LordKOTL
Senior Member
Inspired by This thread--specifically the quoted posts.
I figured this was the best place to put this because filters are indeed accessories.
There has been a lot of questions as to how much a UV filter would flare when put on a given lens, So I conducted an experiment to show how much flare a photo would have along different angles with respect to a singular light source with the lens bare, with a UV filter, with a lens hood, and with both a hood and a filter.
I say this experiment is “unscientific” because I could not eliminate every single variable in laboratory-like conditions. In some cases, the angle between photos is not exact. In some cases, the subject positioning is not exact. In some cases, the focus is not exact. This was done by a layperson (myself) in their garage. It’s not meant to be exact, but rather practical. Your average amateur photographer is not going to be concerned with exact angles where the difference between one shot and another is less than a degree. In my opinion the point of this experiment stands nonetheless.
The Setup
The setup was to mimic a singular directional light source which would be analogous to the sun. I conducted this experiment in my darkened garage where the only significant light source was the single directional light. Any small light leaks were negligible from a practical standpoint. I did this by setting up a Godox AD200 on a light stand at an end of my garage roughly along the seam in the concrete, with the light being pointed along the seam towards the other end of the garage. The AD200 was fitted with the LED head, which was set at full brightness. This provided the source of light which was very bright along a singular direction, but with the light spread out and reflecting off of the garage surfaces. I had the light stand at the lowest height setting. This put it in the frame, but at the high-end of the frame when the camera was pointed right at it.
Along the seam approximately 9 feet away, I positioned my camera (a Nikon D750) on my tripod (a Mefoto Globetrotter aluminum tripod), and leveled it with the bubble level on the tripod head. The tripod was set at its lowest height setting. For the experiment, the only adjustment made was with the pan function on the tripod head. I set the camera up so that 180 was in the direction of the light source and 90 was perpendicular to the light source—that way I could use the tripod head to dial in the angle towards the light source.
For a lens I used a Nikon Micro 105 f/2.8 VR lens, which was freshly cleaned before the experiment.. The main reason for this lens is that the lens’ filter thread diameter matched the UV filter I had—62mm. I also think this lens would be somewhat typical in terms of performance for consumer-grade lenses in terms of flare. This lens also has a lens hood that came with it, so the safe assumption could me made that the lens hood would be optimally sized for blocking any stray light that is outside of the lens’ angle of view from hitting the front element—since the lens is a prime lens and not a zoom lens, and thus the hood doesn’t have to take into account vignetting at it’s widest.
The filter I used was one I bought years ago when I shot film: A 62mm Nikon L37c UV Filter. This was also freshly cleaned before the experiment. In terms of performance, at the time, I believe it was considered mid-range—It wouldn’t break the bank but it did the job it was supposed to.
For a subject, I used the top of a road cone—which I had lying handy.
For the experiment I marked out approximately 8 feet from the camera 7 locations all at different angles with respect to the light source: 90 degrees (perpendicular) to the light source, and then at approximately 15 degree increments towards the light source: 105, 120, 135, 150, 165, and 180—with 180 pointed towards the light source. I marked out approximately 8 feet at each of those angles, and used pegboard pegs to mark out the angle and distance.
After that I snapped a cell shot of the setup (see below), turned on the Godox AD200 and turned the garage lights off so the only illumination was from the Godox AD200 LED head.
I then set the traffic cone over the peg at 90 degrees to be used as a subject. I adjusted the camera pan to 90 degrees, and without a filter or hood, I used live view to set up my focus.
Setup of the experiment - the garage lights were turned on for this shot with my cell.
From there I set the exposure: I used manual exposure with an ISO of 100. Since the camera was on a tripod shutter speed wouldn’t matter. I set the aperture to f/8. I figured this would be a stopped-down enough number for general use (f/8 and be there), but not so stopped down that you get diffraction, etc. The calculated shutter speed was 4 seconds for a “center the needle” exposure, so I used that. f/4, 4sec, and an ISO of 100. This was the setting for all of the photos.
For focus I tired to use CDAF through live view, and when that wouldn’t register, I manually focused using live view zoomed to 100% trying to focus on the top of the traffic cone.
The Experiment
Without a lens filter, I took a shot at 90, then moved the traffic cone to the 105 position, then adjusted the pan to the 105 position, adjusted the cone to the focus point in live view, refocused, then took another shot. I repeated this process for 120, 135, 150, 165, and 180 degrees.
