-= 90D owners unite! Discuss and Post Photos

Very, very nice. How many active focus points? Any other settings tips? MY BIF success rate is low with my 90d, great resolution for macro stills and such. Thanks for any suggestions.

I used 9... and 2 got close!!

Hi Mike.... I don't know how you are set up.. but here's how I am set up

1. I'm using a 1.4 III ex with my 100-400 mk2
2. I'm shooting f9
3. I'm shooting 1/2000 min for swallows
4. I'm shooting Auto ISO (worked out to 800 for this one)
5. For groups of flying birds (there were many out there), I use 9 point
6. I have set my acceleration to +1 (this is with swallows in mind)
7. I shoot BBF (Always on AF-C)

Archibald wrote in post #19059934
Frank, please clarify - what do you mean by "will not take MFA"? I have MFA'd my 100-400 II on the 90D, apparently successfully. It is set at +4 for the bare lens and +13 for the lens with the 1.4X III. These values are similar to what I found with my 7DII. The body seems to AF fine with the lens at these settings.

The greater DOF at shorter focal lengths can easily cover this effect, even at wide open aperture. There is less of a problem for me when using the bare lens alone and with the 1.4x. The slight shift I see wide open with the 2x means the difference between sharp details of a bird's eye and slightly soft and that reflects how much headroom I have when cropping, as was noted in another poster's message in this thread.

When doing the MFA, the first thing I noticed was that the 90d had a lot of trouble focussing on my normal target (a twist and loop of thick fence wire) which no other camera body ever had a problem with. I had to switch to the rusty white top part of a steel post that's about 1" wide on the face I used. The problem with the post is that I can't see the horizontal wire immediately behind it and that makes adjustment in the right direction a little bit harder.

Starting with the factory set zero position, I focus optically on my test subject and then switch to LiveView and look at it at maximum magnification, if it looks off, I put in a guess correction and repeat that until it looks good. I would then proceed immediately to photograph a bird and make a bunch of images, in one case 186 frames of a couple of Dowitchers in a small pond as they interacted - every one of those frames was backfocussed. Lighting was dim overcast and the birds were about twice as far away as when I did the MFA, a common situation that was handled fine by previous cameras. By "fine" I mean not perfect, but I'd get some useable images, at least 10% would be keepers and a lot would be in good focus.

Repeat the MFA, get a slightly different adjustment, and try again - the same. Many more repeats. . .

When I did a bunch of 10 frame sequences of images and adjusted the MFA between them, I could see changes due to big adjustments (5 or more), but still the focal plane shifted a bit in each set on the same number, more or less centered on the new adjustment point. This is not the case with the 5dsR, when it's on target, it stays on target at least 80% of the time.

Well Frank.. Just one question. How much Jim Beam are you drinking between MFA adjustments?

John Sheehy wrote in post #19060082
Perhaps he was talking about the fact that even after getting MFA right on average, he still gets a little variation in controlled, repeatable focus situations.

That could be some play in the combined 90D+lens AF system, but it could also be that the play is there with other combos, too, but not noticeable when the pixel density is lower. He compares to the 5Dsr, which has 60% the pixel density of the 90D.

The higher the pixel density, the more repeatably accurate the auto-focus system needs to be to get a certain high percentage of keepers judged by the sharpest possible 100% pixel views that you have seen from that sensor. The 90D would need much more accurate AF than the 1Dx3 would need to get 100% of images in maximum appreciable focus, judged at 100% pixel magnification. What would be a little off on the 90D might look well-focused on the 1Dx3, or any 20MP FF or 8MP APS-C like the 20D and 30D.

Thanks for the analysis of the AF versus the 1dx3, etc. I was under the impression that Canon's flagship camera got the best of the hardware and other models less so.

I'm not only talking about repeatability in controlled conditions (the MFA setting), but also the unreliability of the setting in the field, as explained in another message earlier. When I do an MFA, then take it to the field and, in common dim lighting conditions (Oregon, you know), I get total AF failure. Mostly with the 2x, much less so with the 1.4x and bare 600 and less so under full sun, full frontal lighting. Unfortunately, I have to use the 2x a lot!

