Interesting highlight halos

Anyone ever see rainbow halos like this in the images?

 

I know bright lights often have halos around them, but this is the first time

I've noticed multicolored "rainbow" halos like this in an image.

 

They were shot using the EF-S 18-55 kit lens, but even so I was still a bit

surprised to see them!<div></div>

<P>I have heard of people getting this effect before, yes. The examples I saw looked exactly like that.<br>

A clue may be that it has always been with the EF-S 18-55 lens.</P>

<P>I wonder if perhaps some of the elements are not actually optical glass (for cost reduction), but maybe some synthetic material which is refracting the light.</P>

Nope, they are all glass as far as I know. I think there is an aspheric element but I'm not sure what type (moulded, ground or plastic covered).

 

However even with plastic elements it's not immediately obvious to me what the optical origin is.

 

Clearly it's splitting the light into it's component colors, so that has to be done either by dispersion (like a prism) or diffraction by reflection from a periodic structured surface. I suppose the sensor could be acting like a diffraction grating, but the optics still escape me!

 

I'll have to see if I can replicate it under controlled conditions and test whether it's only the 18-55 kit lens that produces these halos.

 

Halos on film can be caused by light reflecting back off the far side of the film base (hence anti-halation coatings), but I'm not sure about digital. There are certainly a buch of layers of different optical materials in from of the sensor which could reflect light back which itself was reflected off the sensor. If the original light was bright enough, the reflections, even though faint, might be seen. I wonder if the anti-aliasing filter could do something like this with back reflections. However that would be independant of the lens in use.

I've seen rainbow halos in non complete circles before. That was also with city lights I believe. They were quite far from being a complete circle. More like only a 10 degree arc or something small like that, but it was obvious that it was an arc form with center pointed at the light source.

 

As for what caused what you are describing? There is dispersion through pretty much everything. It's not surprising to imagine a possible light path where out of focus light (due to multiple reflections for example) ending up on the sensor with a rainbow effect. Anyways, keep us posted. I'm interested if you ended up spending the time to draw a more mathematical solution and seeing what it is.

I might've guessed misty rain or a mist on your lens' front element?

Just an off he wall idea..

 

was the due point within 2 or 3 deg of the air temp? I have had this happen with aircraft landing lights at night, but only when its cold and the TAF is forecasting mist for early in the morning.

 

Maybe micro droplets of water in the air ... would be exactly the same as a rainbow - which are a full circle if seen from the air!

 

Simon

I would think about the mist in the air too. But it is interesting. Please keep us posted if you find more.

I was thinking it was an atmospheric optical effect, too. In order to generate a complete circle, the particle causing it has to be symetric, so I'd guess some kind of water droplet.

Murphy's law: "If you tried to by a filter or lens to get that effect you wouldn't be able to!"

I don't think it was in the air. It was around 7pm in New York city on Thursday evening. There had been no rain all day and there was no rain that evening. It didn't look misty to the naked eye and it was still quite warm - no sudden drop in temperature.

 

I'd thought about some sort of atmospheric effect, but I'd kind of dismissed it as unlikely. There are all kinds or large scale atmospheric phenomena that give rise to various arcs and circles around the sun via forward scattering, though I'm not really aware of any that break up the light into component colors. Rainbows are a backscatter effect, where the light source has to be behind you. I don't think there's any form of circular rainbow around the sun (or moon). Many of the atmospheric arcs (like sun dogs) are far from a complete circle and don't usually show such intense colors. They also tend to form at a much larger angular distance from the light source.

<<I don't think there's any form of circular rainbow around the sun (or moon)>>

 

 

http://www.atoptics.co.uk/droplets/cormoon.htm

Hmm interesting.

 

Anyway, further tests showed the effect seems to be only seen with the 18-55 kit lens and it's visible in the viewfinder, which rules out an sort of internal or external sensor or sensor filter reflections.

 

I still don't know the optical explanation, but at least I know what isn't causing it!

Figure 7 on the following page is rather interesting. In this case the light appears nearly monochromatic.

 

http://www.assap.org/newsite/htmlfiles/Lens%20flare%20and%20orbs.html

Looks like some Cokin special effects filters.

Bob - I have done this with highlit subjects against black background in-camera by breathing on the lens and then waiting until it clears to the right amount and then clicking.

 

Were you inside an air-conditioned car or building prior to taking this? Maybe some residual moisture inside on an element?

 

Example here (trying to get an inline jpg to appear):

 

http://www.john.macpherson.btinternet.co.uk/Resources/usa3002a.jpeg

 

or here if I cant get the image to appear you'll have to click this link:

 

http://www.john.macpherson.btinternet.co.uk/fullimagecreatir.html

Another try:<div></div>

And another example - both of these are much more subtle than yours, but its there:<div></div>

It's simply the lens. I did bench tests with a very bright light source (super LED) and the circular rainbow flare is there every time. It shows up with the light both in the center and at the edge of the frame, wide open and stopped down and you can see it in the viewfinder.

 

It's obviously related to the optical construction of that lens, since the other lenses I tried didn't show the effect.

Ha! Thats interesting. But of course Bob you do know that you've probably started some long-living rumour about 'problems' with this lens - wonder how long it will take to appear all over the interweb with demands for a recall! :-)

I don't think this is an issue for Canon. This is a lens made down to an (amazingly low) price, and if under VERY extreme lighting conditions it produces some peculiar effect, so be it.

 

What I find really fascinating is the question of what the physics of this can be. Those of us with enough schoolboy optics to know how rainbows work would have concluded immediately that water droplets weren't involved, without even having Bob's assurance that conditions were dry, because the angles involved are far too small. Is it produced by an internal reflection in a lens element, perhaps? Can anything be deduced from the apparently very consistent way that separation of the colours in the 'rainbow' from the light source is a function of frequency? What happens if the zoom setting is changed? Bob, you're an optical expert, tell us what you think is going on.

Bob,is this lens the one that you modified to fit a 10D? Looking at the pictures on your web site I notice that on the unmodified lens, the rear protrusion appears to incorporate light baffle grooves.

 

Could this be causing the rainbow effect? Probably not but it is an interesting coincidence.

No, this is an unmodified lens.

 

I don't regard it as a "defect". It's a cheap lens and is certainly worth its price. It only shows up under extreme conditions (which is probably why I only just noticed it after owning the lens for 3 years).

 

Maybe they all do it, maybe it's just mine.

 

As noted above, I'm sure some people would pay $50 for a special effects filter that did this!

 

I've no idea as to what the physics/optics of this are. You'd really need to run a pretty sophisticated ray tracing program to figure it out (maybe not even then).

Maybe Canon is field testing a new design?

 

Heard once, .. that Chuck Westfall, had stated, Canon's R&R was working on new lens designs with water in the lens elements? Perhaps the Aspherical element in this one is a little different from the others.

 

Bob, thanks for the info, it is very nice for Canon to give me a "free" Halo filter for my lens!

Bob,

I also owned a 18-55mm MKI that showed the same effect.

The first time I noticed it was on a clear winter day with the sun low within the frame striaght ahead.

It produced the same strong rainbow effect.

I have never experienced anything like it with my other Canon lenses!

 

Could it be the lens coating, splitting the light into 3 different wave lengthes?