will be using the Nikon D800E as a reference camera for this article, but any modern DSLR with Live View capability can be used for the same test (even entry-level DSLRs such as the Nikon D3500 have a Live View mode). Why would you want to test your camera for autofocus issues? Because if your camera or your lenses are defective or have a calibration problem, then you will not be able to obtain critically sharp images.
When testing the autofocus accuracy of a camera or lens, it is always best to have exposure consistency. Therefore, I recommend switching to Manual mode and keeping the same exposure for each shot. Here is a summary of what I would recommend in terms of camera setup:
nikon focus chart download
Download https://1cuiposvspirshi.blogspot.com/?dz=2wYJUb
Now, take a look at the other 3-4 images you captured after this one. Do they look more or less the same? Or do they all look out of focus? Here is a second shot that I captured, which clearly indicates a focus problem:
As you can see, the image not only looks out of focus, but we are also seeing some aberration problems in the image as well (a pretty normal thing to see when an image is out of focus). All three shots were very similar, which clearly indicates a focus problem either on the lens or the camera body.
As I have already pointed out, it is best to test multiple lenses when assessing the autofocus accuracy of your camera. If you only have one lens to test, then it would be impossible to tell whether the lens or the camera have alignment issues. Sometimes it could be both. If you have 3-4 different lenses, perform the above test on all lenses you have and see what you get. If all lenses show very similar results, then your camera is most likely to blame for AF issues. Still, it will be hard to pinpoint the exact source of the problem, because it all depends on how badly calibrated your lenses were before you started the test. See my article on lens calibration.
Just read the article and it makes things clear to this newbie! I am going to set up a test area and see how I make out. I always seem to get shots that are almost in perfect focus but lack a little bit to make them great. Thanks for all this info!
I have calibrated my lenses. I understand why we should turn off the lens fine tuning when we capture the focus chart in Live View mode. But since my lenses are fine tuned should I turn it back on when I capture the second image without Live View? If I do not turn it on I know the images will not be in focus because they are not calibrated.
The basic premise of a chart like this is that when photographed at anangle, the paper slices through the plane of focus. Depending on thesituation, that plane can be very thick, encompassing the whole chart(leaving the whole chart in focus), or, more usefully, it can be very thin,leaving only the areas of the chart that intersect with that thin region infocus.
The illustration below shows proper autofocus and the result one mightexpect from it, where the region of what is actually in focus is more orless centered on the focus target. It also has views for the problems ofback focus and front focus, where for some reason, the lensactually focuses on a region either behind or in front of the focus point(resulting in the in-focus area on the chart to be above or below the focustarget). Mouseover the buttons under the chart to see those views.
Two popular autofocus test charts I've seen and used (and learned muchfrom), by Tim Jackson (no longer online) and LeonGoodman, don't address either of these concerns as well as they could,which is why I developed the chart presented here.
Below is a small section of the chart at full resolution, from slightlyleft of center, showing part of the black bar that is the focus target(with the red line added here just to indicate the vertical centerline ofthe chart).
This provides a clear target for the autofocus. The low-contrast grayextends quite a bit from side to side, but more importantly, it extends thefull height of the page. This means that even when the chart is viewed at asteep angle like that shown in the photo at the top of this page, a largearea of low-contrast gray still presents itself around the target,buffering it from anything else that the autofocus might lock on. Thus,with the precautions discussed later, you're sure that the autofocus locksonto the target bar if it locks onto anything at all.
I've seen autofocus test charts that use lines of random LoremIpsum text, because, as the author correctly notes, we arehypersensitive to the crispness of text, so it makes for good test fodder. The problem with this approach is that we don't really care aboutabsolute crispness, but want to scan up and down to gaugerelative crispness. For this, random lines of spaced text are not asgood because there's no continuity as you scan vertically.
On the other hand, this chart's vertical lines make for something thatwould be smooth and consistent throughout a vertical scan if focus wereperfect throughout, but since focus is not perfect throughout, the smoothand consistent nature of the lines highlight clearly what is and isn't infocus.
