Ryan E. Walters, Cinematographer

Cinematic Excellence at 24 Frames a Second

Perspective On Part 02 of Zacuto's Revenge Of the Great Camera Shootout 2012

On the 15th of this month Zacuto released part two of their Revenge of the Great Camera Shootout 2012. In this second episode of their three part series they reveal the answers to the blind test that was conducted in part one. This test has raised a lot of discussion, debates, and heated arguments on the internet. As I participated in the event by running Red Epic - a camera with a very passionate following by its users, I thought I'd offer a little perspective on the test.

If you haven't watched Part 01 & Part 02 of the test, I recommend that you do so now. Otherwise the rest of this is kind of pointless ...

First Some Clarification:
In Part 01 of the test, the way my interview is edited makes it sound like I used HDRx in the grade. This is not the case. I shot with HDRx on set purely to cover my bases, and I knew that if I didn't shoot with it on, I would definitely be ripped to shreds on Reduser. So I dialed in HDRx to the appropriate levels in accordance to exposing the image correctly. (Normal exposure + HDRx to protect the highlights). Everything is there in the original file. When I got to post, I took a look at my options in the Baselight (which were the same as what was available in RedCine-X at the time) and opted to NOT use the HDRx track. Even though it is a brilliant way to approach HDR motion imaging, I still have not seen a methodology for combining the two tracks that fits with my own personal taste. Maybe this will change in the future. I'm always open to changing my opinion and tastes. :) But until that happens, I treat the camera as if it doesn't have that function.

I also made sure that we used RedlogFilm and RedColor2. At the time this was the latest color science available, and it yields the best base image to grade from. RedColor3 was released soon after the completion of this test, but I don't think that it would have changed the results from night to day. (The differences would be more nuanced). I also ensured that we were working straight off of the R3D's without any other conversions. On set I shot with the camera set to the highest RedCode it would allow at the time with HDRx enabled, which was RC 7:1, so I knew we were working with the highest quality image available to us.

As I was the last of 4 cameras to be graded that evening, by the time I got my chance to grade in the suite, it was fast approaching midnight. By the time we were done with the grade of my lighting setup, it was time to call it a day. (Call time the next day was at 7 am). That meant I was not around to observe the grade for the empirical portion of the test. So I cannot comment on any choices that were made on the handling of that portion of the footage. What I can say, is that I know Bruce Logan, ASC and the good people at Filmworkers Club Chicago are all knowledgable, experienced, and talented people who wanted to make sure everything was done exceedingly well. 

The Backstory:
In preparation for this test Zacuto approached all of the camera manufacturers and asked them to directly participate or offer recommendations on who should participate in the test. This same offer was made to Red. For reasons known only to the team at Red, they decided not to respond to the requests made by Steve Weiss. Because they did not receive a response from anyone at Red, Zacuto reached out to the Reduser.net community for suggestions on who to run the camera. Unbeknownst to me, I was one of the people recommended.

As I mentioned in my previous post about the Shootout, when I was contacted by Zacuto and asked to run the Epic camera, I accepted knowing that I was opening myself up to some very harsh criticism and judgement by the Reduser community. In fact the judgement on the test began even before anyone had seen one frame from the test, as Red called for a REAL camera test, only to then turn around and cancel it 2 days later. Then, when people actually got a chance to see the test for themselves, projected from a 2k DCP, the community at Reduser unleashed their name calling, criticism and conspiracy theories directly on me. (As if I had set out to ruin the test). Although to be fair, some of the people that I have developed relationships with on Reduser over the years, and some who have seen my work, did end up sticking up for me: here, here, here, and here. (And I'm not the only one who has been harshly criticized. Michael Sutton was flogged by Reduser for his posts here and here. So it really is par for the course.)

In preparation for this test, I consulted with some well respected, and experienced people* about my approach and I even conducted my own high dynamic range test, as I had a feeling that the scene was going to be lit in that manner. This testing, consultation, as well as my many years shooting with the Red One, Red One MX, and now the Red Epic all prepared me for entering an environment where everything would be criticized unless Jim Jannard himself shot with the camera. But I was up for it- how could I turn down an opportunity to work along side and rub shoulders with some of my favorite ASC members?