I think the pixel pitch of of the 5dsR is 4.14μm versus 3.23μm for the 90D. That makes 3.23/4.14 = 78%. That is, the 90d has the smaller pixel pitch at 78% of the 5dsR's: to my mind that's fairly comparable, not a gross exaggeration of image appearance at "actual pixels".

Sorry, this is Oregon and I'm an IPA and Pinot Noir kinda guy. Eat and drink local!

FrankKolwicz wrote in post #19060502
The greater DOF at shorter focal lengths can easily cover this effect, even at wide open aperture. There is less of a problem for me when using the bare lens alone and with the 1.4x. The slight shift I see wide open with the 2x means the difference between sharp details of a bird's eye and slightly soft and that reflects how much headroom I have when cropping, as was noted in another poster's message in this thread.

When doing the MFA, the first thing I noticed was that the 90d had a lot of trouble focussing on my normal target (a twist and loop of thick fence wire) which no other camera body ever had a problem with. I had to switch to the rusty white top part of a steel post that's about 1" wide on the face I used. The problem with the post is that I can't see the horizontal wire immediately behind it and that makes adjustment in the right direction a little bit harder.

Starting with the factory set zero position, I focus optically on my test subject and then switch to LiveView and look at it at maximum magnification, if it looks off, I put in a guess correction and repeat that until it looks good. I would then proceed immediately to photograph a bird and make a bunch of images, in one case 186 frames of a couple of Dowitchers in a small pond as they interacted - every one of those frames was backfocussed. Lighting was dim overcast and the birds were about twice as far away as when I did the MFA, a common situation that was handled fine by previous cameras. By "fine" I mean not perfect, but I'd get some useable images, at least 10% would be keepers and a lot would be in good focus.

Repeat the MFA, get a slightly different adjustment, and try again - the same. Many more repeats. . .

When I did a bunch of 10 frame sequences of images and adjusted the MFA between them, I could see changes due to big adjustments (5 or more), but still the focal plane shifted a bit in each set on the same number, more or less centered on the new adjustment point. This is not the case with the 5dsR, when it's on target, it stays on target at least 80% of the time.

Very strange, but others have noted similar effects with the 90D. So it definitely is the 90D. Yet others with other lenses get good results.

What you describe is reminiscent of a problem I had a few years ago with the EFS 17-55mm/2.8 lens. It had significant inconsistent AF at long distances. MFA was inconsistent. I did hundreds of tests with it on at least 4 bodies and the lens had 6 trips to Canon. The odd thing about this one was that most owners of the lens never noticed the problem. But a few did, confirming the effect. Eventually Canon confirmed what I was observing and declared it was considered normal. Fine, but the lowly 18-55mm never did the same.

These AF systems are very complicated and software-controlled. There is no point in trying to figure them out. What matters is the performance in your hands with your gear.

Canon fixed it! But it does seem like a significant defect on some 90d bodies and probably should be a warranty repair for anyone noticing it. I have back-up on that from an employee of LensRentals, where I've been buying used equipment for years (including this 90d, the 600 and the extenders). The employee told me that they did some testing after I informed them of my troubles with both bodies and they found that 2 out of 10 or 15 90d bodies they tested with various 600/4ii and iii lenses. The mark iii version of the 600/4 had no problems at all with the bodies that failed on mark ii 600s in their tests.

Canon CPS repair says that they found that the sensor was misaligned and did $200 worth of fixing. It's definitely worth it for me, disposing of the camera would cost more than that in time and money. And, besides, I like the camera, it solves the single most problematic part of the 5dsR for me - the lack of an articulating LCD; the smaller file size is nice, too, and with no loss in resolution.

Immediately, on checking MFA on my standard fence wire, I could see that the repaired 90d no longer had a problem focussing on the twisted vertical strands that was impossible before. Focus locked-on and stayed put with only slight variation that could easily be due to the 3D nature of the test subject and slight movements between frames.

When comparing AF function among the different setups with/without 2x and 1.4x converters, the loss of accuracy with the 2x was very noticeable, especially compared to the bare lens (f/4 vs f/8), but accuracy with 2x was about as good as the 5dsR that was my previous primary body on that lens.

