Intro
Ever since I acquired my first infrared-converted camera (the 630nm Pentax K-5) I tremendously enjoyed infrared shooting and wanted to expand my horizons a bit more. Some 4 months ago I acquired a Pentax K-3 which – same as last time – was converted to infrared by Sven Lamprecht of IRreCams.de and like before, I have absolutely no issues with the service. Conversion was fast and the shipping and payment solutions offered were very satisfactory. This time I wanted more flexibility so I had this body converted to full-spectrum (280nm I believe) which would enable me to experiment with various wavelength filters.
I had a blast shooting at 550nm (acquired from IRreCams again), 590nm (China Rocolax brand via eBay), 665nm (Heliopan from a German distributor) and finally a 720nm (a trustworthy Hoya quality R72 filter purchased from a Dutch photo-superstore) and using various techniques to extract different colour and/or B&W versions using Darktable as my raw converter of choice and Gimp for post-processing of the final images. Of course the videos provided by IRreCams were excellent guidance but I cannot neglect to mention various Youtube channels and software sources such as Rob Shea and Nicholas both with excellent tutorials on infrared processing. Inspired by those resources and other websites and forums, I also invested some time into working with colour LUTs (Look Up Tables) to complement the channel swap technique I thought I had started to master over the past year. Examples can be found in my Infrared album online.
So, all is well?
No, not really. I still had to choose which camera body to bring along when traveling and although as long as I stay close to home it is no problem to bring more than one DSLR, when on business trips or on family vacations I thought it would be convenient to have a single camera to shoot both IR as well as visible light photography. Enter the concept of a “hot-mirror” filter. Normally, that is the filter which sits over the sensor in any regular camera and prevents near-infrared and ultraviolet light to inversely impact your regular and expected image quality. As I explained in an earlier blog, that is the filter which infrared conversion services remove and replace when converting a camera to infrared. A hot-mirror filter can be used after conversion to re-enable the camera to take regular visible-light images.
First thought – a clip-in filter
My first thought was to get a clip-in filter which sits in the lens mount and allows any lens to be used without external screw-in lens filter (expensive to get in many thread sizes). It’s a bit fiddly to insert but I thought it would offer a convenient way to test my hypothesis of having an “all-in” DSLR for light weight travel. I use Pentax and there only was one manufacturer who offered clip-in filters for Pentax K-mount, STC Optics in the US. Their (no doubt marketing-based and somewhat over optimistic) claim was:
STC
Clip filters are customer designed for each camera model. We support Canon, Nikon, Sony, Fujifilm, Pentax, Olympus, Blackmagic, Z-Cam and Panasonic models.
Our engineers tirelessly work to eliminate the color shift that can occur with lesser brands. We
invest heavily to create coatings that work with wide-angle lenses using interference coatings giving you greater creative options.
Although the shipping cost was horrendous and I was required to pay import VAT to FEDEX after the package arrived (expected but still not very nice as a surprise), the diminutive little plastic box arrived at my doorstep after a few weeks and I started experimenting with it.
Reality sets in
I had read in various locations that those clip-in filters, although seemingly very handy, had a number of more serious downsides. A few of those downsides I had already encountered and some others I started to find out about:
- Clip-in filters are small and somewhat flimsy, the metal is easily bent and the filter itself is extremely thin
- Inserting and extracting the filter from the lens mount is not at all easy. On mirrorless cameras, there is a danger of scratching the sensor, on DSLRs the mirror can easily be touched
- If the filter is installed only slightly at an offset, it may prevent mounting the lens or worse it may lock the lens in place necessitating a costly repair
- As the filter sits in the mount but in front of the mirror on DSLRs, lenses with a protruding rear element cannot be used
- The filter is extremely sensitive to light entering at oblique angles, introducing unwanted reflections and/or loss of contrast and sharpness in (ultra) wide angle lenses under certain situations
Next step
As I had been using the STC filter mainly with my DA35mm/f2.4 and FA43mm/f1.9 Limited series lenses until recently, I had not really had a chance to evaluate its downsides. Most of my non-IR shooting was indoor family gatherings and not so much the architecture and nature shots I take when out and about. But I’m always interested in improving image quality and expanding my knowledge so over the period of a few weeks I had started to think about acquiring a lens-mounted hot-mirror filter and I reviewed various options by Kolari Vision and Life Pixel in the US, Dutch Digital Works in the Netherlands and of course IRreCams in Germany (by the time I’m writing this blog, I could not find the hot-mirror filter as an option in the DDW webshop anymore sadly).
