English - Espaņol
Sharpness - Bokeh
Foto tutorial (English)
Foto tutorial (Espaņol)
Equipment recommendations US-ES
Flaat for Canon
Flaat for Nikon
Flaat for the BMC
Flaat for NEX-5N
Old Picture Style Tests
550D video lineskip
APS-C vs Full Frame
Badly assembled lenses and image quality
Lens mount compatibility chart
ISO on different cameras
High ISO on the 5D3
DIY: DR test chart
RGBWK Bayer sensors
Notes on DoF-FoV
Notes on crop-DoF-FoV
Custom Cropmarks for Magic Lantern on the Canon 550D
How many megapixels do I want?
How many megapixels can I see?
Quick Monitor Calibration Chart
As part of the effort that finally yielded the Similaar suite of Flaat picture styles for Canon DSLRs, I created a DR test chart using my TV. Here's how I did it.
DR test charts are expensive. I'd like to have one, but it's an expense that doesn't make sense for me. Luckily, I have a nice Sony TV (KDL52W4000, permanently hooked to a PC, and reasonably well calibrated: I have a ColorVision Spyder 2 Express), and the manufacturer claims it's got a 33000:1 contrast ratio (16 stops of latitude). Unluckily, those claims are never true, and my measurement says it's around 11 stops, so 1000:1. Just about enough for my T2i (550D).
Thanks to Magic Lantern, my T2i (550D) gives me a reasonably accurate brightness reading at the center spot. I put the camera on a tripod in front of the TV, a few meters away, using a moderately long lens (Leitz Elmarit-R 90mm f/2.8) set at f/4 (to avoid vignetting) and with the focus point a bit before the TV (just so the individual pixels are blurred).
I set my desktop background to pure black (flare was problematic), opened Microsoft Paint (yes, Paint; that PC is a media center, not a workstation, but Paint is enough for this), and created a blank 1000x500 image, which I filled with 1% grey. In the center of that, I created 12 identical rectangles, filled with 20% grey.
I filled the leftmost rectangle with pure white, turned the lights off and blocked the windows reasonably well (doing this at night helps).
And then I started a slow and painful process of filling each successive rectangle with a shade of grey that's displayed by the TV as exactly 1 stop darker than the one to its left (this takes at least an hour, much more if something goes wrong, e.g. because lens flare, however minimal, starts to mess with you on the dark side of the chart).
That process worked as follows:
* point camera at reference rectangle, set shutter speed and ISO so I have an IRE reading of around 50% and no flickering; take note of said IRE reading (e.g. 56%)
* point camera at the next rectangle, and adjust ISO or shutter speed so exposure is pushed up one stop
* reach for the keyboard+touchpad, and fill the rectangle with the appropriate shade of grey that gives me the IRE reading I want (in the example, 56%)
Rinse and repeat. When finished, re-check and adjust everything carefully (this step is also needed if you want to use the chart again a couple of days -or hours- later).
In the end, I was left with one square that couldn't get to the appropriate IRE reading even when filled with 0,0,0. First thing that came to my mind was to carefully stick a piece of black cardboard in front of it, and adjust position of far-away 1W LED light until I get the IRE reading I want. This led to pain and suffering, as minimal flare and reflections in such a dark environment mean everything changes with the slightest thing.
So I just forgot about this square, filled it with the background color to make it disappear, and settled for an 11 stops chart (half a stop less than my camera could handle).
This is a downsampled version of the end result:
It may look really odd on your screen, but when displayed in my TV at optimal settings for DR test chart generation (which required backlight increased beyond what I consider confortable for daily use), each strip in this chart is exactly one stop darker than the one on its left (well, just as "exactly" as my infrastructure would allow).
I also used the following trick: take 7 shots, changing shutter speed by 1/3 stops in each one, then combine them in the PC. This provides a check that the DR test chart is accurate (it could have been spoiled by things like flare and vignetting), and a way to construct a more detailed virtual chart that works in 1/3 stops instead of full stops (as long as I use a manual lens and shutter speed is relatively slow, random variation between shots should not be a problem).
This is how that virtual 1/3 stops chart looks like:
And this is the waveform of said virtual chart shot with my T2i (550D) in Flaat_2 picture style:
The fact that each horizontal green bar is not thick or divided in several thin bars is what tells me the experiment was successful and the chart is reasonably accurate. There may be some issues with the brightest rectangle in my chart, but they're modest issues anyway.