Upload & Sell: Off
I was bored this weekend so I decided to make a little comparison of this fantastic lens with and without the equally fantastic Canon 2.0X TC III that I just acquired.
The results were not surprising in the sense that I expected to see some degradation in IQ, and I did. What was downright shocking is the small amount of degradation there was.
Based on these tests, I would not hesitate to pair up this lens and this TC for any situation where I may need to. The biggest thing I noticed at this ridiculous magnification was a loss of contrast. The loss was slight enough that that in a real-world scenario, it may not even be noticeable and should be totally correctable.
Please feel free to comment and ask questions. I am interested in agreements/disagreements, whatever. I have all the images so if you want to see more, just ask.
A few caveats, conditions, and notes.....
1. These tests were conducted in a studio under controlled conditions. I am not a pixel-peeper and tend to value real-world results more than lab tests. This sort of thing goes against my grain, but I did it anyways. YMMV. If I have time, I may do a similar test with more realistic subject matter.
2. These test were conducted all at 200/400mm. The long end is usually the worst FL for zooms, so rather than be rigorous and fill up a hard drive with images at all FLs, I picked this one. I may do more later.
3. As mentioned above, these tests were conducted in a studio. Distance to subject was roughly 2 meters for the 200nn (no TC) test and 4 meters for the with-TC 400mm test. This was obviously to preserve a similar FOV. It would have been interesting to compare 200 to 200 at the same distance, but since most people use a TC to gain reach, I didn't see the point in this. Either way, neither distance is representative of a typical distance at which this lens may be used. IDK if the results would have been different further away. Comments welcome. Sorry, my studio is not that big.
4. As evidenced by #2 and #3 above, I am somewhat lazy by nature when not getting paid to do something, so in furtherance of this laziness, I used FoCal's excellent aperture-sharpness testing function. This wonderful software will automatically cycle through all the apertures of a particular lens and after a few minutes, hand you a nice PDF (linked below) full of interesting data. All I had to do was introduce Mr. Camera to Mr. Laptop and click a few buttons. The downside: The samples are all center crops. If I get time, I will do some corner comparisons, but I'll have to do those manually. I suspect there is more degradation in the corners than in the center.
5. All Images were made on a 5D2 mounted on a sturdy Manfrotto tripod.
6. EV was 10.0 (pretty bright) as measured by a Gossen Digisix light meter. FoCal seemed to agree as it reported EVs of 9.5 to 9.9 with some unexplained variance shot to shot.
7. The full image is not in the PDF, so I have posted it below so you can see the extent of magnification in the samples. As far as I can tell, FoCal is showing about 3% of the image, which is what you see in the side-by-side shot below. Anyone feel free to correct my math if I'm wrong. Either way, its a ridiculously high level of magnification even for you pixel-peepers.... I think..
8. I have also posted the best shot with and without TC for a side-by-side comparison. "Best Shot" is defined as the best aperture for the lens alone and with TC. FoCal reports this and I agree that F3.2 and F11 are the best apertures respectively without and with TC. As you can see from the No-TC PDF, there is little variance in IQ below F16. This is an amazing lens by itself.
No-TC test: http://geekpix.com/host/pdf/70-200.pdf
With TC test: http://geekpix.com/host/pdf/70-200_2.0x.pdf
Full Frame for test
Side-by-side of best F-stops.