Help with focus.

[sidigirl]

I seem to be having issues with my focus.

This was taken with manual focus


IMG_0003 by sidigirl, on Flickr

And this with auto focus.


IMG_0002 by sidigirl, on Flickr

Which is clearer and is it my eye or does the problem appear to be something else.

Thanks.

 

[Rick58]

I looked at your flickr shots and frame 0001 definitely seems to be a DOF issue. If you look at the bottom left corner of that frame, you'll see the wording begins to come into focus. If you're shooting wide open, try stopping down your lens a couple stops and see if that helps.

 

[Light Guru]

What is the minimum focus distance for the lens you are using and how far away is was that book from your lens?

I have a feeling that you were shooting the book with it being right at the minimum focus distance and thus making it hard for the auto focus to get it crisply in focus.

 

[The_Traveler]

If you look at DOFMaster and calculate the DOF for the approx distance given in the exif, the DOF in front of the focus point is .03 meters or 1.18 inches.
If that book wasn't perfectly perpendicular to the place of focus, it might be in focus at the point you focused and the bottom edges be moving forward out of the field.

To test focus, pin something on a wall, put the camera on a tripod , use time delay and try it that way.

 

[gsgary]

Shutter speed was also slow if you were hand holding

 

[pgriz]

Sidigirl, you’ll need to work with us to help us help you. I appreciate that you took these images after I gave you some suggestions in another thread, but there is a lack of information about exactly what you did that makes it hard to diagnose the issues.

1) Did you shoot the images on a tripod? This is very important, as doing this (along with other vibration-reducing steps such as locking up the mirror, using a remote trigger OR setting the camera on delayed release) will eliminate camera shake as one of the possible causes of blur. It also allows the same image (under different settings) to be compared without distance, movement or other effect making the comparision difficult. As well, shooting on the tripod means that you need to turn the len’s IS feature OFF, as the camera can try to “improve” the focus and move the lens around which will instead contribute to the blur. Some cameras have smart-IS programming, and they sense when the camera is on a triod, but I don’t know if yours handles the IS that way. So turn it OFF when on a tripod.

2) You mentioned you used “manual focus”. BUT, we don’t know the circumstances of HOW you got the manual focus. For instance, did you manually focus through the eyepiece? Did you do the manual focus using live-view with the 10x magnify assist? Did you have IS switched on or off? How did you release the shutter (related to #1 above)? To reiterate what I told you in the other thread – the best and most reliable way to get the focus is to use live-view at 10x magnification because that image comes directly off the sensor, and the 10x magnify give you a much clearer view than does a focus attempt through the viewfinder.

3) Your image was also taken with the on-camera flash. Since the camera’s computer CAN change what your camera does when flash is used, it is better to use constant, steady and good light (Say, a north-facing window on a bright day), to give you a good amount of light to see your target with.

4) For testing purposes, using a wide-open aperture will show issues with focusing the best. Yes, for actual shooting, you’d want to shut the aperture down to a higher f/number to get a greater depth-of-field and to use the lens closer to its sweet spot, but for TESTING purposes, it is best to shoot wide open. Your test shot was done under dim conditions and at f/4.5. Dim light makes focusing difficult to acquire both for visual focusing and for the camera’s focusing efforts.

5) According to the EXIF data on your “Manual” shot, your camera iso was set to Auto, your focus mode was set to AI Focus AF, Your shooting mode was “Single”, you have multi-AF points enabled. These settings basically invalidate your attempt to set the focus “manually” as the camera will refocus as soon as you half-press the shutter button. To truly get a manual shot, you need to turn AF on the lens OFF. As it was, the camera was allowed to pick any AF point it wanted, and it basically ignored your attempt at manual focus setting and did it by itself. To make the test meaningful, you need to enable ONLY the center point (which you will point at exactly the same point you focused on manually), you need to change the AF mode to Single-Shot (which will freeze the focus on the selected point once AF is acquired). Your EXIf data also shows that you did NOT use the Live-view feature while focusing, so you do NOT have the most accurate focusing method used for your manual focus.