Additionally, when some of the lines are dashed, the individual blocksmaking up the line become convenient visual markers of distance from thefocus target, allowing you to quickly compare a block above thetarget with its counterpart the same distance below the target.
Your situation may be different from mine, so I've built seven differentversions of the test chart, each with differing levels of gray for thelow-contrast areas. The 5% version is the most faint, while the 35% is thedarkest:
If possible, print on high-resolution matte photo paper, such as Canon'sMP-101. Besidesgiving a crisper print, the paper is more sturdy, which makes it resistantto curling and warping (both of which are detrimental to its successful usein testing focus).
After printing, test the appropriateness of the low-contrast gray bylighting the chart well, filling the viewfinder with the gray area, andchecking to see whether the autofocus can lock on to it. If it can, youneed to move to a lower-percent gray.
The version you've printed holds promise if the autofocus system can'tlock onto the gray, but the quick test you've just done is onlypreliminary. Be sure to check again under actual conditions after setting upfor the real shot, as described below.
Deciding how far away you should be from the chart is influenced by whatkind of test you want (some lenses display autofocus accuracy that varieswith subject distance), the focal length of the lens, and the nature ofyour autofocus sensors.
It's possible that the active area of the camera's autofocus sensorexactly matches the indicator you see in the viewfinder, but it's notlikely. More likely is that the active area of detection is a bit larger,perhaps asymmetrically so.
The photo sequences below illustrate the pitfalls of not understandingyour camera's autofocus sensors. The right-hand image is just a closeup ofthe central part of the full-frame left-hand image. Mouseover the fourdescriptions below the pictures to see the story unfold...
The final result above appears as if the camera/lens has back focus, but inreality, it's just picking an unexpected spot to focus on. It would be niceif the indicator in the viewfinder exactly matched the active area, butsince that's not the case, it behooves you to understand your sensors.
I'll leave the detailed description of mapping your autofocus sensors tothe link in the previous paragraph, but in short, one way to measure theactive area is to get close enough to the chart so that the low-contrastgray area fills the width of the viewfinder, and place the sensor indicatorright in the middle of it. Having nothing high-contrast anywhere nearby, itshouldn't be able to lock onto a focus, but instead, should hunt arounduntil it gives up. (If it can lock onto the low-contrast gray areas, you'llneed to use one of the lower-percent gray versions offered in the How to Print It section.)
If you have multiple sensors, they may well each have their owncharacteristics, so you may wish to map them all, but for the purposes oftesting your autofocus system, it's sufficient to use only the middlesensor.
Of course, the more you can fill the viewfinder with the chart, the better.Even if the scenario above results in a valid test, chart is too small toreally make out much detail from it, even with the resolution my10-megapixel SLR affords.
After taking into account everything in the previous section, it's a simplematter to take some shots. Take multiple shots at different angles ofattack. A lower angle shows the depth of the in-focus field the mostclearly (and with it, the accuracy of the autofocus system), but requiresthe most care to ensure that the autofocus sensor does not see the top ofthe chart instead of the intended target.
Be very careful not to allow yourself to move the camera between achieving focus lock and actually taking the picture. Some people, for example, have an unconscious habit to move forward a half an inch in the process of taking the shot, and such movement would absolutely destroy any meaning to this test.
Pause occasionally to double-check that the autofocus can't lock on the low-contrast gray (by pointing at the wide expanse of it in the upper half of the chart and confirming that focus can't be found.)
Once I've set up for a particular shot, I take it, then point the cameraat something far and autofocus there, then return to autofocus on thechart, taking a second shot. I then do the same with something near (if I'mnot already near the minimum-focus distance of the lens) and return for athird shot. This way, I feel sure that the autofocus is starting fromscratch each time.
