Jump to content
UltravioletPhotography

IR photo without IR filter


Recommended Posts

Most ND filters are quite transparent above 720nm, so I played with possibility to use stronger ND filter to create pictures with IR look but also with some normal colors preserved.

 

This one was shot with Camdiox ND1000 filter:

 

http://www.fpl.cz/uvp/nonir.jpg

 

It needs scene where leafs are in direct sun light - - in shadow (even when there is lot of light) it delivers image close to result from unfiltered full spectrum camera.

And is important to use as apochromatic lens as possible. Esp. wide lenses will produce heavy aberrations:

 

http://www.fpl.cz/uvp/nonir2.jpg

 

 

Some ND filters: http://www.fpl.cz/tst/ndir.jpg

Would be nice to have their real transmission curves ...

Link to comment

An interesting idea. I have some ND filters which have not been used for a very long time. Now I am enthused to find them and try this myself.

 

That is quite some amount of color aberrations in the crop! It almost becomes an artistic abstraction.

 

Readers, be sure to click Petr's link at the bottom to see the results from a wide range of ND filters.

Link to comment

I don't know the transmission of other ND filters, but the Schott NG filters don't seem to have much difference between visual and IR, so be aware of that.

"NG" filters are the Schott neutral density filters.

post-87-0-91800600-1524681444.jpg

Link to comment

I don't know the transmission of other ND filters, but the Schott NG filters don't seem to have much difference between visual and IR, so be aware of that.

 

Key is _difference_ between transmittance in visual and IR. You are using in your graph scale, on which it isn't so visible. When you display lower portion of transmittance graph for 2.2mm thick NG3 (what probably is base glass on which were build at least BW and VF ND1000 filters from above mentioned test), with linear scale on Y axis, then you can see quite significant difference:

 

http://www.fpl.cz/uvp/ng3trans.jpg

Link to comment

You might try playing around with crossed polarizers, as well--they don't attenuate much beyond 760 nm.

I don't believe that Vario ND filters based on crossed polarizers can produce enough difference in VIS/IR suppression, but I will try it!

Link to comment

Yes, ND filters are a mix of Visual and IR, and transmit a higher ratio of IR compared to to visual, and thus are a nice mix of visual and IR compared to a straight full spectrum, or a straight visual + IR wide open mix.

However, I would not say that ND filters "are quite transparent above 720nm", in fact they will attenuate IR transmission, making the exposure longer.

Above I was using 2.2mm thick as a typical filter thickness, but below I am using the 1mm thick example (as used in the Schott catalog) because it better shows the transmission of all their NG filters.

The same ratio seen in the normalized graph can also be seen in either the diabatic of Linear graphs, however diabatic and linear graph versions also show the transmission attenuation applied to the exposure.

Keeping in mind that this graph shows 1mm thick transmission, so 2mm thick will make exposure time even longer. When using ND filter for visual + IR mixing, you should experience some extended exposure time.

 

post-87-0-00561700-1524706458.jpg

Link to comment

2.2mm thick version.

Not all ND filters may be created equal, so I can't compare these graphs with your ND filter.

The transmission of these Schott NG (ND) filters at even 1mm thick tend to attenuate the IR transmission quite a bit compared to a usual longpass filter.

Notice how the normalized NG1 (black line) doesn't get to "1.00" until 1200nm, yet it isn't even up to 1E-03 at 1200nm.

Also notice how NG11 and NG5 have slightly less transmission in IR than they do in visual (except between 690nm ~ 740nm).

NG5 and NG11 would work for IR + Visual, but would actually reduce the IR more than the visual. One might use GG420 instead for a more flat Visual + IR curve mix.

Notice how NG3 and NG4 (yellow and purple) would be the only two that would work practically for attenuating visual mixed with IR (depending on usual thicknesses),

maybe ND9 (green) if it were very thin, but once again the overall attenuation is very strong, so it would require a long exposure time.

post-87-0-09152500-1524725195.jpg

Link to comment
Andy Perrin
I guess I don't understand what the benefit of doing this is, given that IR filters are pretty cheap overall?
Link to comment

If you already have them - which many of us do - then might as well use them for IR. I never thought of it myself. My NDs are all in some storage bin somewhere currently.

 

I had those big rectangular (Lee or Cokin) filters from early digital days. When the camera dynamic range was smaller, it was useful to have a graduated ND filter for bright skies or sunsets.

Link to comment

I guess I don't understand what the benefit of doing this is, given that IR filters are pretty cheap overall?

It's attempt to have visible IR effect but still at least some level of natural colors. It can't be achieved with any usuall IR filter.

Link to comment

Nice day out there so went out back and compared some IR filters and ND filters.

Settings: F9, ISO 200, AP, Matrix, 0EV, Exposure times noted below. White balanced full frame marquee in NX2 (CNX2).

The RG filters used here are 2mm thick, and the NG filters used here are 2.2mm thick.

 

RG1000 1/20s

post-87-0-29409700-1524789181.jpg

 

RG850 1/100s

post-87-0-67369000-1524789349.jpg

 

RG9 1/400s

post-87-0-75921700-1524789389.jpg

 

RG695 1/400s

post-87-0-25522000-1524789438.jpg

 

NG3 1/25s

post-87-0-04451500-1524789506.jpg

 

NG4 1/25s

post-87-0-86239100-1524789544.jpg

 

Since the two NG (ND) filters looked dark, I added exposure compensation of 1 to them in post (NX2).

ND3 1/25s +EC 1 post

post-87-0-95269500-1524789717.jpg

 

ND4 1/25s +EC 1 post

post-87-0-10061900-1524789774.jpg

 

Full Spectrum (no filter) 1/600s

post-87-0-10004000-1524789596.jpg

Link to comment
I don't think the ND filters are terribly selective for IR in the last examples--we see IR effects just because the sensor is already very sensitive to IR, even with no filter. As I said before, experimenting with crossed polarizers will give more (and more controllable) selectivity.
Link to comment

I tend to think the ND3 example is somewhat pleasing and different, given no swap needed for blue sky, and red barn is a definite mix of visual.

However, it does requite a longer exposure, especially if you don't want the brighter version out of the box.

It works slightly like a UG1 or UG2A dual band IR shot, but not as strong of blue, and the addition of red in this case also from the visual mix,

something you don't get with UG1/UG2A shots.

Link to comment

I have done few more test pictures with different OG/RG filters and with ND1000:

 

http://www.fpl.cz/uvp/ogrgnd1.jpg

 

http://www.fpl.cz/uvp/ogrgnd2.jpg

 

http://www.fpl.cz/uvp/ogrgnd3.jpg

Link to comment
Andrea B.

These are cool comparison thumbnails!! Thank you for sharing them with us.

 

Petr, if you have not already tried them, you might also enjoy the interesting false colors produced by the dual band-pass filters when used unblocked. For example, the BG3 (Blue + IR) or the U330 (UV+IR). The latter records mostly IR if used unblocked, but passes just enough UV/B/G to make things interesting in false color.

Link to comment

From those type of filters I like Hoya B410 (and also Schott UG5 give some interesting results).

Only limiting factor is, that there are not available wide lenses apochromatic enough in such range. :-(

Link to comment

My compliments of the very nice comparisons! Excellent!

By the way, someone was telling me the other day that the new LifePixel Hyper Color filter is a 470nm/475nm longpass, like a GG475 filter.

I don't know other than what they told me, but the shots do look as such.

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...