-= 90D owners unite! Discuss and Post Photos

One of my new favorite macro pics. This jumping spider was walking towards me down the center of a lily leaf. The leaf added a really interesting mix of colors and texture, and the spider is in movement. Usually all my jumping spider pictures are of fairly stationary spiders.

Considering we're all on lock-down here in AZ, I might have the time to try and get a better one!

Archibald wrote in post #19039198
I have no doubt you are having AF problems, and that you know your gear, but I don't understand the problem you are having exactly.

From Dofmaster, the depth of field with the bare 600mm lens at f/11 and at a distance of 8 meters is 7 cm (3"). That is not a lot but should be OK for a typical passerine.

With the 2X on the 600, you would back up to 16 meters for the same framing. The new DOF at f/11 is also 7 cm.

So the DOF seems to be adequate. I ran a calculation to see if maybe there is an issue with the depth of focus (at the sensor plane), but that doesn't seem to be critical. Maybe the 90D firmware is not up to calculating the placement of the focus elements at these focal lengths.

Your calculator is probably based on the "circle of confusion" and assumed viewing conditions. With a very sharp lens, there is no thin band of depths that are pretty much the same sharpness, in the analog projection of the lens, liberally magnified. The center focal plane *is* much sharper than the edges of the "DOF" as defined by such a calculator. I have mounted a Pentax Q (small pixels, like a ~160MP APS-C) on my 400/4DO II, and at f/4 there is a very distinct difference in sharpness over a range of distances that would all fall into the "DOF" by your chart, or on a big-pixel sensor. There is no plateau in the analog optics, other than any compression of sharpness caused by poor lens design or aberrations, which is less of a factor with "very sharp" lenses.

John Sheehy wrote in post #19039406
Your calculator is probably based on the "circle of confusion" and assumed viewing conditions. With a very sharp lens, there is no thin band of depths that are pretty much the same sharpness, in the analog projection of the lens, liberally magnified. The center focal plane *is* much sharper than the edges of the "DOF" as defined by such a calculator. I have mounted a Pentax Q (small pixels, like a ~160MP APS-C) on my 400/4DO II, and at f/4 there is a very distinct difference in sharpness over a range of distances that would all fall into the "DOF" by your chart, or on a big-pixel sensor. There is no plateau in the analog optics, other than any compression of sharpness caused by poor lens design or aberrations, which is less of a factor with "very sharp" lenses.

I agree about the details of DOF and so on. But

1) When doing DOF calculations, one must use accepted methods or you will attract arguments, and

2) More than one experienced user has reported AF or IQ problems with big whites that don't seem to be found with other lenses (including the 100-400mm II).

What is the explanation?

CUBAN BROWN ANOLE

The other reponse has been on target, but I wanted to emphasize that examining high resolution images, like those from the 90d or a comparable 5ds, with the sharpest lenses, at 100%, on a moderately large monitor, makes the image circle calculations look ridiculous. That's not to say that slightly soft images can't make good prints or be suitable for other uses, but when you're pushing things to the limits, like with heavy cropping or large prints, you get to find that *almost* critically sharp either doesn't work or requires a lot more post processing to recover the lost detail and really isn't as good as a critically sharp image would have been.

I happen to need long lenses, thus the 600/1.4 or2x lens, on high res bodies, and I still find that I have to crop quite a bit, quite often, so what's normal for me is really out at the extreme end of what conventional camera gear can do. Macro work is similar, both involve high magnification in the final image and magnification magnifies everything - especially faults, it seems.

Hi Frank.. I have struggled with AF on the 90D as well. These are the steps I have taken. I shoot with the 100-400 ii and the 1.4 ex III always attached.

I shoot birds only.

1) Do you shake? I shake. I now use a monopod and also try to brace myself on a tree or post. I use the OVF
2) Do you use single point or 9 point? I use both. 9 point for BIF
3) Do you use BBF? Makes all the difference in the world to me because the extra "gripping" helps to stabilize the camera
4) Have you turned off iTR AF Tracking? That is: turn off face tracking and tracking sensitivity. Are your shutters set to focus priority?
5) I use Continuous focus All the time and 10 fps. Remember if you use BBF and you take your thumb off the BB you are now in One-Shot and also can use the Manual Override Focus on the lens if needed.
6) Because I have really poor health, I always shoot in lousy afternoon light!!!! haahahah.. I am trying to learn to make allowances for glare and more glare. Lousy glary light will ruin a shot quicker than anything.
7) I am set at f9 always (unless there is a lot of light and I have plenty of time to fiddle)
8) I shoot manual on Auto ISO. If the light is good I'm shooting as fast as I can for BIF
9) My focus seems to be getting better.. but I'm thinking a lot of it has to do with light and stability
10) Oh I set my acceleration to + 1 and it seems to help

Good luck

I'll respond to your questions by the numbers below:

1) I always use some kind of camera support, mostly a home made car window mount. Shake is not the problem, besides, I can tell shake from other problems that effect sharpness.

