After previously looking at the light distribution, or Vignetting, in this article I would like to compare the amount of light that reaches the sensor, as well as its colour.
05.01.2016 – Update: I added the results from my measurements at f/2.8 and wide open and calculated the average values too.
17.01.2016 – Update: After curing the Rikenon 55/1.4 for four weeks, I did a quick comparison against the Color-Ultron and am happy to report that the colour cast is gone. See result below.
What is Light Transmission?
Photographic lenses are mostly sold with indications of maximum f-stop and all my lenses, except for the Yongnuo, in this comparison have an aperture ring to manually select the aperture you want to take a picture at.
The aperture is a simple mathematical formula used to approximate the amount of light a lens will let through. It’s quite genius too, because for any given focal length on any given sensor, a certain f-number will always result in the exact same exposure, provided the shutter speed and the iso sensitivity are identical.
There’s just one tiny problem with that:
Depending on the shape of the aperture, the effective focal length and the actual size of the aperture, the amount of light gathered can actually vary quite a bit (most of these values are roughly rounded (pun intended) for marketing purposes).
Usually, though, this isn’t much of a problem, because you take a picture with one lens only and in both aperture and shutter priority modes the camera measures the light through the lens anyways. Also, if you shoot manually, you usually focus on the Histogram to properly expose to the right, so you don’t really have to bother with these comparisons.
For videographers, on the other hand, the transmission is critical, because you don’t want the brightness to change when switching lenses (Much like the “uniformity” in rendering across lenses of the same series of video lenses). For that reason, manufacturers of video lenses usually indicate the T-stop instead of an f-stop. The T-stop indicates the measured transmission of a lens and is much more precise.
For me, there’s one additional problem, though:
When a lens says f/1.4 but the images shot with it wide open aren’t even brighter than those from an f/1.8 lens, I feel cheated.
What’s a colour cast?
Back in the film days, white balance wasn’t something you could easily correct in post processing and the camera itself had absolutely no option for that. There was film for artificial lighting, but most of the film was intended for daylight (6500 K).
If your lens is perfectly clear, it will deliver a neutral grey when shooting a grey card in daylight. However, this is often not the case and colour casts are especially strong with radioactive lenses (Read more on that topic in my article about radioactive lenses ).
Radioactive lenses often come with a yellow tint from the radioactive decay in the glass:
So, if your lens has tinted glass, it will obviously “filter” the light passing through, passing on that colour tint and rendering images warmer. Some people swear on the use of tinted radioactive lenses for monochrome photography (because yellow filters are very popular in that category).
How did I measure the colour cast / light transmission?
I sat the camera on a tripod and put the grey card (cheap chinese one) on the wall, then I used a continuous light source (45W putting out 5’500 K) to guarantee unchanging conditions for all contenders.
I used the Color-Ultron for reference, because it performed admirably in the past.
This was my setup at a distance of 0,8m with the lamp positioned above and slightly to the left behind the camera to provide even illumination:
I used the following exposure values for all lenses: manual mode with the shutter at 0.5s at iso 100 (best color quality on the sensor) and f/5.6 to have the same aperture on all lenses.
The only thing I did in Lightroom was cropping the center portion, setting the camera profile to Neutral and of course set the white balance identical for all the lenses.
After correcting the white balance in Lightroom. (Temp 5’300K and Tint +22 was suggested by the WB selector on various spots of the grey card) the picture from the Color-Ultron looked like this:
Compared to my first series of colour-cast comparisons, this image looks pretty good to me.
Here’s an overview comparing all the images at identical white balance settings and nicely showing you what an impressive influence the colour cast and transmission loss can have on the final image if not corrected:
Notice how the Rikenon is extremely dark with a strong bronze colour cast. I have no information whether this lens is in fact radioactive, but I will put it under the Jansjö lamp anyways.
The Takumar (radioactive) looks great compared to the rest of the lineup, the colour cast was almost completely removed using my treatment, whereas the Fujinon and Pancolar both still suffer from rather strong yellowing.
The Helios is pretty dark and it’s the same at f/2. Looking at previous tests wide open, it’s obvious that this lens unfortunately can’t even provide a true f/2 aperture.
Two important things to note:
- The Olympus has no aperture stops, because the previous owner must’ve opened it and did a sloppy job at putting it back together. Comparing it to the other lenses at f/5.6 is therefore not really possible.
