Canon R8: AF box confirmation is rarely on the subject in servo mode

[SumanV]

Hi. My R8 + 400mm lens combo is not giving sharp pictures when shooting birds and I realized that the AF point is rarely on the subject. Yesterday, confirmed my suspicion. When I shot bursts, I noticed that only the first picture in the sequence shows the AF box on the subject. The box is not on the subject for the other frames in the sequence. In the EVF, however, the tracking box appears to be on the bird 90-95% percent of the time. I've set a shutter speed of 1/2500 s and an aperture of 7.1 to minimize camera shake. Despite this my pictures are soft and have this AF box problem. One thing to note is that I'm shooting in EFCS shutter mode and the burst is set to continuous (no H+ or H). That should not be a problem though. I'm attaching a screenshot showing the problem (this wasn't shot yesterday).

I am struggling to understand why this is happening. I've seen reviews where people have praised the exceptional tracking of the R8 but I am not seeing it (at least via the images). Why would the tracking be accurate on the EVF but not on the images that were shot? If it's a lens issue, then shouldn't the problems also appear on the EVF? Could there be an error with the AF? It may be that the focusing system in the camera is tracking but the lens is unable to keep up?

Can someone from Canon, or an experience with R8, help me?

SumanAF box appears elsewhere

[Waddizzle]

---

@SumanV wrote:

Hi. My R8 + 400mm lens combo is not giving sharp pictures when shooting birds and I realized that the AF point is rarely on the subject. Yesterday, confirmed my suspicion. When I shot bursts, I noticed that only the first picture in the sequence shows the AF box on the subject. The box is not on the subject for the other frames in the sequence. In the EVF, however, the tracking box appears to be on the bird 90-95% percent of the time. I've set a shutter speed of 1/2500 s and an aperture of 7.1 to minimize camera shake. Despite this my pictures are soft and have this AF box problem. One thing to note is that I'm shooting in EFCS shutter mode and the burst is set to continuous (no H+ or H). That should not be a problem though. I'm attaching a screenshot showing the problem (this wasn't shot yesterday).

I am struggling to understand why this is happening. I've seen reviews where people have praised the exceptional tracking of the R8 but I am not seeing it (at least via the images). Why would the tracking be accurate on the EVF but not on the images that were shot? If it's a lens issue, then shouldn't the problems also appear on the EVF? Could there be an error with the AF? It may be that the focusing system in the camera is tracking but the lens is unable to keep up?

Can someone from Canon, or an experience with R8, help me?

SumanAF box appears elsewhere

---

Where is the AF point located releative to the subject?  Zooming in on the bird has hidden the AF point(s).  Can you post a RAW file of the sample image to a public file sharing platform like Dropbox, etc.

Exactly which 400mm lens are you using? Judging from your screenshot, it seems you are using the EF 400mm f/5.6L USM.  If so, then the lens is the most likely cause of your issue.  It does not focus fast enough for the highest frame rates the camera is capable of using.

If you have downloaded a copy of the User Guide, somewhere in the first 5-20 pages you should find a page with a web address link and a QR code.  Use these to download a Supplemental Information Guide for your camera.  It should contain a table of EF lenses compatible with max frame rates when using H/H+ drive modes with ECFS.

Lenses with original release dates prior to 2009 are typically not included in the lists.  Here is a sample.

[SumanV]

Hi. Thanks for the reply. I will post the RAW images in some time. I set the camera to a low-speed continuous drive mode in EFCS shutter. That gives me 3 shots/s. 

My main issue is that the tracking seems to work fine as seen in the EVF, but the images tell a different story.

[p4pictures]

One important point is that the AF area that can be used with the EF 400mm f/5.6L USM is variable according to the information from Canon https://cam.start.canon/en/H001/supplement_0110.html 

It is possible that the AF cannot use the whole area of the frame, and your subject is rather small and near the edge of the frame. (Judging from the thumbnail on the top left, and the 100% zoom below the image. I don't know what the specific AF area is for that lens 7 EOS R8, but it might be a factor to consider. 

I have some examples myself where the AF point recorded does not match the subject. I was using the similar EOS R6 Mark II (shares the AF system with EOS R8) and a newer EF 70-200mm f/2.8L IS II USM lens. I was photographing red kites in flight, and these birds are similar sized to what you showed. First one is with the image filling the screen, the second is 100% zoom. Bird is sharp too. 