Actually, in this case, there might be the slightest bit more sharpnessto the numbers above the midline, but it's so slight that it could wellbe because the autofocus picked the top edge of the target stripe ratherthan the bottom edge. That's how good the autofocus was with this shot.
35fe9a5643
When testing the autofocus accuracy of a camera or lens, it is always best to have exposure consistency. Therefore, I recommend switching to Manual mode and keeping the same exposure for each shot. Here is a summary of what I would recommend in terms of camera setup:
nikon focus chart download
Download https://1cuiposvspirshi.blogspot.com/?dz=2wYJUb
Now, take a look at the other 3-4 images you captured after this one. Do they look more or less the same? Or do they all look out of focus? Here is a second shot that I captured, which clearly indicates a focus problem:
As you can see, the image not only looks out of focus, but we are also seeing some aberration problems in the image as well (a pretty normal thing to see when an image is out of focus). All three shots were very similar, which clearly indicates a focus problem either on the lens or the camera body.
As I have already pointed out, it is best to test multiple lenses when assessing the autofocus accuracy of your camera. If you only have one lens to test, then it would be impossible to tell whether the lens or the camera have alignment issues. Sometimes it could be both. If you have 3-4 different lenses, perform the above test on all lenses you have and see what you get. If all lenses show very similar results, then your camera is most likely to blame for AF issues. Still, it will be hard to pinpoint the exact source of the problem, because it all depends on how badly calibrated your lenses were before you started the test. See my article on lens calibration.
Just read the article and it makes things clear to this newbie! I am going to set up a test area and see how I make out. I always seem to get shots that are almost in perfect focus but lack a little bit to make them great. Thanks for all this info!
I have calibrated my lenses. I understand why we should turn off the lens fine tuning when we capture the focus chart in Live View mode. But since my lenses are fine tuned should I turn it back on when I capture the second image without Live View? If I do not turn it on I know the images will not be in focus because they are not calibrated.
The basic premise of a chart like this is that when photographed at anangle, the paper slices through the plane of focus. Depending on thesituation, that plane can be very thick, encompassing the whole chart(leaving the whole chart in focus), or, more usefully, it can be very thin,leaving only the areas of the chart that intersect with that thin region infocus.
The illustration below shows proper autofocus and the result one mightexpect from it, where the region of what is actually in focus is more orless centered on the focus target. It also has views for the problems ofback focus and front focus, where for some reason, the lensactually focuses on a region either behind or in front of the focus point(resulting in the in-focus area on the chart to be above or below the focustarget). Mouseover the buttons under the chart to see those views.
Two popular autofocus test charts I've seen and used (and learned muchfrom), by Tim Jackson (no longer online) and LeonGoodman, don't address either of these concerns as well as they could,which is why I developed the chart presented here.
Below is a small section of the chart at full resolution, from slightlyleft of center, showing part of the black bar that is the focus target(with the red line added here just to indicate the vertical centerline ofthe chart).
This provides a clear target for the autofocus. The low-contrast grayextends quite a bit from side to side, but more importantly, it extends thefull height of the page. This means that even when the chart is viewed at asteep angle like that shown in the photo at the top of this page, a largearea of low-contrast gray still presents itself around the target,buffering it from anything else that the autofocus might lock on. Thus,with the precautions discussed later, you're sure that the autofocus locksonto the target bar if it locks onto anything at all.
I've seen autofocus test charts that use lines of random LoremIpsum text, because, as the author correctly notes, we arehypersensitive to the crispness of text, so it makes for good test fodder. The problem with this approach is that we don't really care aboutabsolute crispness, but want to scan up and down to gaugerelative crispness. For this, random lines of spaced text are not asgood because there's no continuity as you scan vertically.
On the other hand, this chart's vertical lines make for something thatwould be smooth and consistent throughout a vertical scan if focus wereperfect throughout, but since focus is not perfect throughout, the smoothand consistent nature of the lines highlight clearly what is and isn't infocus.