2) Single point, except for a rare flight shot.

3) I don't use back button focus, it's too painful with severe arthritis in my thumbs.

4) iTR AF is of no use for birds and it's OFF.

5) OneShot Mode seems to give better AF performance than AIservo, but I'm still going to have to change the way I work to use LiveView AF as much as possible and switch to a 5dsR for times when I have to follow a difficult subject with the optical viewfinder.

6.) I work in all kinds of weather, especially in Oregon's gray, wet winter. I have not found glare to be a problem in full sun.

7) Wide open with the 600/2x is f/8, mostly I try to use f/11 and 1/500th second. My mantra is "one stop down and adjust as needed".

8) Likewise. BIFs are rare for me.

9) Sure, good light and high shutter speeds do wonders for image sharpness.

10) Don't know, off hand, how mine is set.

a few in flgihts from last week with the og 500f4IS...some are probably a bit off, i need to stop shooting wide open i think and give myself a bit of room for error

IMAGE LINK: https://flic.kr/p/2iLo​Q3U

IMAGE LINK: https://flic.kr/p/2iKv​CCu

IMAGE LINK: https://flic.kr/p/2iKs​Svc

FrankKolwicz wrote in post #19038359
I've had a 90d since December and have had trouble with the autofocus, even after sending it and the 600/4/1.4x tele to Canon for evaluation. Their recommendation was to use OneShot mode and only use AIServo for birds moving toward or away from me. However, most birds are moving in some way, even if slightly, and getting critical head/eye focus with shallow DOF in OneShot mode is a random procedure, not really reliable, but AIServo is a lot worse, often missing critical focus for long series of images of the same bird, in the same spot, with normal head and body movements at shutter speeds that yeilded sharp images when I switched to LiveView focus.

The AF problem is that it does not put critical focus where I've aimed it reliably, most often backfocussing, no matter how many times I do a micro-focus adjustment. Those adjustment settings, too, have varied at times, not related to operating or ambient temperature.

All my images (RAWs) are examined at 100% in LightRoom or DPP4 or other viewing/editing software and I can often see that best focus is either in front of or behind the subject when it isn't right on. In fact, I can often see that the subject is toward the front of the DOF zone that would have yeilded a critically sharp image, if focus was not too far back.

My experience with Canon products goes back about 30 years and I've been doing nothing but bird photography for the last 7 years, almost daily. Before the 90d, I've been using 5ds and 5dsR bodies with the 600 and 300/2.8 lenses for a couple of years and have not had this problem with them. The 5ds bodies have been easy to do an MFA and reliable in use compared to LiveView images on the same subjects.

If Canon thinks this level of AF performance is acceptable, they are wrong, but the 90d has several features that fix problems for me using the 5ds and I'll still try to use it as much as possible with LiveView and switch to the 5dsR for times when I have to use an AF/optical viewfinder.

I had the same behavior with the 90d when using the 600 III, where out of a certain number of burst shots, the AF would suddenly miss 3-5 frames in a row. Not by a small amount either. Because of this, I sold the camera within 2 weeks. Really, really wish it would have worked out, because I loved everything else about it.

MatthewK wrote in post #19040609
I had the same behavior with the 90d when using the 600 III, where out of a certain number of burst shots, the AF would suddenly miss 3-5 frames in a row. Not by a small amount either. Because of this, I sold the camera within 2 weeks. Really, really wish it would have worked out, because I loved everything else about it.

So it is inconsistent focus with that lens (and 90D). And that probably has something to do with the placing of the focus elements and the algorithms in the firmware that look after that.

What I've seen, and querried Canon about, is that the focus tends to revert to backfocus after being MFA'd repeatedly. It's not far off, just a bit to the front of the sharpest plane visible, but it is consistently so, 90% or more of the time.

