Rebel T2i VS. 5D Mark II (pt.2)

For part two I’m going to get a little (ok a lot) more technical. This will probably scramble your brain (has given mine a good sizzle), but may shed an informative light. I know I was surprised!

The second test consisted of placing a Macbeth Color Checker on the wall, and lighting it (evenly) with strobes. I set a custom white balance on both the Rebel T2i, and the 5DMarkII, and captured one image each at ISO 100. All settings were identical with the one exception being the white balance measurement (dictated by the camera’s WB measuring capabilities).

I then loaded the files into Adobe Camera RAW, not applying any adjustments. Going square by square I took note of the Deviation in the RGB numbers of the image capture, and the “known” set of numbers for the Color Checker.
NOTE: I am not testing the “visual” accuracy of the cameras (since everyone sees everything differently) I’m testing the technical accuracy of the cameras.
Hover over the images to see which is which.

Here’s where it gets rough….

The 5d’s Deviation is as follows:
Red +16.125
Green +29.125
Blue +33.375
Absolute White AVG 237 (Chart =255)
Absolute Black AVG 33 (Chart = 0)

The Rebel’s Deviation:
Red +13.792
Green +29.292
Blue +31.5
Absolute White AVG 235 (Chart = 255)
Absolute Black AVG 45 (Chart = 0)

So what do those numbers mean?
The Red, Green, and Blue numbers represent the average point deviation from the Color Checker’s known value for those color channels, the lower the better. The only difference is the Absolute White and Absolute Black averages (I’ll explain this in a sec)
It may not look like it by the numbers but these are pretty standard numbers from any camera out there today.
“Superior” deviation results are considered to fall in the 10-20 range.

Now as far as how the cameras compared to each other…
The Rebel was more accurate in producing the Red Channel than the 5d, but to be fair, 3 points is really negligible (most color monitors fluctuate 3-5 “points” throughout the day while running). If it were 10, then there would be problems.
The 5D has an infinitesimal edge in the Green Channel
And the Rebel edges out the 5d in the Blue channel as well.

Now for the White and Blacks.
The white balance deviation on both cameras were virtually identical, and was only 1.5 points! So VERY accurate custom white balance computing.
When measuring the greyscale accuracy there are other things to consider.
One.. in the RGB colorspace, each channel’s value represents not only amount of color but brightness of that color.
So changing the brightness of an image effectively changes the RGB colors, and therefore the deviations.
With the neutral tones (which measure identical RGB values for each channel…ie “middle grey” is represented in RGB as R=128, G=128, B=128), I found that on both cameras the deviation increased as the values moved away from “middle grey”.
The “middle greys” on both cameras read 129.5 (accounting for the aforementioned 1.5 color deviation). The differences came at the extreme ends of the scale.

Now as you can see by the images above, the 5d’s image has more contrast.
Its black point number came in at an average of 35 (Chart standard is 0 ), but the Rebel’s only got to 45.
Nearly one stop less!
Both cameras white points fell in the 235-237 range. Negligible.

So does this mean the 5d has a greater dynamic range? No.
Both cameras capture the same white point, but the 5d’s image also represented a darker black. Is this greater range…NO, but it is a higher native contrast. The dark values of the 5d were consistently lower than the Rebel’s which means, straight out of the camera, the 5d’s file will have more visual contrast. But if you guys are familiar with “Exposing to the Right” you’ll know that more information lies in the brighter tones, than the darker ones. So the 5D isn’t “technically” capturing more information. Based on the “expose to the right” standard… it’s a disadvantage. The fact that the Rebel’s black point in comparison to the known value range of the color checker is brighter, means it has more information to manipulate. It may not look as good straight out of camera, but it has more latitude when it comes time to editing.

Ok.. enough science babble…

What does all of this shit mean?

It means that if you really want to split hairs over color that you can’t see… get a Rebel.
If you shoot RAW and refuse to edit it beyond your RAW processing defaults… Get a 5D (but only if you prefer more contrast).

Both cameras performed virtually the same technically speaking. Visually straight out of camera 99.9% of people would pick the 5d as better looking, but based on the numbers, it’s a toss-up.
There is no clear advantage to either camera based on my test results.

For the sake of argument, I’ll give the slight edge to the 5d for being visually pleasing straight out of camera (hey, that’s the end goal anyway right?). But in all actuality in this age of Photoshop, it’s not much of an advantage.

Round 2 goes to: 5D Mark II

Stay tuned for the 3rd and final installment where I break it down feature by feature and announce a winner!

Part 3