Disclosure
I ended up asking mr. Lamprecht if he would be interested in a test of his newly released “Normal Plus” filter which is promoted as having
- Improved color rendering, especially neutral colors when using the camera’s internal automatic white balance
- Hydrophobic surface coating for easier cleaning
- Slim filter ring reduces the risk of vignetting (a front thread is available on both versions)
In the interest of neutrality, the agreement is that I will return the filter to him at the end of my testing so both Sven Lamprecht and myself are very conscious of wanting to avoid any kind of bias into this blog-post. I also offered to let him have a say in whether and how I will post my results and asked him if he’d like to review the raw files prior to publication and his answer was as follows:
To be clear, this is your work and blog post, you can publish and write anything that you want. I don’t want to influence your outcome on this in any way!… I might be interested in a summary of your text, or something like that, for publishing it in our blog or newsletter. But we can discuss this as soon as you have published your article.
So, that should tell you all how I decided to go forward and how objective my testing will be, given the limitations of my technical abilities and the time available to me. My basic assumption is that I’m free to come to any conclusion I see fit without any financial or other pressure or any intervention from the commercial party, IRreCams.
Mister Lamprecht also came up with some really good advice as to my suggested process communicated early on in an email and (summarized) he said the following:
- shoot all tests in AWB mode
- compare the outcome between the camera/filter combinations
- correct the WB in in darktable for all images in a next step
- see how the colors perform in this optimal condition.
- test the filters in their known weak spots: backlight situations and inside a forest.
- start with comparing color rendition
- eventually also compare sharpness
My process
Hardware:
- my regular 24 megapixel Pentax KP acting as the control
- the full-spectrum K-3 as the testbed for filter evaluation.
- the same lens on both bodies, the Pentax DA16-85mm/f3.5-5.6
- the STC clip-in filter
- IRreCams “Normal Plus” filter in a 72mm thread size fitting directly on the front thread of the lens
- NOT USED: I decided to forego the use of the dedicated lens-hood in order to exaggerate and highlight any flare and/or contrast issues in back-light situations
Software:
- To convert the raw files, I used Darktable version 4.6.1
- To line up the handheld images and create side-by-side comparison images, I used Gimp 2.10.38
- To create (labeled) thumbnails in a size suitable for web publication I used ImageMagick version 7.1.1.32-1
- The platform was Manjaro, updated a few days ago to the newest stable 24.0 Wynsdey release
Shooting procedure (in my local park/forest on an open spot at ca. 10:30 AM on a sunny day):
- With my back placed firmly against a tree, I shot a series with the KP consisting of:
- A wide shot (16mm) with the sun just outside the top of the frame roughly straight ahead (lots of sky in the image)
- A wide shot (16mm) at 90 degrees to the sun (some blue sky visible)
- A wide shot (16mm) with the sun at my back (mostly trees and green foliage)
- The same angles again but now with the lens set at the 35mm marking
- Without moving from my position and only unmounting the lens off the KP and onto the K-3 with the clip-in filter installed
- The same 6 shots as above so three at 16mm and three at 35mm and at angles of 0, 90 and 180 degrees to the sun
- Again, without moving, I unmounted the lens, removed the clip-in filter (very fiddly indeed) and remounted the lens this time with the normal plus filter installed
- Boring but the same 6 shots again with this combination
I did not bring a tripod as I was already weighed down with 2 DSLR bodies, a camerabag, extra batteries, a lens and a dog…
I did try to keep framing consistent and where possible, I overlaid the images afterwards in Gimp layers to line them up at least a little.
Raw conversion:
- as suggested my mr. Lamprecht I performed a basic raw conversion with WB adopted from the camera initially and a minimum of modules activated to minimize any impact
- in a second step I left everything the same except for the WB module where I opted for darktable to apply a WB computed from the image area (auto-WB)
Preliminary outcome
Without wanting to make a final judgement as I need some time to evaluate the 18 resulting images in detail but I can share them for you to have a look for yourself. You can find them in my online album surprisingly titled “Hot mirror test“. What I can do for now is to also share a few of the images I combined in Gimp either side by side or, when there’s a lot of sky involved, stacked vertically. In all of them the sequence is:
- Control shot KP, no filters, no hood
- K-3 shot with clip-in filter, no hood
- K-3 shot with 72mm normal plus filter, no hood
Backlight, WB uncorrected, 16mm:
Backlight, WB uncorrected 35mm:
Shot at 90 degrees to the sun, WB uncorrected:
Shot at 90 degrees to the sun, WB corrected in darktable, different crop to include both foliage as well as sky:
