1. Reminder: Please user our affiliate links to get to your favorite stores for holiday shopping!

Native m4/3 Lens Infrared performance / IR hotspots

Discussion in 'Native Lenses' started by Sanpaku, Aug 26, 2012.

  1. Sanpaku

    Sanpaku Mu-43 Regular

    79
    Jul 24, 2012
    Lenses optimized for visible light often exhibit low-contrast and central hotspots at smaller apertures in the infrared. The anti-reflective glass coatings and matte black finishes that work well in normal photography may be reflective at longer wavelenghs.

    I've attempted to collect all extant information on native m4/3 lens infrared performance from a prior list by camera conversion service Kolari Vision, a comprehensive internet search, and tests with my own lenses. While no lens offers perfect infrared performance, better performing lenses offer decent contrast and the near absence of central hotspots at larger apertures. My criterion in this list (where information is available), is good performance at f/8 and below.

    Good Performers

    Olympus 12mm f/2
    Olympus 60mm f/2.8 Macro (hotspot @ f/9)
    Olympus 9-18mm f/4-5.6
    Olympus 14-42mm f/3.5-5.6 ED
    Olympus 14-42mm f/3.5-5.6 II R
    Olympus 40-150mm f/4-5.6
    Panasonic 8 mm f/3.5 (hotspot @ f/10)
    Panasonic 14 mm f/2.5
    Panasonic Leica 45mm f/2.8 Macro
    Panasonic 7-14mm f/4 (hotspot @ f/9)
    Panasonic 14-42mm f/3.5-5.6
    Panasonic 14-45mm f/3.5-5.6
    Panasonic X 45-175mm f/4-5.6

    Poor Performers

    Olympus 12-50mm f/3.5-6.3
    Olympus 17mm f/2.8
    Olympus 45mm f/1.8
    Olympus 75mm f/1.8 (low contrast & hotspot @ f/4.5)
    Panasonic 14-140mm f/4-5.8 (hotspots from 14-25mm)
    Panasonic 45-200mm f/4-5.6 (low contrast)
    Panasonic Leica 25mm f/1.4 (hotspot @ f/5)
    SLR Magic Hyper Prime 12mm f/1.6
    Voigtlander 25mm f/0.95

    Mixed reviews (please let me know if you have these)

    Olympus 45mm f/1.8 (one report of hotspot @ f/9)
    Panasonic 12-35mm f/2.8

    For my own tests, I used a full-spectrum GX1 with 850nm (Wratten 87C, B+W 093, Schott RG850) and 790nm (Wratten 87, Lee 87) long-pass filters, taking multiple exposures at all apertures of blue sky (> 90° from the sun's direction) framed by light colored clouds or leaves. Loss of contrast and hotspots show very strongly in these conditions at wavelenghs above 800nm.

    *** currently editing a mosaic showing how hotspots vary with aperture for inclusion here ***

    If anyone has used other native m4/3 lenses with infrared filters, or on infrared/full spectrum converted cameras, I'd like to add your report in order to make this listing as comprehensive as possible.
     
    • Like Like x 7
  2. RevBob

    RevBob Super Moderator

    Jun 4, 2011
    NorthWestern PA
    Bob
    Anyone have some insight into why most of the higher-end lenses produce hot spots?
     
  3. kwalsh

    kwalsh Mu-43 Top Veteran

    775
    Mar 3, 2012
    Baltimore, MD
    I believe it is the lens coatings. Some of the higher performing coatings in the visible may be more reflective in the IR. It is also possible that some of the fancier glasses have similar issues. Please be aware I am mostly speculating here based on what I've seen of coatings in optics other than consumer photographic optics.

    Ken
     
  4. kwalsh

    kwalsh Mu-43 Top Veteran

    775
    Mar 3, 2012
    Baltimore, MD
    I have not found the 45-200 to be a particularly good performer on an IR converted G1. It might not have an obvious hot spot but it is very low contrast as if the entire frame is covered by a giant hot spot.

    The other issue here is that hot spots are only part of the problem. The optics are often poorly corrected in the IR and so many lenses are quite soft in the IR while others are very good. I'd say the 14/2.5 is one of the best performers for contrast and shaprness in the IR.

    Finally, consider many optics will look worse used on an IR converted camera than a unconverted but R72 filtered camera. This is because in any cases R72 on unconverted is really mostly letting in a narrow gap of wavelengths between the IR blocking cutoff and the R72 cutoff. These wavelengths are close to visible so you get good optical performance but long exposures. On a converted camera the exposure is dominated by longer uncorrected wavelengths and so softer images and more hot spots.

    Ken
     
    • Like Like x 2
  5. twokatmew

    twokatmew Mu-43 All-Pro

    Jun 1, 2012
    Lansing, MI, US
    Margaret
    Very interesting, kwalsh!

