Question about infrared conversion and expected autofocus with lenses

Discussion in 'Open Discussion' started by Amin Sabet, Oct 18, 2012.

  1. Amin Sabet

    Amin Sabet Administrator

    Apr 10, 2009
    Boston, MA (USA)
    I recently bought a used IR-converted (720nm) Panasonic G2 and have been experimenting with it. It was converted by Spencers and the conversion was optimized for the 14-45mm lens it came with. It focuses that lens perfectly just as it perfectly focuses all my telephoto lenses (45mm, 75mm, etc). However, I noticed that my lenses which are wider than the 14mm end of the 14-45 won't focus to infinity. The Rokinon fisheye won't get anywhere close to infinity, and even the Panasonic 14/2.5 prime, which is a bit wider than the 14mm end of the kit zoom, won't quite focus to infinity although it seems to get to about the hyperfocal distance.

    The really interesting thing is that the 14-45mm zoom, which has great edge sharpness on my other MFT cameras, has poor edge sharpness on the IR-converted G2, while the 14mm prime that doesn't quite focus to infinity has good edge sharpness on both the IR-converted G2 as well as my other cameras. The net result is that for the G2-IR camera, the overall performance with the 14mm prime is a hair less good for the center at infinity (since it doesn't focus to infinity) but better everywhere else when compared to the 14-45mm zoom for which the camera was optimized at the time of conversion.

    Here's an example showing the soft edges and corners with the 14-45mm zoom at 14mm and f/5.6:

    Here's a test shot of the zoom vs the prime, both at 14mm and f/8 (warning, full-res files).

    I'm not disappointed - I got a great price, and the performance with both lenses is more than acceptable. In fact, the performance with the 14mm prime was surprisingly good to me. However, this is my first IR converted camera, and I was wondering whether these types of issues are normal for converted cameras.
  2. RevBob

    RevBob Super Moderator Subscribing Member

    Jun 4, 2011
    NorthWestern PA
    Now I'm curious to hear the answer. I'm seriously considering having my E-P1 converted - I don't use it much but don't want to give it up, either. :cool:
  3. RobWatson

    RobWatson Mu-43 Hall of Famer

    NIR trickyness

    NIR conversions really need to account for dispersion and the index of refraction of the materials between the sensor and the lens (in addition to the microlens array). Very significant differences in optical thinkness can result at those long wavelengths in many optical materials. Much to a proper conversion that just swapping out an IR blocking filter.

    The first thing one will notice in a less than top notch conversion is lack of infinity focus (optical distance has changed due to the above factors). Also, since the backfocus point has now shifted once well corrected abberations now become less well corrected and much more noticable - especially towards the edge. Also likely may notice in increase in vignetting.

    Also note that lenses (even some of those designed for IR/NIR) show different focus positions for IR/NIR focus distances . IR/NIR lenses that are designed specifically for IR/NIR will have correct distance markings-they will not focus correctly for visible ranges!
    • Like Like x 1
  4. Amin Sabet

    Amin Sabet Administrator

    Apr 10, 2009
    Boston, MA (USA)
    Hi Bob,

    I was reading this page on that site: Focus Calibration Options | LifePixel Digital Infrared Photography IR Conversion, Modification & Scratched Sensor Repair

    Based on that, I was thinking that maybe my converted camera is more limited than it has to be. I've read other reports of people not being able to focus their 7-14mm lens to infinity on a converted body, while still others say it was no problem. My Rokinon fisheye doesn't even come close.
  5. Amin Sabet

    Amin Sabet Administrator

    Apr 10, 2009
    Boston, MA (USA)
    So a truly top notch conversion of a Micro 4/3 should work well with all lenses at infinity without any edge softness beyond what would be seen on a visible light camera?

    Like I said, I'm perfectly happy with what I've got - the performance with the 14mm pancake seems very good and the price was terrific - just curious as to what is actually possible.
  6. RobWatson

    RobWatson Mu-43 Hall of Famer

    The upshot is even a perfect conversion still leaves issues with standard lenses. Dispersion effects in the lens are the source of the issue.

    Generally IR/NIR lenses can be designed using one visible wavelength (usually green) and two IR/NIR (or more than 2) so focus can be achieved at both visible (with chromatic problems) and IR/NIR wavelengths.
  7. Amin Sabet

    Amin Sabet Administrator

    Apr 10, 2009
    Boston, MA (USA)
    Can a perfect conversion allow all Micro 4/3 lenses to focus to infinity in IR?
  8. RobWatson

    RobWatson Mu-43 Hall of Famer

    One could manage this focus shift by fidding with a variable back focus distance to offest the NIR focus positions. The camera mo would need to include such a back focus mechanism as well as the 'standard' filter work.

    This way non-NIR lenses could be made to work 'better' with converted cameras.

    You are correct in stating that even a perfect camera ceonversion will still have issues since the original lens designs for non-NIR lenses will have significant shifts in focus due to unmanaged dispersion effects at the NIR wavelengths.
  9. RobWatson

    RobWatson Mu-43 Hall of Famer

    Sort of ... back focus adjustments are still required. To my mind a perfect conversion includes a variable back focus adjust mechanism.