I then reset the cone to the 90 degree position, and added the lens hood. I followed the same process for every angle. I reset the cone to 90 degrees, removed the lens hood and added the filter, and did the same process. I reset the cone to 90 degrees and finally added the hood to the filtered lens, and repeated the process for a final time—giving me 4 shots at each of the seven 15-degree increments: bare lens, hood only, filter only, and both hood and filter.
The Results
First, a caveat: No, this wasn’t laboratory perfect, the angles are *slightly* off and through human error the cone wasn’t in the exact same place, but the results in my opinion are “close enough” to establish a trend. Even in one photo where I missed the focus shouldn’t skew the results too much. Again, practical—not laboratory scientific. A typical amateur wouldn’t care if they were shooting at exactly 15 degrees (165 in this experiment) towards the sun, but the results I believe would be similar if they were off a degree or two.
Processing the photos
I processed all of these photos to the same level—the same white balance as-picked off of a grey point in the 1st picture (the nut on the car lift’s support beam—since it’s chrome steel.) All photos used Nikon's STANDARD picture control as-set in Capture NX-D (Latest Picture Contro
, and all photos sharpened via Unsharp mask in Capture NX-D, with an intensity of 50 and a radius of 5. This was set to all photos simultaneously, and all were processed as full-sized JPGs
90 degree results:
Bare Lens
UV Filter Only
Lens Hood Only
Lens hood and UV Filter
At this angle my conclusion is that the results vary only so slightly that it is relatively insignificant. Post-process can accommodate for any imperfections. The shots with the filter have a very slightly reddish-orangish tinge (whether inherent in the filter or caching a flare from my car lift or the cone I can’t tell). The two with the hood seem slightly darker and more contrast (subjective interpretation). This could be because the hood is blocking side-light, but the difference is miniscule to my eye.
At angles like 90 degrees or more to a light source I would say filter or no filter is of no significant factor: Using a hood might get you slightly more contrast but not using one won’t ruin the image.
105 degree results:
Bare Lens
UV Filter Only
Lens Hood Only
Lens Hood and UV Filter
My conclusions for 105 degrees (15 degrees from perpendicular towards the light source) would be the exact same as they were for 90 degrees. Slightly more of an orange-red tinge for the filter, slightly darker and more contrasty for the hood. Post-processing can take care of anything and at angles close to this there’s no significant factor.
120 degree results:
Bare Lens
UV Filter Only
Lens Hood Only
Lens Hood and UV Filter
My conclusions for 120 degrees (30 degrees from perpendicular towards the light source) would be the exact same as they were for 90 and 105 degrees. Slightly more of an orange-red tinge for the filter, slightly darker and more contrasty for the hood. Post-processing can take care of anything and at angles close to this there’s no significant factor.
135 degree results:
Bare Lens
UV Filter Only
Lens Hood Only
Lens Hood and UV Filter
My conclusions for 135 degrees (45 degrees from perpendicular towards the light source) are similar what they were for 90, 105, and 120 degrees, with one possible exception. Slightly more of an orange-red tinge for the filter, slightly darker and more contrasty for the hood. Post-processing can take care of anything and at angles close to this there’s no significant factor. It looks like there might be a slight indication of a blue flare around middle-right of the photo which is more pronounced in the hoodless-shots, but this might be an artifact of the LED light source. I think this is negligible but this might be the point where using a filter starts to flare.
150 degrees:
Bare Lens
UV Filter Only
Lens Hood Only
Lens Hood and UV Filter
My conclusions for 150 degrees (30 degrees away from the light source) start to sway towards the filter flaring blue. My eyes are catching a more pronounced blue flare in the photos with the with the filter than without—and like all the other shots the hood shots are slightly darker and contrasty. At this point I would definitely say, “ditch the filter if you want no flare”.
(To note, I did miss the focus in 2 of these shots, but I don't feel that the focus would have affected flare to a significant factor)
165 degrees:
Bare Lens
UV Filter Only
Lens Hood Only
Fens Hood and UV Filter
My conclusions for 165 degrees (15 degrees away from the light source) is that flare here is a major concern irrespective of filter vs. none—at this point you pretty much have to be shooting *for* some flare as it’s unavoidable with this lens. However, there are things to note:
While the hooded unfiltered shot seems to be the most towards the light source and thus more flare as based on the background, both it and the bare lens have a muted green flare slightly high right in the frame. Both filtered shots have this flare pronounced. In fact, both filtered shots have significant greenish veiling flare. There doesn’t seem to be much difference between hooded and not (especially with the slight angular deviance), but I would conclude for this—the filter definitely causes image degradation whether it’s hooded or not.