    I tried IR for the first time just the other day using a Hoya R72 filter and the Oly 14-42 IIR lens, no hot spots whatsoever. Shortly before I saw this thread, I was just thinking I'd have to try out the 17/2.8 and the 45/1.8. I also own the 40-150R, and I plan to try all three lenses soon. I'll post back with samples.

    In the meantime, here's one from this past Friday. E-M5, 14-42IIR @ 14mm, f/7.1, 12 seconds, though I'm thinking I'll stick pretty close to f/5.6 in the future. :smile:

     
    • Like Like x 2
  6. RevBob

    RevBob Super Moderator

    Jun 4, 2011
    NorthWestern PA
    Bob
    Cool photo - didn't realize that you could do this without converting the camera.
     
    • Like Like x 1
  7. Sanpaku

    Sanpaku Mu-43 Regular

    79
    Jul 24, 2012
    Most camera internal hot mirrors have optical transmission spectra that look a bit like Schott BG039 glass, which admits more than 50% of visible light between 320nm and 610nm, while cutting red and very near IR significantly. Camera conversion services Kolari Vision and MaxMax both sell color correction filters to allow visible light white balance on full spectrum converted cameras, apparently using the BG039 glass:

    2dl7vbo.

    As an aside, note that BG039 isn't that great a match, but its the closest available inexpensive filter glass. The proprietary camera hot filters are mostly like the similar but more permissive Schott BG040 stacked with a light yellow UV blocking filter like Schott GG420.

    Meanwhile an IR pass filter like the Wratten #89B (Hoya R72 filter, B+W 092 filter, or Schott RG715 glass are near equivalents) has a transmission spectra that admits more than half of near-IR radiation with a wavelength greater than 720nm. When the camera hot mirror and #89B deep red filter are stacked what you get in effect is a rather poor bandpass filter for wavelengths greater than 680nm, peaking at about 4% transmission at 720 nm.

    2h36dz6.
    My graph in logarithmic scale, assuming camera hot mirror is equivalent to stacked BG040 (1mm) and GG420 (3mm). Wratten #89B IR filter is the RG715 (3mm).

    The native CMOS response and Bayer filter dye tranmission spectra determine the camera sensor color response to the passed near infrared. In the 680-775nm range passed by Wratten #89B filters on unconverted cameras, green pixels have about a third of the response of red ones, and blue almost none, hence the very red cast. Past 850nm (Wratten #87C, B+W 093 filter) the spectral response of all three sensor pixels colors is similar, and hence most images are rendered in monochrome.

    That 720nm peak is at a longer wavelength than the visible light absorptions of photosynthetic pigments so plants still have that ghostly, etherial look ("Wood's effect") characteristic of IR photos.

    psnpigmentspec.

    The main advantage of doing an IR or full-spectrum conversion for IR digital photography is that by eliminating the camera internal hot mirror, shutter times are about the same as for visible light photography (vs the 5-6 stop loss for an R72 on unconverted camera), permitting handheld, windy day, and flash IR photography.
     
    • Like Like x 2
  8. RevBob

    RevBob Super Moderator

    Jun 4, 2011
    NorthWestern PA
    Bob
    Excellent explanation, thanks. :smile:
     
  9. Sanpaku

    Sanpaku Mu-43 Regular

    79
    Jul 24, 2012
    Bumping to note that I found some other online reports (reported good performance by the 12-35mm is especially encouraging to me), and updated my response to RevBob with a graph from actual filter data.
     
    • Like Like x 2
  10. tornado

    tornado Mu-43 Veteran

    270
    Nov 6, 2012
    Thanks for this thread. I recently started exploring IR work with an IR converted Canon 40D. Any comment on elimination of the anti-alias filter from the converted sensor? This didn't come up in my research prior to having the conversion done. Not sure it a big deal...?
     
  11. tornado

    tornado Mu-43 Veteran

    270
    Nov 6, 2012
    Found this info on anti-alias filters on an IR conversion site I had not know about IR Techniques


     
    • Like Like x 1
  12. laptoprob

    laptoprob Mu-43 Regular

    38
    Jan 28, 2010
    Just found out the otherwise great Hyperprime 12/1,6 creates a hotspot with my IR converted GF1. And it doesn't seem to focus well.
    The Pana 14/2,5 is stellar though...
     
  13. kwalsh

    kwalsh Mu-43 Top Veteran

    775
    Mar 3, 2012
    Baltimore, MD
    Just adding a quick note on the 12-32, cross posting from DPR:

    Did a very quick check (viewfinder/LCD check only) of the 12-32 on a 720nm modified G1. There were no obvious hot spots to be seen at both wide and small apertures and at 12mm and 32mm. Please note that this was a very quick check for obvious hot spotting. It is quite possible there is a more subtle hot spot that might show up in post processing with increased contrast and a more uniform subject. This means the 12-32 is certainly usable for IR but I'm not sure if it is truly clear of a hot spot or loss of contrast. So it is at least "acceptable" I'd say for IR and with more testing might prove to be great.
     
