SL3 takes pictures slightly rotated

[dheldb]

I’m writing form Chile. I bought an EOS REBEL SL3 camera and noticed that the pictures were taken slightly counterclockwise rotated when using the autofocus points to define the horizontal level at the viewfinder. I went to the local technical service, where they confirmed the problem and replaced the camera with a new one, but that new camera has the same problem. After discussing with the technical service, I realized that there isn’t more they can do and I accepted the replacement.

Did someone else face the same problem? The way to find this out is e. g. to take a picture at the see, or a lake, and using the autofocus points at the viewfinder to set the see level horizontal, take the picture and then check the result at the screen or download the picture to your computer and check it there. The rotation is less than 1°. For pictures at places where there isn’t important to have the horizon horizontal, this isn’t a problem, but when you take a picture to the see, and the see horizon appears slightly rotated, this is somehow annoying. Now, every time I take a picture at the see, I have to calculate how much I have to rotate the camera to compensate this situation.

Just as reference, I have had other Canon reflex cameras, but with none of these I faced this situation.

According to my experience, considering that both SL3 cameras I checked have the same problem, it seems to be a manufacturing problem.

Thank you for your comments.

[stevet1]

dheldb,

Have you tried turning on your grid lines rather than depending 9n your autofocus points?

In your menu settings, it's in the "wrench" menu, page 4.

Steve Thomas

[dheldb]

Thank you Steve. Yes, I tried. When using the grid lines at the camera screen and I take the picture in that way, the problem doesn't happen. Moreover, if I use a tripod and set the horizon horizontal with the auto focus points at the viewfinder, and then switch to the screen with the grid shown, not having taken the picture, it can be seen there that the image is slighly rotated.

[stevet1]

dhelb,

Perhaps your autofocus points are slightly offset, and trying to level your shots using those gives you a picture that is slightly rotated.

Maybe you could eliminate those, or just not display them, and just use your grid lines.

Steve Thomas

[dheldb]

The autofocus points are rightly aligned with the frame of the image seen through the viewfinder. It rather seems to be a pentamirror alignment problem in relation to the camera sensor.

Thank you Steve

[jrhoffman75]

If you use the grid lines, which are intended for framing and alignment, do you have the same problem?

[dheldb]

If I take the picture using the camera screen with the grid, and align the horizon horizontal with the grid lines, then there isn’t the problem.

Moreover, if I use the viewfinder with the focus points when focusing to take the picture and align the horizon horizontal with those focus points, and then switch to see the image through the camera screen with the grid, at the screen it can be noticed that the image is slightly rotated. I. e., the viewfinder “sees” the image slightly rotated compared with the camera screen.

After reading your comment that the focus points that are shown in the viewfinder are an LCD overlay, I rechecked if maybe those LCD overlayed focus points may be slightly rotated, and not the complete image seen at the viewfinder, for what I searched an object with several parallel lines and did the test with it, doing one of those lines pass through the focus points and another parallel close to the image frame seen through the viewfinder, and switched again the image to the camera screen. Even though my tests at my home aren’t as exact as the ones that can be done in a laboratory, the result I get shows quite clearly that the focus points are well aligned to the image frames seen through the viewfinder. I. e., the problem isn’t the focus points but the complete image seen through the viewfinder.

Therefore, it seems that there is an alignment problem of the pentamirror system (including the reflex retractable mirror) that makes that the image shown through the viewfinder is slightly rotated.

[deebatman316]

Can you please post a few example pictures in the forum so forum members can better assist you.

[dheldb]

Thank you Demetrius. I don't have now pictures with the see horizons slighly rotated (once I detected the problem, I took the pictures trying to compensate this problem by slightly rotate the camera). But now I did a test at my home, using a tripod and a table edge as horizontal horizon. It is not a professional test, but I tried to do my best to show the real problem. In that picture, the table edge shoud be horizontal (I used the autofocus points at the viewfinder to set it horizontal). To see the picture rotation, I drawn a thiny red line on the picture that shows the horizontal level. There it can be seen that the table edge is slighly counterclockwise rotated compared with the horizontal red line.

[jrhoffman75]

---

@dheldb wrote:

Thank you Demetrius. I don't have now pictures with the see horizons slighly rotated (once I detected the problem, I took the pictures trying to compensate this problem by slightly rotate the camera). But now I did a test at my home, using a tripod and a table edge as horizontal horizon. It is not a professional test, but I tried to do my best to show the real problem. In that picture, the table edge shoud be horizontal (I used the autofocus points at the viewfinder to set it horizontal). To see the picture rotation, I drawn a thiny red line on the picture that shows the horizontal level. There it can be seen that the table edge is slighly counterclockwise rotated compared with the horizontal red line.

---

It's not surprising to me that this is happening. I think its a matter of manufacturing tolerances.

Shrink the picture you posted down to the size of your camera viewfinder and the little gap at the left of the image will be almost, if not completely, unnoticeable.

The focusing points that show in the viewfinder are not actual focusing points, they are an LCD overlay. There is a limit to the precision that can be achieved. Even the 1D X Mark III, which has an electronic level in the viewfinder, has a +-1 degree accuracy limit and that is using actual sensors in the camera. I put a protractor on your image and it looks like your deviation is about 3 degrees.