Blog

Musings by Snehal aka Fearless, a filmmaker with a hankering for technology and science.  

Single Chip Camera Evaluation (SCCE) screening at Hollywood DI

I was really happy to get an email from Neil Smith of Hollywood DI announcing a screening of Zacuto's camera shootout between 11 digital cinema cameras available on the market and 2 Kodak film stocks.  I missed the NAB screening and looked forward to seeing the video in a controlled calibrated environment with a 2K projector. 

The Single Chip Camera Evaluation (SCCE) is a 26 minute film of tests administered by Robert Primes ASC, between these cameras:  Arri Alexa, Sony F35, RED ONE MX, Sony PMW-F3, Panasonic AF-100, Weisscam HS2 Mk2, Phantom FLEX, Canon 5D MkII, Canon 1D MkIV, Canon 7D, Nikon D7000; and film stock: Kodak 5219 and Kodak 5213.

I have to admit, there were a lot of surprises for me after seeing the cameras in a dozen or so tests that pit them side by side using the same setups and conditions.  The evaluations varied from shooting spinning disks of white lines which showed how rolling shutters create distortion to filming of scenes under various light conditions and skin tones.

The camera footage was "hardly" color corrected as Neil pointed out and the best system for recording was used when possible.  Scott Lynch, from Zacuto, said this meant the higher end cameras were coupled with a Codex or something comparable using 444 when possible (dual HD-SDI required).  The DSLR cameras were all recorded using in-camera technology since none provided a clean output HD signal without some sort of overlay.

We watched the film twice, once all the way through and then we watched a second time, pausing so that the tests could be discussed.  I found the detailed discussion about the SCCE led by Neil and Scott to be of immense value plus Zacuto was recording our thoughts for a SCCE documentary.  There were a lot of talented DPs in the room that had their opinions and Aaron Peak, Hollywood DI's resident colorist was on hand to answer post-production queries. 

Gale Tattersall, DP of House, was in the audience as well and had more than a few things to say about Alexa and Canon DSLRs.  Apparently House is going with Alexa for it's main camera next season, recording in ProRes using Log-C.  It's the same workflow that Rodney Charters told me he was using on another show right now.  For the TV guys, its really important to have a straight-forward, solid workflow that is easy to use and doesn't require high-end colorists to finish correctly.  The Canon DSLRs have become ubiquitous for a lot of TV DPs and they carry this little around from job to job.  As Gale pointed out, it's the only camera he uses for in-car shots since it's easy to position in tight spaces and can quickly be Velcro-ed to the hood if needed.

So the surprises???  ...well let me just give you a quick rundown of my reaction to each camera based only on these tests.  Mind you, this was by no means a scientific endeavor since a lot of variables were uncontrolled, but it was a good way to see how the cameras behaved in basic working conditions.  The price point of the cameras ranged from $1000 to $250,000 for the body only, no lenses or accessories.

Alexa - this was a class leader in many ways.  It has a very pretty image that is clean even under magnification.  It's not contrasty, there are no artifacts to speak of and the dynamic range is incredible.  It's a lot more sensitive to light both above and below the "sweet spot" than film can ever be.  The rolling shutter artifacts mimics that of Arri's own film cameras so it doesn't bother me in the least.  The only test where the Alexa came up a bit short was color rendition but the SCCE was not done using ArriRaw so I would be interested to see what those results would look like with the latest iteration of Alexa at 2K ArriRaw.  BTW - The lower resolution does not hinder the test camera in least.  The image under every lighting condition was beautiful, surpassing the look of film in my opinion - yeah that's right, I said it!  Pricepoint - mid-range.

Sony F35 - very very good camera.  Image quality is fantastic.  Sony really understands HD video and this camera does the company justice.  Sony has some of the best color rendering technology on the market and this shows in the F3 as well, the little brother to the F35.  Skin tones were nice, the camera was one of the better one when it came to rolling shutter artifacts, and it has a good dynamic range with nice low-light performance.  It does OK in the highlights but is very good in the mid-range.  Expose it well and this camera rocks. The only drawback to this camera is the incredibly high price point - 250K! 

RED MX - It's OK.  Aaron reminded us about something important when dealing with RED cameras; they require a lot of massaging in post.  You basically have to color the footage no matter what.  It's great for filmmakers because if given the time, you can make RED footage look awesome.  There is a ton of resolution and REDCODE has a lot of latitude in both color and exposure.  That being said, basic imagery coming out of the MX looked sharp and not all that sensitive in the low or high end of exposure.  It does great in the middle but so so with skin tones.  I know a lot of people are interested in pushing the resolution equation, but I'm not sure if the tradeoff is worth it.  More resolution means tiny pixels, which means less sensitive pixels that have less color choices.  That's a reality no matter how you slice it.  Sure you can go back and paint the pixels all you want in post, but that is time consuming and not always affordable.  I foresee a lot of TV shows going with Alexa or even the F3 over RED MX in the near future.  EPIC might be a different story.  I need to use it a few times to draw my own conclusions.
 