[SumanV]

Thank you very much! It would be nice of Canon if they explained what variable area means. But going by the link you shared, most RF lenses have variable AF area. So I am assuming that it may not be a bad thing after all!

In the example you shared, the AF box is at least near the subject. That is not so in my case. Also, the image appears a lot sharper than mine. When I shoot a burst, the AF confirmation box appears on/near the subject for the first image. The rest of the images in the sequence are, for the most part, blurry and the confirmation box does not appear anywhere close to the subject. In the EVF, however, the camera shows that it is able to track as there is a blue square on the subject.

Do you also notice a discrepancy between the tracking shown in EVF and the AF confirmation box from the images?

I'll shoot with my RF lens and check if I have this issue. If the problem does not appear then will this be a confirmation that my EF lens has a problem?

[Waddizzle]

Brian raises a good point about AF area. The preview image doesn’t show AF points.  But your main display area can.  Your subject is too far away for a clear shot.

 

Do you experience AF issues with other lenses?  

[SumanV]

@Brian and @Waddizzle: I now understand what you mean. The subject is close to the top of the frame (in the thumbnail). Is it possible that in this case, the EVF also shows that the tracking has failed?

[wq9nsc]

I love the EF 400 f5.6, it was the first telephoto prime I bought and I still have it and use it occasionally.  It is very sharp and focuses rapidly, even a few years ago I was still using it for some daytime outdoor soccer matches when I didn't want to use my heavier 400 f2.8 but now the versatile EF 200-400 f4 with built in 1.4X is my long lens of choice for outdoor daytime however I belied the relatively low cost EF 400 f5.6 provides the same high image quality and fast focus acquisition as the much more expensive sports zoom.  I believe the 400 f5.6 was always the best value in Canon's EF line for wildlife and sports.

But it does impose some limitations with camera focusing systems regardless of architecture.  For example, Canon used a letter rating system with DLSR cameras and the EF 400 f5.6 is a "group E" lens.  My 1DX III uses a 191 point AF sensor in viewfinder mode and 155 of these points are increased accuracy cross-type point with a dual cross center point.  But with the E type lens only the center array of 65 sensor points behave in the cross type mode and the remainder are horizontal sense only.  The AF manual doesn't state any additional limitations for this lens when using it in live view mode with the 1DX III and AF coverage extends to 90% horizontal of the image sensor and 100% vertical when using the image sensor for AF.  But there are likely additional limitations depending upon the image sensor in the camera when working with "slower" aperture lens like these much like there are with the separate AF sensor and these lenses in a DSLR.

Looking at the small part of the frame occupied by the bird in your image, you are shooting at a very long distance from the bird.  In a case like this, you may find better results by capturing focus and then using focus lock while then tracking the bird rather than having the camera constantly hunting focus. 

At 350 feet, the depth of field for critical focus with a 400 f5.6 on your camera is 48 feet.  I don't like using the two button focus approach because I shoot sports with multiple bodies often switching bodies within a single play.  So I have my AF tied to the shutter release but I reprogram the AF button on the body to AF stop so that if I want to lock focus I use that button (or the handy focus stop buttons that are on the far end of my long white prime lenses).  That way I can grab perfect focus in difficult situations like making sure that the keep in soccer remains focus locked regardless of the traffic around the keep.  It may prove equally useful when tracking a bird at great distance when it is only covering a small area of the sensor.

Rodger

I suspect the frame rate limitation on the R8 is a combination of the power draw of the lens and its aperture stepper and focusing motors in conjunction with the small battery used by the R8 and the higher power budget of a mirrorless body.  The 400 f5.6 has no problem keeping up with the 16 FPS high speed capture mode of my 1DX III and I expect it would be fine at the 20 FPS rate in live view but I never use live view for action.  But the larger bodies 

[SumanV]

Hi. I have uploaded some raws on my google drive. You would see that the AF confirmation box is rarely on the subject. Is this a problem? Do you think that the subjects are in focus? 

https://drive.google.com/drive/folders/101jHjXsxDWTVpd8EbpL5IU_mdopeT3-D?usp=sharing