180 degrees:
Bare Lens
UV Filter Only
Lens Hood Only
Lens Hood and UV Filter
My conclusions for 180 degrees (Straight towards the light source) is that again, you see significant more flare with the filter hooded or not. This would be a special case where you’d be shooting into a light source, but the filters do cause a lot of extra flare. The most pronounced is that while you see a secondary reflection slightly below and to the left of the light source of the LED array in all of the shots, in the filter only you’re also seeing a tertiary reflection low and slightly left of center that is in none of the other shots—which would indicate that in that shot the filter is reflecting even more off of the sensor than in the other shots (this looks barely visible in the hooded and filtered shot, but barely. This could be an angle deviance or the hood is blocking some stray light—I can’t tell which). The hooded, unfiltered shot seems the darkest, but I would call this a special case—where you’re intentionally shooting into a bright light source. Still, if you are shooting like this, I would recommend no filter.
Conclusion
The 1st major conclusion is acknowledging that “your mileage my vary”. Different lenses and different filters likely will produce different results. Practically, this proved possibility, not being definite. I would assume a wider-angle lens might start to have flare problems at angles further away from the light source. I would assume zoom lenses might be slightly less contrasty when zoomed in since their hoods have to be designed to not vignette at wider angles. I would assume that newer filters with better coatings might flare less.
However, I would also assume that a cheap filter would flare as-much or more than my L37c. So, in terms of protection there is a trade-off: A good filter, even one which is years old might have a negligible impact on image quality up to 45 degrees towards a light source (on a 105mm lens, which on full-frame has an angle of view of approx. 19 degrees horizontal, 12.7 degrees vertical, and 22.8 diagonal). However, closer to the light source you risk more veiling flare and ghosting. So, do you risk lack of protection and lose any possible veiling flare, or do you keep the protection and try to fix (if possible) in post? That’s up to the photographer, proper.
My advice to any photographer doesn’t change though—if you have a filter on and the picture looks sub-par due to possibly veiling flare? Take the filter off and try again—that might solve the issue and can’t hurt—especially if you’re shooting at an angle close to a light source.
I would also recommend to any filter-user to experiment yourself with your gear and lenses how close to a light source you can shoot with the filter before you start to get flare. If you know what you and your gear can get away with before an important shot it can save headache and ruined shots before they happen.
Thank you for your time in reading this.
--
There's no shame in using auto or semiauto modes--no matter what the salesdroids at Best Buy tell you.
I figured this was the best place to put this because filters are indeed accessories.
There has been a lot of questions as to how much a UV filter would flare when put on a given lens, So I conducted an experiment to show how much flare a photo would have along different angles with respect to a singular light source with the lens bare, with a UV filter, with a lens hood, and with both a hood and a filter.
I say this experiment is “unscientific” because I could not eliminate every single variable in laboratory-like conditions. In some cases, the angle between photos is not exact. In some cases, the subject positioning is not exact. In some cases, the focus is not exact. This was done by a layperson (myself) in their garage. It’s not meant to be exact, but rather practical. Your average amateur photographer is not going to be concerned with exact angles where the difference between one shot and another is less than a degree. In my opinion the point of this experiment stands nonetheless.
The Setup
The setup was to mimic a singular directional light source which would be analogous to the sun. I conducted this experiment in my darkened garage where the only significant light source was the single directional light. Any small light leaks were negligible from a practical standpoint. I did this by setting up a Godox AD200 on a light stand at an end of my garage roughly along the seam in the concrete, with the light being pointed along the seam towards the other end of the garage. The AD200 was fitted with the LED head, which was set at full brightness. This provided the source of light which was very bright along a singular direction, but with the light spread out and reflecting off of the garage surfaces. I had the light stand at the lowest height setting. This put it in the frame, but at the high-end of the frame when the camera was pointed right at it.
Along the seam approximately 9 feet away, I positioned my camera (a Nikon D750) on my tripod (a Mefoto Globetrotter aluminum tripod), and leveled it with the bubble level on the tripod head. The tripod was set at its lowest height setting. For the experiment, the only adjustment made was with the pan function on the tripod head. I set the camera up so that 180 was in the direction of the light source and 90 was perpendicular to the light source—that way I could use the tripod head to dial in the angle towards the light source.
For a lens I used a Nikon Micro 105 f/2.8 VR lens, which was freshly cleaned before the experiment.. The main reason for this lens is that the lens’ filter thread diameter matched the UV filter I had—62mm. I also think this lens would be somewhat typical in terms of performance for consumer-grade lenses in terms of flare. This lens also has a lens hood that came with it, so the safe assumption could me made that the lens hood would be optimally sized for blocking any stray light that is outside of the lens’ angle of view from hitting the front element—since the lens is a prime lens and not a zoom lens, and thus the hood doesn’t have to take into account vignetting at it’s widest.