- The Tessar has a very sticky helicoid and focusing is almost impossible. Also, it seems as if something is terribly wrong inside, because it said 1,2 metres where all other lenses were at 0,8-0,9m on their distance scales. Also, at f/5.6, the Tessar was in fact brighter than the brightest lenses. Therefore, I took another image at f/6.3 on the aperture ring for the numbers part below.
I assumed the “true” White Balance to be the one measured from the Color-Ultron and used the eyedropper for each lens to adjust the White Balance individually.
Additionally, I adjusted the exposure (after correcting the White Balance) for each image crop so that the histogram was in the exact same place for all lenses.
In the chart below you can see these “measured” values for each lens. I coloured the numbers to indicate what they mean.
You see that the Rikenon has a strong yellow cast, pushing the Temperature by over 600K. Combined with the rather low shift towards green, this results in that hefty bronze tone. Unfortunately, the lens swallows almost 1 full stop of light, which makes it slower than an f/1.8.
The Pancolar still has the strongest colour cast of all lenses, with +700K warmth and -29 towards a green tint.
Looking at the table, you will notice that most lenses can keep the temperature very close to neutral (+/- 100 is harmless in normal shooting conditions).
But most of the contenders suffer from noticeable green tint. (Example: shifting the tint by more than 5 usually leads to weird skin tones in portraits.)
I was shocked to find the Planar to be 0,22 stops slower, because it’s great in all other areas. In fact, I was so surprised, that I took several images and looked at the histogram closely. I will do another series of measurements wide open, to see the maximum light gathering capability of these lenses.
Except for the Helios (0.6 stops) and the Rikenon (0.86 stops) no lens is suffers more than 1/3 of a stop of light loss compared to the others. That’s a good result in my opinion.
* Keep in mind, that the Tessar was shot at f/6.3 for the comparison, because it was too bright at f/5.6, and that the Olympus has no fixed aperture values.
Measurements at f/2.8
Here’s the chart with all the lenses set to f/2.8
The Olympus, again, was set to approximate f/2.8 because it lacks fixed f-stops on the aperture ring.
The Color-Ultron isn’t the brightest lens at this aperture setting. The Takumar is brighter by as much as 0.2 stops and the Yongnuo by more than 0.1 stops.
The Rikenon once more is almost a full stop darker than the brightest lens in the comparison.
Another interesting thing is the fact that in this round no lens was producing colder colours than the color-ultron. Actually, the Helios and the Tessar were significantly warmer than the Ultron this time around.
Measurements wide open
In this round (shot alongside the f/2.8 measurements) I had to adjust the shutter speed to match the selected aperture values. Based on my previous measurement of the Nikkor 50mm f/1.2 Ai-S – that led to believe that it’s actually only 1/3 stop faster than the f/1.4 lens – I could easily do so with my camera set to 1/3-stop increments.
Notice how the transmission value is at or near 0 for many lenses once shot open. This confirms the theory that the aperture shape has an influence on the transmission (because wide open the aperture has no influence anymore).
Also notice that the Rikenon performs like an f/1.8 lens instead of an f/1.4. I really hope that I can improve that with my Jansjö lamp.
The fact that some lenses now display a noticeable shift towards blue is also quite irritating. I attribute that to the fact that at such large apertures the ambient light (the test was not performed in a completely dark room) starts to have an influence on the images.
Just for the sake of comparison, I also did an average of the three sets of measurements:
Please keep in mind that I do not have any professional equipment to measure the transmission of my lenses and that my comparison is based solely on the White Balance tools and Histogram in Lightroom.
Small variations are always possible, but changes over 150K in Temperature and 5 in Tint are easily noticeable. I would also claim to achieve an accuracy below 0,1 EV in the Histogram.
17.01.2016 – Update
I had the Auto-Rikenon 55mm f/1.4 under the Jansjö lamp for a full week now. After this treatment I took some more pictures with the Color-Ultron and the Rikenon at various apertures, then I followed the same protocol as above and visualized the results in a chart.
When looking at the average above, you can quickly see that the Colour cast went from more than +550 to almost zero! The tint was reduced ever so slightly and the transmission also improved by more than 1/3 of a stop at the given aperture values.
I compared the Rikenon to the Zeiss Planar T* 50mm f/1.4 wide open and the Rikenon is only 1/3 of a stop darker there. (vs. 2/3 of a stop before)
Now I have to look up further information, because it’s possible that the Rikenon contains radioactive glass after all. Unfortunately, I don’t have a GM counter at hand at the moment, so I can’t conduct any measurements.
I decided to put the Helios under the curing lamp just to see if it would improve too. We’ll see in a week.