Additionally, when some of the lines are dashed, the individual blocksmaking up the line become convenient visual markers of distance from thefocus target, allowing you to quickly compare a block above thetarget with its counterpart the same distance below the target.
Your situation may be different from mine, so I've built seven differentversions of the test chart, each with differing levels of gray for thelow-contrast areas. The 5% version is the most faint, while the 35% is thedarkest:
If possible, print on high-resolution matte photo paper, such as Canon'sMP-101. Besidesgiving a crisper print, the paper is more sturdy, which makes it resistantto curling and warping (both of which are detrimental to its successful usein testing focus).
After printing, test the appropriateness of the low-contrast gray bylighting the chart well, filling the viewfinder with the gray area, andchecking to see whether the autofocus can lock on to it. If it can, youneed to move to a lower-percent gray.
The version you've printed holds promise if the autofocus system can'tlock onto the gray, but the quick test you've just done is onlypreliminary. Be sure to check again under actual conditions after setting upfor the real shot, as described below.
Deciding how far away you should be from the chart is influenced by whatkind of test you want (some lenses display autofocus accuracy that varieswith subject distance), the focal length of the lens, and the nature ofyour autofocus sensors.
It's possible that the active area of the camera's autofocus sensorexactly matches the indicator you see in the viewfinder, but it's notlikely. More likely is that the active area of detection is a bit larger,perhaps asymmetrically so.
The photo sequences below illustrate the pitfalls of not understandingyour camera's autofocus sensors. The right-hand image is just a closeup ofthe central part of the full-frame left-hand image. Mouseover the fourdescriptions below the pictures to see the story unfold...
The final result above appears as if the camera/lens has back focus, but inreality, it's just picking an unexpected spot to focus on. It would be niceif the indicator in the viewfinder exactly matched the active area, butsince that's not the case, it behooves you to understand your sensors.
I'll leave the detailed description of mapping your autofocus sensors tothe link in the previous paragraph, but in short, one way to measure theactive area is to get close enough to the chart so that the low-contrastgray area fills the width of the viewfinder, and place the sensor indicatorright in the middle of it. Having nothing high-contrast anywhere nearby, itshouldn't be able to lock onto a focus, but instead, should hunt arounduntil it gives up. (If it can lock onto the low-contrast gray areas, you'llneed to use one of the lower-percent gray versions offered in the How to Print It section.)
If you have multiple sensors, they may well each have their owncharacteristics, so you may wish to map them all, but for the purposes oftesting your autofocus system, it's sufficient to use only the middlesensor.
Of course, the more you can fill the viewfinder with the chart, the better.Even if the scenario above results in a valid test, chart is too small toreally make out much detail from it, even with the resolution my10-megapixel SLR affords.
After taking into account everything in the previous section, it's a simplematter to take some shots. Take multiple shots at different angles ofattack. A lower angle shows the depth of the in-focus field the mostclearly (and with it, the accuracy of the autofocus system), but requiresthe most care to ensure that the autofocus sensor does not see the top ofthe chart instead of the intended target.
Be very careful not to allow yourself to move the camera between achieving focus lock and actually taking the picture. Some people, for example, have an unconscious habit to move forward a half an inch in the process of taking the shot, and such movement would absolutely destroy any meaning to this test.
Pause occasionally to double-check that the autofocus can't lock on the low-contrast gray (by pointing at the wide expanse of it in the upper half of the chart and confirming that focus can't be found.)
Once I've set up for a particular shot, I take it, then point the cameraat something far and autofocus there, then return to autofocus on thechart, taking a second shot. I then do the same with something near (if I'mnot already near the minimum-focus distance of the lens) and return for athird shot. This way, I feel sure that the autofocus is starting fromscratch each time.
Actually, in this case, there might be the slightest bit more sharpnessto the numbers above the midline, but it's so slight that it could wellbe because the autofocus picked the top edge of the target stripe ratherthan the bottom edge. That's how good the autofocus was with this shot.
35fe9a5643
0 new messages