The filter I used was one I bought years ago when I shot film: A 62mm Nikon L37c UV Filter. This was also freshly cleaned before the experiment. In terms of performance, at the time, I believe it was considered mid-range—It wouldn’t break the bank but it did the job it was supposed to.
For a subject, I used the top of a road cone—which I had lying handy.
For the experiment I marked out approximately 8 feet from the camera 7 locations all at different angles with respect to the light source: 90 degrees (perpendicular) to the light source, and then at approximately 15 degree increments towards the light source: 105, 120, 135, 150, 165, and 180—with 180 pointed towards the light source. I marked out approximately 8 feet at each of those angles, and used pegboard pegs to mark out the angle and distance.
After that I snapped a cell shot of the setup (see below), turned on the Godox AD200 and turned the garage lights off so the only illumination was from the Godox AD200 LED head.
I then set the traffic cone over the peg at 90 degrees to be used as a subject. I adjusted the camera pan to 90 degrees, and without a filter or hood, I used live view to set up my focus.
Setup of the experiment - the garage lights were turned on for this shot with my cell.
From there I set the exposure: I used manual exposure with an ISO of 100. Since the camera was on a tripod shutter speed wouldn’t matter. I set the aperture to f/8. I figured this would be a stopped-down enough number for general use (f/8 and be there), but not so stopped down that you get diffraction, etc. The calculated shutter speed was 4 seconds for a “center the needle” exposure, so I used that. f/4, 4sec, and an ISO of 100. This was the setting for all of the photos.
For focus I tired to use CDAF through live view, and when that wouldn’t register, I manually focused using live view zoomed to 100% trying to focus on the top of the traffic cone.
The Experiment
Without a lens filter, I took a shot at 90, then moved the traffic cone to the 105 position, then adjusted the pan to the 105 position, adjusted the cone to the focus point in live view, refocused, then took another shot. I repeated this process for 120, 135, 150, 165, and 180 degrees.
I then reset the cone to the 90 degree position, and added the lens hood. I followed the same process for every angle. I reset the cone to 90 degrees, removed the lens hood and added the filter, and did the same process. I reset the cone to 90 degrees and finally added the hood to the filtered lens, and repeated the process for a final time—giving me 4 shots at each of the seven 15-degree increments: bare lens, hood only, filter only, and both hood and filter.
The Results
First, a caveat: No, this wasn’t laboratory perfect, the angles are *slightly* off and through human error the cone wasn’t in the exact same place, but the results in my opinion are “close enough” to establish a trend. Even in one photo where I missed the focus shouldn’t skew the results too much. Again, practical—not laboratory scientific. A typical amateur wouldn’t care if they were shooting at exactly 15 degrees (165 in this experiment) towards the sun, but the results I believe would be similar if they were off a degree or two.
Processing the photos
I processed all of these photos to the same level—the same white balance as-picked off of a grey point in the 1st picture (the nut on the car lift’s support beam—since it’s chrome steel.) All photos used Nikon's STANDARD picture control as-set in Capture NX-D (Latest Picture Contro
, and all photos sharpened via Unsharp mask in Capture NX-D, with an intensity of 50 and a radius of 5. This was set to all photos simultaneously, and all were processed as full-sized JPGs90 degree results:
Bare Lens
UV Filter Only
Lens Hood Only
Lens hood and UV Filter
At this angle my conclusion is that the results vary only so slightly that it is relatively insignificant. Post-process can accommodate for any imperfections. The shots with the filter have a very slightly reddish-orangish tinge (whether inherent in the filter or caching a flare from my car lift or the cone I can’t tell). The two with the hood seem slightly darker and more contrast (subjective interpretation). This could be because the hood is blocking side-light, but the difference is miniscule to my eye.
At angles like 90 degrees or more to a light source I would say filter or no filter is of no significant factor: Using a hood might get you slightly more contrast but not using one won’t ruin the image.
105 degree results:
Bare Lens
UV Filter Only
Lens Hood Only
Lens Hood and UV Filter
My conclusions for 105 degrees (15 degrees from perpendicular towards the light source) would be the exact same as they were for 90 degrees. Slightly more of an orange-red tinge for the filter, slightly darker and more contrasty for the hood. Post-processing can take care of anything and at angles close to this there’s no significant factor.