To repeat my earlier message to you (and to add some additional information):
1) Set your camera up on a triod. If you don’t, you will have a hard time comparing images otherwise.
2) Set your target (the magasine you used was fine) in a bright area with constant, even lighting at least 6 ft away from the lens. This will make it easier to see the focus, and this will put your target far enough away that the minimum focusing distance should not come into play.
3) Turn off the IS and AF on your lens. Your information screen should show MF (manual focus).
4) Change your AF mode to “single-shot”.
5) Change your AF selected to center ONLY.
6) Use ISO 100 as this gives the “cleanest” image with the lowest noise – it makes it easier to interpret the image detail when it is blown up a lot.
7) Set your Shooting Mode to Av (Aperture priority) and set your aperture to f/3.5 (which is wide open on your lens). Set your focal lenth ring to 18mm (widest aperture). Later in the test you can change those settings, but it gives us a clear baseline to use for comparisions.
8) Set your “drive mode” to self-timer, 2 seconds. This is to allow the camera to avoid shutter-click vibrations.
9) You should se t your recording method to RAW. This will store the image in RAW format and will not apply any of the in-camera sharpening, adjustments, etc. Again this is to enable us to see what is actually going on, without trying to figure out what the camera did while recording the JPG file.
10) Your target point should be something clear and high-contrast (such as the word “used” in the second to last paragraph in the text you shot. This should be exactly at the center of the image, because that is where the AF center point is located, and you will need to be sure that the AF and manual focus aim at the exact same point.
11) For the manual shot, you need to enable “liveview”, move the focusing rectangle to the center where your target text is located, then select the (+) magnify button to invoke the magnify function, press it again to move it to the 10x magnify mode, and then adjust the focus ring on the lens manually until the word “used” is in crisp focus. NOW gently depress the shutter button, take your hands off the camera, and let the camera count down the 2 seconds and release the shutter. This is your “Manual” exposure.
12) for the AF exposure, turn the AF lever on the lens to ON, half-press the shutter – that should signal to the camera to focus on the center AF point, and once it acquires the focus (red dot lights up), you can complete the full shutter button press. Hands off, let the camera count down, and take the image. This is the “AF” shot.
13) Upload both images to your comuter, use DPP (the Canon software that came with your camera) to open up the raw image, and go to 200% image view to see how the word you’ve focused on came out in both images. To allow share this with us, use the trimming tool to create a rectangle about 600x400 centered over the target, and without doing any additional processing, convert that rectangle to a JPG image, which you’ll label “manual”. DO NOT resize the image. Do the same with the AF image, and label it “AF”. Now we have something we can compare for AF focusing.

What you see from the long narrative above, is that it takes quite a bit of care to ensure that you are testing the single variable you need to know about, and everything else is kept constant. Otherwise, we just don’t know what it is that is going on during the “test”.

Edit: according to your EXIF, #0002 was manual focus and #0003 was autofocus. Focus on both appears sharpest at the upper left corner, so I'm guessing that the AF was picking the closest spot. However, that doesn't explain your why your #0002 shot was not sharpest at the center.

 

[The_Traveler]

What a great answer by pgriz.

 

[sidigirl]

Thanks, I didn't get a chance to get the shots done today, but I will tomorrow.

 

[sidigirl]

Photos 1 and 2 were shot with manual focus through the view finder after setting the camera as described by pgriz. I have a wireless remote and had the camera set up on the tripod.


IMG_0001 by sidigirl, on Flickr


IMG_0002 by sidigirl, on Flickr


IMG_0003 by sidigirl, on Flickr


IMG_0003 by sidigirl, on Flickr

I also went back over one of my books and reset the diopter to the focus points as Canon suggests. I'm hoping I have it set right now.


Thanks so much for all the help.

 

[pgriz]

I went through the images and at the maximum size, there's little or no difference that I could see between them. If the second set (#0003 and #0004) were set by the AF mechanism, then it appears to be pretty much dead-on. Looking back at your earlier attempts, it may have been that you had multi-AF points enabled that threw the exercise off. The multi-AF points are programmed to pick the point that is closes to the camera to be the one with the best focus. By using only the central focus point, you know exactly which point the camera is focusing on. Anyway, I am hoping you are getting a better feel for what the camera is doing (and maybe even why).

 

[Light Guru]

Cindy I don't think that the instructions given by pgriz were the best in addition the execution of them halls has some issues.

His instructions don't take into account that your lighting conditions will not match his and thus your images are badly under exposed making it hard to tell.

During your execution you are focusing on a image that is behind glass, glass with a lot of glare.

If you are really worried about focus issues you should be using something like this