Now, I wonder why Canon didn't find and fix this problem the first time I sent it in.

FrankKolwicz wrote in post #19062051
Canon fixed it! But it does seem like a significant defect on some 90d bodies and probably should be a warranty repair for anyone noticing it. I have back-up on that from an employee of LensRentals, where I've been buying used equipment for years (including this 90d, the 600 and the extenders). The employee told me that they did some testing after I informed them of my troubles with both bodies and they found that 2 out of 10 or 15 90d bodies they tested with various 600/4ii and iii lenses. The mark iii version of the 600/4 had no problems at all with the bodies that failed on mark ii 600s in their tests.

Canon CPS repair says that they found that the sensor was misaligned and did $200 worth of fixing. It's definitely worth it for me, disposing of the camera would cost more than that in time and money. And, besides, I like the camera, it solves the single most problematic part of the 5dsR for me - the lack of an articulating LCD; the smaller file size is nice, too, and with no loss in resolution.

Immediately, on checking MFA on my standard fence wire, I could see that the repaired 90d no longer had a problem focussing on the twisted vertical strands that was impossible before. Focus locked-on and stayed put with only slight variation that could easily be due to the 3D nature of the test subject and slight movements between frames.

When comparing AF function among the different setups with/without 2x and 1.4x converters, the loss of accuracy with the 2x was very noticeable, especially compared to the bare lens (f/4 vs f/8), but accuracy with 2x was about as good as the 5dsR that was my previous primary body on that lens.

Now, I wonder why Canon didn't find and fix this problem the first time I sent it in.

Wow, hard to believe, and not what I expected. But it sounds great... better test it again tomorrow, though!

FrankKolwicz wrote in post #19062051
Canon fixed it! But it does seem like a significant defect on some 90d bodies and probably should be a warranty repair for anyone noticing it. I have back-up on that from an employee of LensRentals, where I've been buying used equipment for years (including this 90d, the 600 and the extenders). The employee told me that they did some testing after I informed them of my troubles with both bodies and they found that 2 out of 10 or 15 90d bodies they tested with various 600/4ii and iii lenses. The mark iii version of the 600/4 had no problems at all with the bodies that failed on mark ii 600s in their tests.

Canon CPS repair says that they found that the sensor was misaligned and did $200 worth of fixing. It's definitely worth it for me, disposing of the camera would cost more than that in time and money. And, besides, I like the camera, it solves the single most problematic part of the 5dsR for me - the lack of an articulating LCD; the smaller file size is nice, too, and with no loss in resolution.

Immediately, on checking MFA on my standard fence wire, I could see that the repaired 90d no longer had a problem focussing on the twisted vertical strands that was impossible before. Focus locked-on and stayed put with only slight variation that could easily be due to the 3D nature of the test subject and slight movements between frames.

That's good news. I could never recommend the 90D to someone who required 1Dx-series AF ability, or even say that it was as good as the 7D2 for large-pattern AIServo, but always thought I could recommend it as a single-point sharp-shooter's machine with the best focal-length-limited IQ that Canon provides.

F/8 AF requires pretty favorable AF conditions with the 90D. When there is no contrast, it can hunt, and when the light is low, it can be pretty slow even if it doesn't hunt. The results are amazing, though, when everything comes together. I had a Spotted Sandpiper walking the rocks at a damn last week, at 20 to 30 feet from me, in good light, and I left the 2x on the 90D and 400/4DO II, and it focused more than 50% of the time on the eye on this very active bird, even when it was moving towards me (single-point AIServo). In other situations, I might have noticed the TC's drag on the AF and stepped back to 560mm or 400mm, for acceptable AF. IMO, loss of AF is the only reason to crop instead of using a TC.

Someday, we may be viewing our wildlife photos on very hi-res displays, and those big-pixel captures that so many people seem to drool over are going to look compromised compared to the best photos with high pixels-on-subject.

BLUE ANGELS

Bullfinch UK Wildbird

Canon 90D 100-400mm IS USM Mk2

iso:125 s: 1/125 f:5.6

Great Blue Heron.

The 90D is in the POTN database!

Goldfinch UK Wildbird