« Cheap Nikon Good | Main | Open Mike: High-End Loudspeakers (OT) »

Saturday, 14 July 2012

Comments

I think I can see the entire sensor rectangle. Not cool at all!

I think Mike is incorrect on this, not simple lens flare but the "red dot" issue that some have reported with one or more of the Pens. Google Olympus red dot or red spot and you will find a bunch of discussions. I have never seen it myself ( E-PL2, E-M5), but I've never shot directly into the sun.....

Since I've owned both the K-5 and the E-PL2, I've noticed that Pentax does an excellent job of minimizing flare with their lenses, better than the Olympus kit lens. OTH the Olympus 12mm is superb.

That's not lens flare. It's a known problem with the E-PL2 that shows up under rare circumstances.
Google "e-pl2 red spots" for some discussion.

An even better idea is to turn the camera around and photograph the light that such beautiful sunsets cast for just a few moments! Flare problems gone!

I agree that it looks like flare to me, too. Lens coatings really have become extremely adept at minimizing flare during the digital age. A couple of weeks ago I was using an old Canon FD lens and was really struck by how flare-prone it was, especially it seemed in LED lighting. I had to be -very- careful not to pick up a ding and also to avoid any glare from my subject. We've come a long way in the lens design and coatings arena!

The other remark I wanted to make was regarding how funky today's lens flares are compared to the old-time lenses. Look at Peter's sample, the rectangular pattern displayed. I was accustomed to the angular repetition of colored cusps that moved from the lens center as the light source/lens relationship changed. But I've noticed odd patterns like this on relatively new camera designs. Perhaps it's the newer lens designs? Perhaps the sensor really is also involved (since, unlike film, it's loaded with little lenses). Whatever the engineering explanation, today's flare is not your grandpa's flare!

This issue was complained about with the usual fervour the internet crowds are capable of. This thread here has more: http://www.seriouscompacts.com/f2/fyi-e-pl2-red-dot-problem-1218/

Like Mike, I'm also kind of disappointed that modern-day coatings are super competent against flare; back then we had to guard against it even in the studio, using hoods and flags and whatnot. We learnt to use it creatively.

Of course, these days, people use Photoshop to add in "creative flare". Which is nothing like the real thing and overhashed....crap.

This is actually not lens flare, but a particular anomaly of the EPL 2 low pass filter.

There's been some discussion of red spots with the E-PL2 in other forums, for example: http://www.flickr.com/groups/olympusesystem/discuss/72157627929644003/

I don't think anyone has provided a satisfactory explanation, although I lean toward Mike's lens flare version. Flare can result from reflections from any surface so red reflections from the sensor could be the cause.

Beautiful lens flare. I think this makes the photograph interesting. I would explore it more. The iPhone has some unique flare. I have been exploring it for some time now. I have found that when taking a photograph with the bright sun in it exposing for something else other than the sun you get this beautiful flare.

Noto to editor: I don't know if this is appropriate for this blog, but I'm including a link only to illustrate my point, not for self-promotion.

http://jpgmag.com/photos/3376176

Sorry to Spam you with another comment, but here's a better discussion: http://visualsciencelab.blogspot.com/2011/01/can-we-talk-very-frankly-about-serious.html. Or just for fun, search Google for "e-pl2 red spots."

Those red dots in the first pic definitely do not look like lens flare. More like a reflection off a focus screen or sensor or something. Lens flare would have a single set of reflections on a single axis, generally, not multiple reflections around multiple axes....at least in my experience.

Ed

The sensor flair as you call it is a pretty well known phenomenon. I believe it was discussed in a post on TOP some time back.

I have a few examples of this taken with various 12mp Olympus m4/3 cameras. You have to have the light source at just the right angle to produce it. It is not generally a problem for me.

Actually, it isn't flare as such. The E-PL2 is, as is some other cameras, such as the Sigma DP1, known to produce this pattern of red dots when shot against the sun. Apparently the fault lies with some part of the sensor and is not related to the lens, as is ordinary flare.

You're not supposed to take pictures of the sun. You could go blind. I just read that in my 7D manual. Now stop it.

There is no 14-50mm kit lens (as claimed) for the Olympus E-P2.

I'm not convinced that the red dots are due to lens flare. The number of dots, and their regular spacing in rows and columns, hints at some effect from the sensor array. I hope Ctein has a theory on this.

Interesting article, thanks. I have a very nice, very flare-prone Olympus 7-14 zoom (14-28 full-frame equivalent). It flares dramatically anywhere near the sun, and of course anywhere pointed away from the sun you can see your shadow, so that leaves a little sliver of shot possibility with neither. Or cloudy days. But it sure is sharp.

