Jump to content
UltravioletPhotography

Color shifts and other issues


Stefano

Recommended Posts

I have some problems with the colors in my UV photos. I always WB in camera with a paper tissue. If I take a lot of photos in "multishot" mode (you keep the button pressed and the camera keeps shooting), the first one usually has less vibrant colors than the others. In videos I usually get stronger colors than in photos. Here https://www.ultravio...-nectar-guides/ You can see an example.

 

For every image, the filter was ZWB2 (2 mm) + chinese BG39 (2 mm). "vivid" color in camera settings.

 

F-stop: f/2.8, ISO 400, 1/8 s exposure.

post-284-0-55053300-1581517350.jpg

 

Same settings.

post-284-0-42055900-1581517439.jpg

 

Same settings. Taken the day before. The first photo from this row was almost B&W.

post-284-0-31101600-1581517549.jpg

 

F-stop: f/2.8, ISO 1600, 1/8 s exposure.

post-284-0-60347600-1581517711.jpg

 

Same settings.

post-284-0-00023900-1581517724.jpgù

 

F-stop: f/2.8, ISO 400, 1/8 s exposure.

post-284-0-18384900-1581519143.jpg

 

Same settings.

post-284-0-08595100-1581519213.jpg

 

Sometimes, the WB changes a bit, suddently (I increased the ISO, so the sky probably "blew out" and under a typical UV white balance white becomes yellow), but even dark objects became more yellow. In these images you can also see another problem: I have fringing at the sides. Maybe my UV filter is too thick (4 mm).

 

F-stop: f/2.8, ISO 400, 1/13 s exposure.

post-284-0-61909300-1581517889.jpg

 

F-stop: f/2.8, ISO 800, 1/8 s exposure.

post-284-0-90408300-1581517834.jpg

 

And now, these two images show that, sometimes, I can get stronger colors.

 

F-stop: f/2.8, ISO 800, 1/8 s exposure.

post-284-0-47834400-1581519350.jpg

 

Same settings.

post-284-0-21118600-1581519367.jpg

Link to comment

The dandelion picture is an indication there is a slight NIR leak. Or your lens doesn't go very deep into the UV(A).

 

That being said, w/b-ing UV photos can easily result in minor variations as the false colours are as a whole not tightly defined. This in particular is evident when using lenses that "give up the ghost" in the 370-380 nm region.

Link to comment

The dandelion picture is an indication there is a slight NIR leak. Or your lens doesn't go very deep into the UV(A).

 

That being said, w/b-ing UV photos can easily result in minor variations as the false colours are as a whole not tightly defined. This in particular is evident when using lenses that "give up the ghost" in the 370-380 nm region.

 

That is true, given the dark black/gray of the background grass, the warm brown of the center should be black, thus indicating a little Red/IR/700nm range leak.

Link to comment
I like the pics! Are you setting the white balance in camera? If not, white balancing JPEGs is sometimes problematic. But it could also be an IR leak.
Link to comment
I have a far red LED. It should peak at 736 nm, or near that. I will try to force a leak with it (it isn’t a fair way to test filters, but it can give an idea).
Link to comment

So still a mystery why my first photos in “multishot” mode are not as good as the others? Can you see the (slight) difference between the first one and the second one in my examples? Are you experiencing the same issues with your cameras? (I don’t think so).

I usually keep the button pressed to make more photos and have better colors in the others. This is how I took my flower photos above. They come from rows of ~3-6 photos.

Link to comment
Hm, I haven't tested what multishot does. It may be a question of auto noise reduction in the camera or something messing things up?
Link to comment
Are you using a fixed WB setting, or are you resetting it for every shot? The latter strategy could certainly cause erratic results. If you are using a fixed setting and not varying lens and filter, the issue must be changes in the lighting itself, which may or may not be under your control.
Link to comment

For this photos I used the same WB. In the third one, even if it was taken the day before (with maybe a slightly different WB) the color changed “by itself”. I have other examples of changes in color that I cannot control.

 

Images #8 and #9, for example, were taken seconds apart. The color changed because I changed the ISO.

Link to comment

I tried to push a leak with an incandescent source. I used a 5 W/21 W bulb, running the thin filament (5 W) at 32 V. Yes, it doesn't melt at 32 V, but at 39 V.

Both images f/2.8, ISO 1600, 1/8 s exposure.

 

ZWB2 (2 mm) + BG39 (2 mm).

post-284-0-42648300-1581548021.jpg

 

ZWB2 (2 mm) + BG39 (2 mm) + pen ink filter.

post-284-0-07013800-1581548082.jpg

 

Yes, you can push a leak. You can tell it's IR because far red light appears yellow under a UV white balance. But can this leak contaminate my images?

Link to comment

Test for Red/IR/700nm leak:

Stack your UV stack, shoot pic that has optimal settings/exposure.

Stack the same with a longpass filter, let's say 590nm or above, but below where you think the leak might start, mid 600nm+(?).

Shoot another shot with the longpass stack. Use the exact same settings and exposure time as the first optimized shot.

The second shot will isolate all IR that is present and mixed into your first UV "only" (leaky) shot.

