100-400 sometimes has a "glow" around distant subjects. Any idea why?

Sometimes, when shooting things at a distance - birds or planes, for example - I see a sort of halo/shimmer/glow around the subject. I do not see the problem at shorter distances. White subjects seem to really suffer the most, to the point of making the shot worthless. I do not have a filter fitted and everything is clean, shiny and calibrated. As far as I can see this is not a problem of poor focus or shake/blur. I have no idea what the problem is and thus no idea how to fix it. I've also seen the same problem using the lens with my 1D3, again with a white subject such as a swan or egret, typically against a background of grass or rushes.

Here is an example shot today with my 50D at 400mm, 1/2000, f/5.6, 400 ISO. I'm pretty sure IS was on, but not 100% certain. I've seen this sort of thing even with IS off and the lens on a tripod, so I am really puzzled. This was shot raw and processed through Lightroom with no edits. There is no highlight clipping at all. The angular velocity was so slow that I would hardly even call this a panning shot, so if IS was on, I doubt that mode 1 or 2 should matter.

HOSTED PHOTO
please log in to view hosted photos in full size.

HOSTED PHOTO
please log in to view hosted photos in full size.

The thing is, here are a couple of 100% crops taken with the same camera/lens in the last 24 hours. They don't have a white subject, but the first subject is a good distance away and I see no sign of the problem.

What is the problem with white things at a distance? Anybody?

HOSTED PHOTO
please log in to view hosted photos in full size.

HOSTED PHOTO
please log in to view hosted photos in full size.

Looks like a fairly severe case of halation to me. Underexposing on these bright subjects does help to control it, or alternatively stopping down a bit.

Kind regards, Wim

Thanks, Wim. As a raw shooter who likes to ETTR I do not find it acceptable to have to underexpose in order to combat this phenomenon. The leading edge of the bird's head is only reading at around 84%/84%/84% for RGB in Lightroom, yet still shows the effect. There is an area on the bird's back that gets up to around 98%, but nothing close to clipping.

The raw histogram (attached) shows at least 2/3 stop of room on the green channel before clipping, more on red and blue. At 400 ISO, and with a capture that is not overexposed, it is clear I am not suffering from blooming. I would hope there is a more satisfactory solution than simply underexposing. There is some clipping on the white feathers of the kingfisher example I posted, easily recovered, I'm sure, but no sign (or very little sign) of the phenomenon. The kingfisher was shot at f/6.3, so stopped down a little, but to be honest, f/5.6 is hardly a fast aperture, even if it is as wide as the lens goes. When shooting BIF you need the wide aperture to get the shutter speed you want while keeping ISO at low enough levels to control noise.

All the links I've followed so far on Google for "halation" suggest it is to do with back scatter from the backing of a film emulsion. Should/could such a thing apply to a digital sensor? Perhaps this is some phenomenon other than halation? Perhaps my lens is flawed and needs to go to Canon to be fixed. Perhaps I am flawed and have a fault with my technique. If technique is the problem then I'm happy to take that on board, but I cannot accept that I must stop down or underexpose in order to prevent this aberration, not from an L lens and 1 series body. Softness I understand - this, I do not.

HOSTED PHOTO
please log in to view hosted photos in full size.

Halation (something I had with my 28-75 until I sent it in to be fixed) or sensor blooming. Probably the former.

Cheers, bohdank. I take it the fix was good? I have a Mack warranty on the lens, which is now just over two years old, but there's only a point to sending it in if it is a fixable problem. If it's "within specs" then I won't waste my time or anyone else's by sending it in for nothing.

tdodd wrote in post #8336377
Cheers, bohdank. I take it the fix was good? I have a Mack warranty on the lens, which is now just over two years old, but there's only a point to sending it in if it is a fixable problem. If it's "within specs" then I won't waste my time or anyone else's by sending it in for nothing.

frankly i don't see a problem unless you are consistently getting bad results.

ed rader

Yes, it was fixed. I would send it in, if I were you. Provide samples such as the above.

Prior to sending my lens in I communicated with the Service Manager via email and sent a couple of crops to illustrate the problem. He said to send the lens in.

In my case, I emailed a crop of large store sign with white lettering on a medium/dark blue backround. It clearly showed the problem.

ed rader wrote in post #8336394
frankly i don't see a problem unless you are consistently getting bad results.

ed rader

Often enough to give me cause to doubt myself or my equipment. Of course, I don't always shoot white things at great distances, so 99% of the time there is no problem. But as things stand I may as well kiss goodbye to shooting small white things at great distances, like birds, or planes. At around 20m, no problem. At 50m+ or 100m+, problem.

I shall see about sending it in. Thanks for the replies.

EDIT : Actually, I just checked another shot from today at 400mm, 1/2500, f/6.3, 400 ISO, some whites clipped (gone completely with +8 highlight recovery in LR), but no sign of the problem. Might that suggest the problem is not halation? Or maybe I do need to stop down to f/6.3 for such troublesome conditions. Anyway, here it is, full frame and 100% crop, no edits....

(I know the composition is poo. I wasn't shooting this as a keeper, just a reassurance check on AF accuracy.)

HOSTED PHOTO
please log in to view hosted photos in full size.

HOSTED PHOTO
please log in to view hosted photos in full size.