This particular type of flare appears to be more common in digital cameras. I encountered particularly ugly - by that I mean unattractive aesthetically - flare when I had Nikon D100/200/300s and Nikon zooms, esp the 18-200. I remember some discussion starting ~5-6 years ago regarding this and it may be a result of reflections off the sensor or filter back to the rear element of the lens. I searched and found many discussions on this topic.
You may find this 2005 article interesting: http://www.shutterbug.com/content/digitally-optimized-zoom-lenses-do-they-really-make-difference
Remember that film is quite non-reflective and very diffuse, but some of the elements in the sensor can be quite reflective. Fresnel reflection off the boundary of glass and air is ~5%!
I wonder if the D800 and D800E are different? That would tell us a lot.

Funny, but back when I got into photography, it used to be assumed that pictures with the sun in them would show flare; now, coatings are so good that people can have the opposite assumption.

Right!

Dear Mike,

I'm with you 100% on this one. The technicalities don't matter. (Yes, I am pretty sure I understand what causes this and no I'm not going to discuss it because it DOESN'T matter. It's an irrelevancy.)

This isn't an “issue,” it isn't a “problem,” it isn't a “defect.” It's just something that happens on occasion that falls under the heading of “cameras can't make good photographs of everything you see.” Anyone who thinks a camera should never fail to make a satisfactory photograph under any circumstance, else it be deemed faulty, had better give up photography now. There has never been such a camera and there isn't one yet. Maybe there will be in the future, but in the meantime, take up painting.

Understand I'm using "camera" very loosely here: I really mean cameras, camera bodies, films, digital sensors, and lenses. The whole shebang that goes together to make a photograph. They're not ideal. There will be times when they fail. Those times aren't very often anymore. I probably run into those times more than most, because I'm always trying to figure out what I can get away with. What you do about those times is to learn to avoid them; It's part and parcel of learning what your equipment can and can't do. That's been inherent to the craft since day one; learning what your equipment can and can't do for you.

It's just like the old joke goes:

Doctor, it hurts when I do that.

Then don't do that!


pax \ Ctein
[ Please excuse any word-salad. MacSpeech in training! ]
======================================
-- Ctein's Online Gallery http://ctein.com 
-- Digital Restorations http://photo-repair.com 
======================================

I am just asking here without commentary: Mike, would that veiled flare in your indoor shot be reduced if you had a polarizer over the lens? So would a polarizer on the first photo remove the red dots from the images?

I am reading the 1938 Kodak book on Polarizers to refresh my approach on some upcoming product photography. Kind of a fun read with it mentioning the recent first practical film polarizer developed by a "Mr. Edwin H. Land". Given his number of patents, how many of his photographic developments, beyond the cheaply manufactured polarizer, are still in use?

CHEERS...Mathew

Dara J: "You're not supposed to take pictures of the sun. You could go blind. I just read that in my 7D manual. Now stop it."

Nuts. It seems like blindness is a potential consequence of so many enjoyable activities.

Flare would be circular in pattern when the sun is centered in the frame, right? A LPF coating anomaly would occur in a grid pattern, like what you see in example 1, imho.

Not sayin I am right. But that was what I could see from the examples and the fact that other PEN models with the exact same lenses don't exhibit the same issue.

Flare, on its own, is wonderful, and I have nothing against it, love it actually. But gridded flare, thats where I draw the line.

I once brought my Nikon D300 and the D3 to a family gathering, one shot with the D300 showed flare in a dark hotel foyer with large windows directly behind the subject. Swapped the D300 for the D3 with the same lens, a 50mm f1.4 afd, and there was no flare, this led me to surmise that there was a difference in the coating on the sensor. Either that or possiblely a different refraction index on the low pass filter.

The little Olympus is an inexpensive camera, so the sensor coatings might be added within specific tolerances in order to keep the price per sensor down. The voigtlander m lenses give off lovely flare, very dreamy...

I took another look at the original photo and have a thought - inspired by optics classes many year ago. The regularity of the ghosts look suspiciously like diffraction. Note how the ghosts are all on a grid? That's indicative of diffraction. Shooting into the sun means you will have minimal exposure with the smallest aperture which is what causes diffraction. The shorter lens focal lengths of some of these digital cameras means that apertures are smaller. Does Peter C. know what the camera settings were?
Want to see what diffraction can look like:
http://www.flickr.com/photos/close2far/3276646969/

All the sun / lens flare I'm familiar with takes the form of images of the aperture on a single axis passing through the center of the image circle and the image of the sun, so unless he's using an something like an imagon, I don't think that's lens flare.