    • Like Like x 1
  14. SteveJB

    SteveJB Mu-43 Rookie

    15
    Nov 5, 2012
    Baltimore
    Steve
    Good thread, all. There's some really great info here. I just picked up an E-PL1 from the 'bay that was converted to 750nm (not sure where the conversion was done). I'm finding that images with my 14-42 II R are a fairly soft and lacking fine detail. Careful processing can bring some of it back, but images are still not what I would call sharp. They seem soft and a much more noisy than I would expect for ISO 200.

    Given that this is reported to be a good lens for IR work, does anyone know what the problem could be - the lens? Poor quality filter used in the conversion? Certainly more testing is needed, but these initial findings are a little disappointing.
     
  15. leuallen

    leuallen Mu-43 Regular

    27
    Mar 26, 2010
    Mackinaw, Il
    I have been shooting IR frequently the last month. You have the 12mm 2.0 listed as a good performer. I have found the opposite. At f8 I get a consistent hot spot in the center. Fixable in SilverEffects Pro using control points - all my IR are converted to BW using SE2. The 45-175 worked fine. I don't recall any problems with the 14-45.

    This is a pain as I keep three camera bags with cameras (EM1, GX7, and G1 720mm infrared converted) with a useful range of lenses for each. This way I don't have to swap lenses for each kit. Just pick kit and go. If I swap lenses I often end up in the field missing a lens that I wanted. I guess I'll have to revert to swapping the 12-35 to get the 12mm focal length. I hope that it does not have this problem.

    The 50-200 4/3 lens works well but is manual focus on the G1.

    Larry
     
    • Like Like x 1
  16. tornado

    tornado Mu-43 Veteran

    270
    Nov 6, 2012
    Anyone know if the Olympus body cap pancakes gave hotspots?
    I will be getting my EM5 back from LifePixel after Super Color filter conversion and I want to test it out in Japan in a few weeks during vacation. Might puck up a pancake if it looks promising...

    Sent from my BlackBerry z10 via mu43 App.
    My BlackBerry Messenger pin:24BDF528
     
  17. Petrochemist

    Petrochemist Mu-43 Top Veteran

    651
    Mar 21, 2013
    N Essex, UK
    Mike
    Strangely I've found my 14-42 to give severe hot-spots with an unconverted camera, but to be much better on a full spectrum conversion. The 45-200 has also been fairly good on the conversion, although both show weak hot-spots & reduced contrast the contrast is easy to add back in.

    Here's a typical shot with the 14-42 from Saturday (Full spectrum camera + 650nm filter):
    13498899544_15da0fc932_z.
    Andrew Jordan - Infra red panning by Analyst 1, on Flickr

    and the 45-200 (full spectrum camera +720nm filter)
    13509093595_c2102742b5_z.
    IR BTCC - Neal by Analyst 1, on Flickr

    I've never had any issues with the Oly 17mm f/2.8 (mainly used with a BG3 filter giving blue & IR)
    Here's an example of that (Full spectrum camera + BG3 filter):
    13392716463_33f7149dd5_z.
    Dovercourt park IR by Analyst 1, on Flickr


    I know there have been multiple versions of the 14-42 Could it be the version makes a big difference in IR performance?
     
  18. Petrochemist

    Petrochemist Mu-43 Top Veteran

    651
    Mar 21, 2013
    N Essex, UK
    Mike
    Perhaps I should mention that for those interested in IR photography there is a new dedicated forum (not related to this one) that can be found at http://infrared-photography.freeforums.net/
    The old IR forum died from SPAM and inactive moderation.
     
    • Like Like x 3
  19. tornado

    tornado Mu-43 Veteran

    270
    Nov 6, 2012
    Thx...signed up.
     
  20. tornado

    tornado Mu-43 Veteran

    270
    Nov 6, 2012
    I've recently read that hotspots can be very dependant on the specific scene...if sun is in frame or otherwise causing flare. Using a good lens hood or otherwise shading the front element can improve the issue significantly.
    Also, hotspots can be specific to the aperture since the blade coatings may be the cause of IR reflections within the lens body.
    So, even if a lens is listed as a poor performer, it might be in need of more rigorous testing to determine if a variety of shooting conditions give hotspots. Of course, the opposite may also be true...a lens listed as a good performer may give poor results under certain conditions.
    Finally, some have reported that shooting with a front mounted IR pass filter gives different hotspot results than using a camera with an IR converted sensor. Since the lens hotspot lists don't typically indicate the camera setup used for testing, we're left to speculate.
    Upshot is: if you have fav lens you want to use for IR, you'll need to make some tests and see what's working for your setup & shooting conditions.
     
    • Like Like x 3