120 degree results:
Bare Lens
UV Filter Only
Lens Hood Only
Lens Hood and UV Filter
My conclusions for 120 degrees (30 degrees from perpendicular towards the light source) would be the exact same as they were for 90 and 105 degrees. Slightly more of an orange-red tinge for the filter, slightly darker and more contrasty for the hood. Post-processing can take care of anything and at angles close to this there’s no significant factor.
135 degree results:
Bare Lens
UV Filter Only
Lens Hood Only
Lens Hood and UV Filter
My conclusions for 135 degrees (45 degrees from perpendicular towards the light source) are similar what they were for 90, 105, and 120 degrees, with one possible exception. Slightly more of an orange-red tinge for the filter, slightly darker and more contrasty for the hood. Post-processing can take care of anything and at angles close to this there’s no significant factor. It looks like there might be a slight indication of a blue flare around middle-right of the photo which is more pronounced in the hoodless-shots, but this might be an artifact of the LED light source. I think this is negligible but this might be the point where using a filter starts to flare.
150 degrees:
Bare Lens
UV Filter Only
Lens Hood Only
Lens Hood and UV Filter
My conclusions for 150 degrees (30 degrees away from the light source) start to sway towards the filter flaring blue. My eyes are catching a more pronounced blue flare in the photos with the with the filter than without—and like all the other shots the hood shots are slightly darker and contrasty. At this point I would definitely say, “ditch the filter if you want no flare”.
(To note, I did miss the focus in 2 of these shots, but I don't feel that the focus would have affected flare to a significant factor)
165 degrees:
Bare Lens
UV Filter Only
Lens Hood Only
Fens Hood and UV Filter
My conclusions for 165 degrees (15 degrees away from the light source) is that flare here is a major concern irrespective of filter vs. none—at this point you pretty much have to be shooting *for* some flare as it’s unavoidable with this lens. However, there are things to note:
While the hooded unfiltered shot seems to be the most towards the light source and thus more flare as based on the background, both it and the bare lens have a muted green flare slightly high right in the frame. Both filtered shots have this flare pronounced. In fact, both filtered shots have significant greenish veiling flare. There doesn’t seem to be much difference between hooded and not (especially with the slight angular deviance), but I would conclude for this—the filter definitely causes image degradation whether it’s hooded or not.
180 degrees:
Bare Lens
UV Filter Only
Lens Hood Only
Lens Hood and UV Filter
My conclusions for 180 degrees (Straight towards the light source) is that again, you see significant more flare with the filter hooded or not. This would be a special case where you’d be shooting into a light source, but the filters do cause a lot of extra flare. The most pronounced is that while you see a secondary reflection slightly below and to the left of the light source of the LED array in all of the shots, in the filter only you’re also seeing a tertiary reflection low and slightly left of center that is in none of the other shots—which would indicate that in that shot the filter is reflecting even more off of the sensor than in the other shots (this looks barely visible in the hooded and filtered shot, but barely. This could be an angle deviance or the hood is blocking some stray light—I can’t tell which). The hooded, unfiltered shot seems the darkest, but I would call this a special case—where you’re intentionally shooting into a bright light source. Still, if you are shooting like this, I would recommend no filter.
Conclusion
The 1st major conclusion is acknowledging that “your mileage my vary”. Different lenses and different filters likely will produce different results. Practically, this proved possibility, not being definite. I would assume a wider-angle lens might start to have flare problems at angles further away from the light source. I would assume zoom lenses might be slightly less contrasty when zoomed in since their hoods have to be designed to not vignette at wider angles. I would assume that newer filters with better coatings might flare less.
However, I would also assume that a cheap filter would flare as-much or more than my L37c. So, in terms of protection there is a trade-off: A good filter, even one which is years old might have a negligible impact on image quality up to 45 degrees towards a light source (on a 105mm lens, which on full-frame has an angle of view of approx. 19 degrees horizontal, 12.7 degrees vertical, and 22.8 diagonal). However, closer to the light source you risk more veiling flare and ghosting. So, do you risk lack of protection and lose any possible veiling flare, or do you keep the protection and try to fix (if possible) in post? That’s up to the photographer, proper.
My advice to any photographer doesn’t change though—if you have a filter on and the picture looks sub-par due to possibly veiling flare? Take the filter off and try again—that might solve the issue and can’t hurt—especially if you’re shooting at an angle close to a light source.
I would also recommend to any filter-user to experiment yourself with your gear and lenses how close to a light source you can shoot with the filter before you start to get flare. If you know what you and your gear can get away with before an important shot it can save headache and ruined shots before they happen.
Thank you for your time in reading this.
--
There's no shame in using auto or semiauto modes--no matter what the salesdroids at Best Buy tell you.