I would guess that it is a signal processing artifact. I bet those spots have something to do with a combination of interger overflow in the raw data compression exaggerated by tone mapping in the post processing.

I wonder if that shows up the same way in all demosaicing software.

With all due respect for some people, but a reality check is needed here. The lens costs what, $200 extra as part of a kit, it's pointed straight at the sun (in a place where the sun is bright, I might add as a true northerner), a flare phenomenon happens and people are all over it on the net. Seriously, focus on photos themselves or start a long cycle of purchasing lenses that cost four figure sums; no lens and camera system is perfect. I was trying out a $5000 kit yesterday and yes, it had technical faults, but I would still classify it as a superb kit; there are always compromises.

The second thing is that if people start saying that it's not flare then please explain concisely but accurately what you think it is then. I do have a science degree, I'm seeing a purely optical phenomenon and would genuinely like to hear if there is compelling evidence that it is something else.

Well, looking at that again on a monitor larger than my iPhone, it looks a little irregular to be bit wrapping in the signal processing , but unless there is some repeating rectangular grid structure in the lens, it's coming from somewhere else.

I've seen weird non circular lens flare from an old C-mount Angenieux lens with a beamsplitter viewer where the reflections from the focusing aid would show up on film, and some whacky artifacts of the compendium hood on a Hasselblad but without some sort of unusual accessory filter or rectangular lens hood that is generating a refraction pattern, this is coming from behind the lens.

One image with many flares:

http://www.photokinesis.info/Other/Share/i-GR7G8n5/0/L/IMG0208-L.jpg

Here is a vivid example of an intense red-dot pattern that showed up on an early-morning shot I took with the Canon s90. That's not all, in addition to the symmetric pattern of red dots you can see veiling flare, ghosting flare, and 6-point diffraction star from the aperture.

Permission to use.

My Tamron SP 17-50MM F/2.8 Di II LD Aspherical (IF)has significantly less flare than any of my other 3 lenses, all Minolta from the the early 80's. When shooting under conditions that will cause flaring, I either expect it or take precautions to avoid it, Sometimes it works other times it doesn't.

Thanks everyone. Sorry, it IS the 14-150mm ED kit lens, slip of my fingers.

My field is/was electronics. I don't know enough to be sure or give an explanation, but I still think it's a sensor artifact rather than just lens flare. It's my gut feeling here, and I've got plenty of gut!

The regular pattern to the dots is not like any lens flare I've ever seen. I'm now going to my files to dig out all the examples of flare I can find. I have one absolute beauty in mind right now. It's flare I liked enough to deliberately aim for it. Coming up asap.

Jeremy Fagan, desirable red dots :-) Like it, like it. Trouble is, they need to be visible to everyone else. I know, I may make my own logo out of this pattern.

Hi!

@ Jeremy Fagan: "Some people pay a fortune...."

Quite possibly the only time I have ever honestly written "LOL."

Dean

More on the diffraction pattern theory:
I couldn't stop thinking about this so I did an experiment. I set the aperture of a Oly 14-42 lens to f/22 at the shortest focal length and shot a picture of it. (http://www.jimhayes.com/photo/Oly-Lens.jpg)
The aperture is about 1mm and looks hexagonal. That's very small.
Note in the photo above the sun is right in the middle of the photo and the red dots are surrounding the central image of the sun symmetrically. The red dots also appear to be about the same size as the sun. When the sun is so low on the horizon, the image of the sun is quite red due to the high attenuation of the bluer light in the 6-7 air masses (6-7 times the path length of air one would see looking straight up). Diffraction images are generally bright in the center and get fuzzier to the edges. I'm betting this is a diffraction pattern.
I'll bet you will see this pattern anytime you shoot directly into the sun, but unless the sun is peeking though dark clouds you may not see it.
In normal shooting, the same phenomena occurs but it only causes a slight blurring of the image because it's integrated across the entire image.
Finally, I searched for some references but most require some knowledge of physical optics. However this animation may be interesting http://www.youtube.com/watch?v=Txgo5Nsv4Lg

I've seen this type of flare before (red dots, repeating pattern, when bright sun is in frame.) It occurred from time to time with a friend's cameras, a Canon 550D and later a 50D. They both flare the same way with the sun in the frame.

I haven't played with her cameras often enough to find out whether the flare is due to the lens or the sensor, though.

xfmj,
That's very cool.