I'd go out and MFA every lens/extender/camera body combination I use (600/1.4 or 2x/90d or 5dsR) and then, 10 minutes later, find most of the 90d images are back focussed, as if I had set them that way. BUT, the 5DsR is right on.

It's as if the 90d just doesn't hold MFA and reverts to the factory setting for some other lens.

Happy with my 90D so far. A recent shot from Brands Hatch using a 100-400iiL.

MatthewK wrote in post #19040609
I had the same behavior with the 90d when using the 600 III, where out of a certain number of burst shots, the AF would suddenly miss 3-5 frames in a row. Not by a small amount either. Because of this, I sold the camera within 2 weeks. Really, really wish it would have worked out, because I loved everything else about it.

There seems to be a number of lens-specific anomalies with the 90D. I wonder if they will be fixed eventually with firmware upgrades. My 400/4DO II used to AF fine (if slow) on my 7D2 in live view with the Canon 2x and the Kenko Pro 300 DG 1.4, and although the 1.4 is a Kenko, that model is a straight-forward reporting TC, but the focus just goes beserk on a camera which should, by all accounts, have much better live view AF (it is excellent, in general).

I don't need 2.8x so much with the 90D's pixel density (even the red and blue channels are pretty much oversampled like that, as TC aberrations become significant), but it still bugs me that this fails at f/11, while my v1 100-400 focuses great in live view with a 2x at f/11, and even f/16 works well (if slow) with some combos!

I have no proof, but I have a gut feeling that the 90D is not providing the correct amount of power with some lens configurations, because I have received "check the contacts" messages when the contacts are clean.

It is a shame that Canon's best sensor for small and/or distant subjects doesn't have a DSLR body with complete/pro AF.

FrankKolwicz wrote in post #19040645
What I've seen, and querried Canon about, is that the focus tends to revert to backfocus after being MFA'd repeatedly. It's not far off, just a bit to the front of the sharpest plane visible, but it is consistently so, 90% or more of the time.

Is this only when you have a big, distant background , or also when the field is relatively shallow?

If the former, I wonder if the AF algorithm isn't weighting focus in some way, which doesn't make sense for many types of photography, where you want quantum snapping of focus to the main subject.

The lens stops reporting at some critical moment? Just a guess in the dark.

John Sheehy wrote in post #19040786
There seems to be a number of lens-specific anomalies with the 90D. I wonder if they will be fixed eventually with firmware upgrades. My 400/4DO II used to AF fine (if slow) on my 7D2 in live view with the Canon 2x and the Kenko Pro 300 DG 1.4, and although the 1.4 is a Kenko, that model is a straight-forward reporting TC, but the focus just goes beserk on a camera which should, by all accounts, have much better live view AF (it is excellent, in general).

I don't need 2.8x so much with the 90D's pixel density (even the red and blue channels are pretty much oversampled like that, as TC aberrations become significant), but it still bugs me that this fails at f/11, while my v1 100-400 focuses great in live view with a 2x at f/11, and even f/16 works well (if slow) with some combos!

I have no proof, but I have a gut feeling that the 90D is not providing the correct amount of power with some lens configurations, because I have received "check the contacts" messages when the contacts are clean.

It is a shame that Canon's best sensor for small and/or distant subjects doesn't have a DSLR body with complete/pro AF.

In that same vein, I also didn't have any luck with the 600III (with or without the 1.4TC) on the EOS R, whereas everyone else reported being more than happy with the camera + other lenses. I'm starting to possibly suspect that maybe the 600III is the culprit here, and that it might require more power to drive properly. Because really, the only camera I had good usage with that lens is the 1DX2, which we all know has the more powerful battery.

John Sheehy wrote in post #19040788
Is this only when you have a big, distant background , or also when the field is relatively shallow?

If the former, I wonder if the AF algorithm isn't weighting focus in some way, which doesn't make sense for many types of photography, where you want quantum snapping of focus to the main subject.

The lens stops reporting at some critical moment? Just a guess in the dark.

John:

I seems like it happens most often at longer distances in wide open spaces. I've been doing mostly waterfowl this winter because of conditions at my usual sites and these tend to be at longer distances with reflections of sky on water. As of the last few days, warblers have started showing up in my backyard and the few images I did today at close range have been fine. I expect to do a lot more in the coming couple of months, so I should be able to test that idea thoroughly.

I don't have answers to the other questions and don't know how to check them or even what they mean, exactly.