Sony PMW-F3
- SURPRISE!  This is the one camera that continually made me sit up and take notice.  It's incrediblly affordable for the kind of ability it has.  I've seen earlier footage from the F3 which I thought was very "video-like" (it's a term I like to use which means that the image reminds me of NTSC-style interlaced video even when the camera is set to shoot like film in 24p, etc.)  But when I saw the SCCE, it got me hooked!  Like it's bigger brother the F35, the camera performs best in it's exposure sweet spot.  Not as good in the highlights as Alexa, AF-100 or even the Canon DSLRs, yet this does not stop the camera from looking amazing in almost every other lighting condition.  Built with a PL mount and costing around 17K, this beauty is a bargin.  Size-wise it's perfect for 3D, crash cams, rigging, handheld, steady-cam and a lot of other things.  Do not dismiss this camera!
Check out Vincent LaForet's reactions to this test and other cameras from NAB posted on his blog and be sure to watch the F3 S-log test he did with Next Level Pictures.

Panasonic AF-100 - Just OK.  Kinda sharp, with weird compression-type artifacts when the image is blown up.  It has strange yellowish tinge that appears when the highlights are overexposed, even though it can see into the white a lot compared to more expensive cameras.  The micro 4/3rds sensor size kind of makes it annoying to lens since it's a 20% zoom from Super35.  I do think that this camera does really well in the mid-range and there is a lot to be said for the gamma curves that are built into the menu system.  I've actually shot some great looking footage with the AF-100, but I wonder if it really is better than using a DSLR aside from the fact that the form factor makes sense for professionals.

Weisscam HS2 Mk2 - Didn't like it in just about every test.  Very "video-y" image for my tastes.  To be fair there were no high-speed tests done as a part of the SCCE so I don't know what it's true capabilities may be.  But in normal operation it is neither sensitive nor pretty. 

Phantom FLEX - Pretty cool.  For a camera designed for high-speed use, the FLEX looked really nice in most of the tests.  It might not have the dynamic range of Alexa, film, RED or some of the other cameras but it preforms well with the right exposure.  It's not overly sharp or contrasty and it was the winner in the rolling shutter artifact test - helped by it's Global shutter technology.  It's not so bad in low light either and would be perfect for a nighttime stunt car chase done in slow-mo.

Canon 5D MkII - The Canon DSLRs did suprisingly well in these tests including the papa of the bunch, the 5D MarkII.  I went in expecting the other cameras to blow away the Canons, but I was wrong.  The 5D does really well with exposure and overall picture quality.  It is strong in low light, more so than film under certain conditions. It does suffer from a softer image, but a lot of that has to do with the sensor size and short depth of field (making the 7D preferable to a lot of film/TV folk).  Of course the shutter artifact test where spinning white lines look like they are jazz trumpets made the audience giggle more than once, but so what - stop whipping the camera around and you won't notice it.  The colors were overly saturated in the tests but they were using a form of Neutral picture style (Shane's settings I believe) and Gale mentioned that using the new LUTs from Technicolor might change color behavior.

Canon 1D MkIV - wow, what a piece of crap compared to it's less expensive siblings the 5D and 7D.  Wow.  I mean every test except super low light.  Well, Darren did use it in Black Swan during the subway sequences and 99% of the people that saw that film had no idea.  So what do I know, right?

Canon 7D - Winning!  I love this camera and this test reinforced everything.  It's plenty sharp even when blown up but still looks pretty, it has a great dynamic range, it hold a LOT of info in the highlights, does pretty good in lower light and color is decent - just a little bit more saturated than reality.  I for one, know that this camera has a lot of possibilities and uses.  The lensing is easy, focus is not as much a problem like the 5D, the resulting picture is very film-like and the camera is tiny.  Pitted next to cameras that cost 150 times as much, I would say that picture is maybe 20% off in quality from cameras at the top of the heap.  1700 bucks for a camera that holds it's own.  That's saying a lot. 

Nikon D7000 - uh, no.  It's like the image from a PD-150 but in HD.  Big fail for Nikon.  I can't believe Canon still knocks it out of the park when it comes to still cameras that shoot video.  It's been a few years now guys, come on.  Sony has to get 24p in their Alphas and Nikon has to wake up.  Panasonic is the only one that is biting the heals of Canon with the GH2.

Kodak 5219 and Kodak 5213 - well, this is the hard part.  Considering we all questioned how the film stock was evaluated since even the exposure tests were controversial (was 500ASA stock rated at 800ASA?), it's hard to say exactly what I think of film in general compared to the digital technologies.  Kodak film stock is definitley more sensative in the highlights than in the darks.  Film is grainy and not as sharp as a lot of the digital cameras.  But at the same time, film is varied and diverse.  There is so much stock out there and all of it behaves differently.  I think it was wise to have film as part of the SCCE because it gave us a base to work from when comparing.  Film does have a look or feel that we are trying to achieve with all these other tools, but maybe we've gotten beyond it by now.  I was much more impressed with the Alexa, F35, F3, RED, AF-100 than I was with the shots done with film.  Plus in the back of my mind, I know just how much a high-quality digital video file can be manipulated to look the way I want.  Film is beautiful, but so is digital.

Alight, that's all folks.  Hope my notes help in some small way.  Do comment on this post, I would love your feedback and thoughts especially if you have seen the SCCE video.

Go to http://www.zacuto.com/the-great-camera-shootout-2011 for more information on the SCCE and the Great Camera Shootout 2011.

Snehal Patel
Director/DP
www.FearlessProductions.tv