Michael Ballhaus: Over and Under Exposure Tests
Thursday, February 25, 2010 at 06:46PM The guy who started his career shooting feature films for RW Fassbinder and now shoots for Scorsese (amongst others) has taken the time to create exposure tests of the Canon 5D Mark II, Sony PMW-EX3, RED One, Sony F35, and Arri D21. The beginning of this test, the exposure bracketing, is an incredibly useful resource as it shows the true limitations limitations of these cameras. Disregard the Side by Sides at the end of the video as I see that several of the examples have been heavily manipulated in post production. Was this accidental, overlooked, or intentional? Don't know. Don't care. It's good to see the over and under examples on the 5 different cameras though.
Michael Ballhaus : HMS Kameratest - Testmaterial from zoom magazin on Vimeo.
I spent the afternoon at Kodak today and saw a film projection of Daryn Okada's 5219 demo and I've got to say, it made me feel a little fatigued by digital. That 5219 is just incredible. It sees so much and is so forgiving. It sees into the highlights 2 stops more than 5218 even which still was a great stock. You don't have to light so carefully and be so protective of your shadows and highlights. There's also a certain smoothness to the tonality that I've just never seen in a digital camera. We spend so much time these days looking at these tiny web videos that it's difficult to really evaluate the quality of a medium until you're seeing it massive in front of your face. I make my living with digital cameras and love the process but, DAMN, I'm just blown away by the quality of this stock. Even rated at 2000 ASA/EI and processed normally, the grain is remarkably tight. I hate this reigning attitude of film vs. digital and "why can't film just hurry up and die." Why can't both technologies co-exist peacefully? Just as manufacturers are developing new ways to digitally capture and process motion pictures, Kodak is developing stocks that are of equally remarkable technology. Yeah it's a cost issue and sadly, shooting 35mm color neg is not an option for most independent productions. But I'm glad that Kodak (and Fuji) remain committed to creating excellent products that still very much fit into the current paradigm.
Reader Comments (12)
Its 5D not D5
its also complete nonsense..
check the difference between how the 5d and EX3 look at the start of the clip compared to the so called side by sides where their saturation and contrast have been completely boosted beyond a reasonable level
That's a good point on the saturation. I just watched the full screen exposure bracketing where I don't think there's any trickery involved. I hadn't actually noticed that on the side by side but you're right, it's heavily manipulated. What's kind of test is that?
Well if you can get past the side by sides, its still good.
yes seems like that but after knowing how far wrong the ending is it kind of loses all credibility..
not sure what they were trying to show by manipulating the results, why they didnt manipulate the full screen sections aswell, or if it was some extreme mistake made in the testing!
originally discovered over here at dvinfo dvinfo.net/forum/canon-eos-5d-mk-ii-hd/474062-5d-showing-its-weaknesses.html#post1494815
everyone makes mistakes. Even legendary German cinematographers.
The test is credible and there is no trickery involved. In accordance with the ASC's recommendation for testing the usable latitude of film stocks, when conducting the tests the cinematographer should shoot 3 stops over and 3 stops under, then evaluate the results. Following this initial viewing and analysis, the cinematographer then should ask his color timer to try to correct the footage back to 'normal' exposure. This is where we see the image falls apart. What Michael's test demonstrates is that at 2 stops overexposure, trying to correct the 5D footage back to normal exposure significantly increases contrast and saturation of the image. Sure, one could simply desaturate the image in the post-processing, but this is not fixing the problem, only covering it up. The detail is still lost. I did the same test with a HPX-170 and received similar results, 2 stops over resulted in a strong shift in contrast and saturation once corrected.
Having not conducted these tests personally it's difficult to comment on the post-processing involved, but knowing Michael's track record and considering his ASC membership I feel I can assume that it was not his intention to mislead anybody with these side by side test results, only to inform. We live in a world of ever-changing technology where different workflows are being created and discovered daily. I consider it quite courageous on Michael's behalf to attempt these exposure tests on 5 different digital cine cameras each with it's particular quirks and post-processing techniques for pulling the best image out of the camera.
Before discrediting Michael Balhaus and the merits of his tests or knowledge, I would respectfully write to him to suggest better ways of processing the footage in question. Send him your tips and tricks, because these tests are all in the spirit of learning and informing so that we as cinematographers know how far we can push the limits of our tools during production.
The test is credible and there is no trickery involved. In accordance with the ASC's recommendation for testing the usable latitude of film stocks, when conducting the tests the cinematographer should shoot 3 stops over and 3 stops under, then evaluate the results. Following this initial viewing and analysis, the cinematographer then should ask his color timer to try to correct the footage back to 'normal' exposure. This is where we see the image falls apart. What Michael's test demonstrates is that at 2 stops overexposure, trying to correct the 5D footage back to normal exposure significantly increases contrast and saturation of the image. Sure, one could simply desaturate the image in the post-processing, but this is not fixing the problem, only covering it up. The detail is still lost. I did the same test with a HPX-170 and received similar results, 2 stops over resulted in a strong shift in contrast and saturation once corrected.
Having not conducted these tests personally it's difficult to comment on the post-processing involved, but knowing Michael's track record and considering his ASC membership I feel I can assume that it was not his intention to mislead anybody with these side by side test results, only to inform. We live in a world of ever-changing technology where different workflows are being created and discovered daily. I consider it quite courageous on Michael's behalf to attempt these exposure tests on 5 different digital cine cameras each with it's particular quirks and post-processing techniques for pulling the best image out of the camera.
Before discrediting Michael Balhaus and the merits of his tests or knowledge, I would respectfully write to him to suggest better ways of processing the footage in question. Send him your tips and tricks, because these tests are all in the spirit of learning and informing so that we as cinematographers know how far we can push the limits of our tools during production.
Please excuse the double post.
If I cared about his test that much, I would write to him. I don't care. We all do camera tests. There are dozens of test shots and comparisons on this blog some of which are probably full of mistakes. It's part of the testing and learning process. Criticism and scrutiny are also part of the process.
Personally, I think it's good that somebody pointed out something suspect in the video because it raises a good point, that you shouldn't take anything at face value even if a brand name is attached. All I saw was Michael Ballhaus' name and I posted it to the blog without question. Upon closer inspection of the work, I agree that side by sides are a little strange. I still think it's a great test and is very useful which is why I didn't delete it from my public blog.
You create something and you put it out there and then people form opinions on it.
Ultimately, it's up to the end user to best decide how to interpret the product.
If you really want to participate in a discussion on this, I suggest going to the thread on DVInfo.net where the debate started.
dvinfo dvinfo.net/forum/canon-eos-5d-mk-ii-hd/474062-5d-showing-its-weaknesses.html#post1494815
Patrick. The first of the side by sides at the end is at normal exposure. It is not at +2 and then corrected. You can see that at normal exposure the 5d has suddenly completely changed and it is not at all accurate to the image quality capable with this camera. Why this has happened i don't know but it is either deliberately misleading or a very bad mistake. How this mistake could be made without noticing and taking into account that the earlier tests do not show this i have no idea.
Michael Ballhaus is NOT a legendary german dop, but an enormous marketing talent for himself. Technically he is in fact quite weak and sure doesn't hesitate in manipulating "tests". MB is and has allways been a selling-machine of his own …
Who cares?