Mike

First a brief note on this statement: "...forums aren't the right place to "decide" technical issues, because all they do is create group consensus."
It's absolutely true, as is the conclusion that sometimes consensus is often erroneously obtained. One example is the conclusion that there is an "equivalente aperture", analogous to what happens with crop factor, so that an f2.8 aperture with translate to f5.6 when the lens would be mounted on a micro 4/3 camera. It's rubbish, of course - but a wide spread one.
More to the point: if my readings are anything to go by, the red dot issue is a kind of flare unrelated to the heavily blown light shown in the last photograph, as they might have different causes. The latter could be a result of digital dynamic range, in that the sensor has a linear response, as opposite to film's smoother and more extended DR curve. Hence the tendency shown by digital cameras to clip highlights easily. (Cf. Michael Freeman, 'Mastering Digital Photography', Ilex, p. 627.) However, this theory must be taken 'cum grano salis': as in all areas of human knowledge, nothing's written in stone.

This is been a 'major' issue with Sigma's compact cameras (DP1, DP2) for a long time. My guess is that it's a combination of internal reflections playing off of the microlenses, and the IR cut filter may be contributing to the coloring. It's interesting in that the Sigma has no OLP filter, and it's IR filter is not bonded to the sensor (but rather sits farther up in the optical path).

However, I agree with the sage advice of "don't do that."

Cheers to Ctein. Not only did he write what I was thinking, he did it better as well.

I'm curious as to whether the OP is "pretty disappointed" in Olympus or himself for not choosing equipment better suited to his preferred subject matter. My guess is the former. Since "red dot hysteria" has been forum fodder for months, perhaps it should be the latter? I bet his local camera store would have gladly let him shoot into some strong lights (and what better place to find them!) to simulate this type of backlit shot.

Sometimes, a poor outcome is someone else's fault. Not always.

. . .These three shots are from the reject pile from my recent vacation. . .

It is manifestly unfair that your rejects look better than my keepers.

Lens flare is a common artificial artifact of video games and CG movies

Totally unrelated (but related to the post title, which got me excited). Isn't it interesting that UFO sightings have dramatically decreased now that everybody has a camera with them ALL the time, and that cameras can shoot at incredible ISOs?

I've been using an E-PL2 as my primary camera for about 18 months and I've had the "Red Dots" show up a few times and I've even tried to cause it a few times. My experience breaks down this way:

1. The problem seems to be the worst with small apertures. I never notice it a f4 or larger. I do see some red flare occasionally but no distinct pattern.

2. The problem seems to be worse at longer focal lengths. It is much more likely to appear and be noticeable with a 45mm lens than a 17mm lens.

3. The problem also seems to be worse with a very strong light source with a lot of red in the light surrounded by a darker area. Sun through the clouds, a single bright light in a dark room, the sun between two trees all see to show the pattern more than a low sun on a clear day. It may be there other times but not enough to be noticeable but it does require a strong light source. I have never had the problem with mercury vapor or fluorescent lights.

I had an E-PL1 and I have an E-PM1 and don't get the red dots under the same general conditions. Since I rarely shoot at high apertures or directly at the sun, it hasn't been a big issue for me. I just consider it to be like lens flare and shoot to avoid it.

Per what Ctein wrote and Oskar Ojala.. one of the comments from the OP at the DPReview thread posted at te end states :

"While I understand I can mostly work around this...I don't think I should have to deal with that on a $600 camera system."

Expectations are sometimes a lot higher than I'd imagine...

I may the only person seeing 'UFOs in the Sunset' and hearing 'Tubas in the Moonlight.'

http://www.youtube.com/watch?v=qm0a7btAtsI

The mind is a wonderful thing!

I think it is a religious sign for us to begin to worship the allmighty Red Dot :D

Well, from what I've seen (30 yrs with film, 12 with digital) you can only get radial flare patterns from a cylindrically symmetrical lens.

A rectangular array of light spots requires a rectangular array of diffracting elements to cause it.

Regular lens flare is reflection, not diffraction, and is a quite different thing (except the diffraction from the edge of the diaphragm blades, which creates the 'star'.

I've got to disagree, this is a terrible technical issue, and totally unacceptable.

I'd suggest any impacted users send their E-PL2s to me for immediate disposal; I'll make sure they're dealt with appropriately. It's the least I can do.

I was not sure what it was when I first saw it, but this is the strangest flare that I have seen, and the light source, the sun, was not even in the frame. It was almost directly overhead though, and I did not have a hood.

http://www.flickr.com/photos/tanguerochino/7247091512/in/photostream

What amuses me is that as stills cameras and lenses get progressively better at suppressing flare, movie makers are busy adding fake flare back in to their CGI creations to make them look more "real". The example which springs immediately to mind is the last Star Trek film, where every time a light source crept into shot you saw a great flare effect, even though in many cases these were obviously painted in after the event.

The comments to this entry are closed.