Show both, compare the UV and leak shot, leak shot should optimally be black, if it is not black then whatever faint image you have is the amount of the leak.

 

Again:

Shoot the dandelion with your UV stack.

Shoot it again using the same settings/exposure, except with a longpass filter added.

The second shot should ideally be black if there is no leak.

The second image is the leak.

 

"Pushing" the leak test shot makes no sense, it only makes the leak look like more than it is, keep it the same settings and exposure time as the UV shot.

Link to comment

We have no idea what cut off your pen ink filter has. So not the best to run a test with. Get a cheap gel or a free Lee sample pack. You just have to pay for shipping for the Free Lee sample pack and it will provide you with many good filters that will cover you lens.

Use that to make controlled tests.

Link to comment
I don’t have other longpass filters other than the pen ink filter and other accidental infrared filters (remote control cover, etc.) My ink filter should cut at ~700 nm, with a somewhat sharp cut (see the spectrums here: https://www.google.c...ClT_jZhUluHs46A, page 1). I will investigate the leak, but I don’t think that my filter has less than OD 4.5 in the red/infrared. Yes, it’s Chinese glass, but it is still a 2 mm + 2 mm stack, which gives more than OD 5 with high quality glass.
Link to comment

It has some IR leaking, that is why the center of the dandelion is warm, it should be black. You can't WB it warm if it isn't leaking, if you did then everything would look warm. So OD4.5 would not have a warm leak like that.

The only thing is, I see nothing else in any of the other photos that suggests any leak, even in the dandelion photo the grass is colorless in the background.

So who knows... but the center of the dandelion looks warm, should look black.

 

You could PM the RAW file to me, and I could WB and see what I think... if you have a RAW?

 

As far as a longpass filter, seems like any red camera filter should do the trick, #25 or #29 which will pass red and IR but not UV.

Link to comment

I don’t have a RAW file or a non-WB version. I WB in-camera, and the original image is exactly what I posted, but in size 4000*3000. I have a red piece of plastic, I don’t think it is a reliable UV-block filter, but I haven’t tested it

 

Edit: I tried shining a 365 nm UV LED torch through it and apparently it is at least 50% transparent in UV. Unusual if you think that this "filter" appears red to the naked eye, this is a proof that things in UV are not always obvious, sometimes you have unexpected results.

Link to comment

My first chinese BG39 broke into five pieces five minutes after I started playing with it because it fell from ~40 cm on my tiles. I always kept those pieces, and they are very useful. Today I took some photos with an additional 2 mm of BG39. I have more than 2 images, but for now I will post just one pair.

Both images f/2.8, ISO 1600, 1/8 s exposure.

 

ZWB2 (2 mm) + chinese BG39 (2 mm).

post-284-0-89829800-1581780799.jpg

 

ZWB2 (2 mm) + chinese BG39 (2 mm) + chinese BG39 (2 mm).

post-284-0-18995200-1581780883.jpg

 

Apart from reduced sensitivity, do you notice a better IR blocking?

Link to comment
Does the fact that I removed the IR-cut filter from the sensor without replacing it with anything play a role here? It seems that my lens can not focus on the sides (and corners). Keep in mind that I have a 4 mm thick filter on a small lens. The color variation could also be caused by the thickness of my filter. I can experiment with visible and infrared light (without any filter) just to see how my lens behaves.
Link to comment

I don’t have a RAW file or a non-WB version. I WB in-camera, and the original image is exactly what I posted, but in size 4000*3000. I have a red piece of plastic, I don’t think it is a reliable UV-block filter, but I haven’t tested it

 

Edit: I tried shining a 365 nm UV LED torch through it and apparently it is at least 50% transparent in UV. Unusual if you think that this "filter" appears red to the naked eye, this is a proof that things in UV are not always obvious, sometimes you have unexpected results.

I don’t have a RAW file or a non-WB version. I WB in-camera, and the original image is exactly what I posted, but in size 4000*3000. I have a red piece of plastic, I don’t think it is a reliable UV-block filter, but I haven’t tested it

 

Edit: I tried shining a 365 nm UV LED torch through it and apparently it is at least 50% transparent in UV. Unusual if you think that this "filter" appears red to the naked eye, this is a proof that things in UV are not always obvious, sometimes you have unexpected results.

 

Probably your red plastic filter works basically the same as a red longpass filter, blocking anything below the red cutoff.

However, as I pointed out before, LED's are not a good way to test filters, lenses, etc.. LED's are a good way to illuminate targets, subjects...

but you will get misleading results when testing filters and lenses with them.

So shining an LED though the red plastic filter is not the same as stacking the U filter + BG filter + Red Plastic filter. Try that, it might be more interesting than the double BG stack.

It doesn't matter if the LED can be seen through the red plastic filter, like I showed before, UV LED's can be seen through 850nm filters,

this is because the OD isn't strong enough to block the intensity of the UV.

Stacking that red plastic with your UV-"only" stack may show you a faint image of the dandelion center, and nothing else.

Use the same settings and exposure time as used for the UV-only shot of the dandelion.

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...