*Unfortunately I cannot provide names, out of respect for them not wanting to get drawn into the fray. (And I don't blame them ...)

The Results Are In ...
After watching episode 2, I find all of the previous bickering, nay saying, and name calling rather humorous. As you watch the episode, listen to how many people say that the letter C (Epic) is in their top picks. Even Francis Ford Coppola had it in his list, and it was above the Alexa. So everyone on Reduser should be breathing a sigh of relief. After all, there were people who placed C ahead of F (Alexa), and there were people who didn't even have F on their list. I would think this would be good news to the ears of the Reduser community. Unfortunately, this little fact doesn't seem to matter much, as all of the news and talk is about Camera B, the GH2. Should this really be surprising? Of course not- I expect a camera like the Epic to do well. So when it does well, I think, yep it did what it should have done- that is not news. But when a camera like the GH2, which costs almost nothing, does well and better then I expected, that is news. And yet I don't see Canon, Canon Users, Arri, Arri Users, Sony, or Sony Users getting all bent out of shape that a $700 camera supposedly "beat" their camera systems in a subjective test ...

Now that the results are published, I feel that I can finally share my answers without swaying opinions. My top 4 in order of preference were: Alexa (F), Epic (C), FS100 (I), GH2 (B). The next two were a toss up for me: C300 (E), F3 (A). I didn't care for the F65** (H) or the 7D (G), and I really didn't like the iPhone (D). My wife, who is not in the movie industry, chose, in order: GH2 (B), Epic (C), and the C300 (E). Most of the rest of the cameras were all okay to her. She really didn't like the 7D (G) and the iPhone (D). I mention her opinion here because we both had the opportunity to view this test at Skywalker Ranch (Yes, that's us sitting in front of Mr. Coppola), and then again on our 50" plasma that I have in my grading suite.

And that brings up a BIG point for me- viewing environment matters. The 7D and the iPhone looked considerably worse projected on the big screen then when I watched it on my 50" plasma, and they looked even "better" when I watched the video again on my 20" Apple Cinema Display. I haven't watched it yet on my iPhone, but I assume that they will "improve" in quality yet again if I watch it on that device. So where you watch this test, the size of screen, and at what compression rate you see it at will all play into how you perceive the final image. Everything is not as simple as it first appears. Which finally leads me into ...

**I've always felt that Sony cameras have had the hardest time rolling off into the highlights of any brand of camera. And this was no different in the F65. What stood out to me most was the red shirt of the gentlemen by the window. It burned up and had that Sony highlight look to me that the other cameras, and even the other Sony cameras didn't have. I am confident that if different choices were made with the camera it would produce image that I liked better. It is a more then capable camera system.

Some Perspective On The Results:
- Goals Of The Test VS. The Real World
I acknowledge that this is a test, and that due to the fundamental nature of testing it is inherently limited in scope so as to control as many parameters as possible. However, I also think that a fundamental issue is being overlooked. And is that the approach of letting each master cinematographer make the camera look its best is inherently flawed because it does not reflect the real world approach to lighting. In the real world, in my opinion, it is the job of the cinematographer to use the tools he or she has in service to the story. And, in serving the story, the technology and tools should be used to their fullest potential.

This test is a unique experience that I have never encountered before, nor will it be likely that I encounter it again. When I light and use a camera system, it is in service to the story and the project; the goal has never been in making the camera look good. This is a subtle and nuanced difference, but an important one to make. I may choose to completely blow out the highlights of a particular shot to create a dramatic silhouette that underlines an important beat in the scene. That choice is made in service to the story, not to the camera. But in the end, if the image supports the story then the images I've created with camera end up looking "good" even though they may be technically incorrect.***

And this is the struggle of what I faced on set with this test. Cinematographers don't light in a vacuum. Everything around the production influences the choices we make when we create images. And it was hard to fight this natural tendency on this test. When I stepped on set, it felt like the test was all about dynamic range- capturing everything in the scene, when in reality lighting is about more than just capturing the most dynamic range. My natural bent was to really blow out the window in the background, as the translight screamed out "Hey, I'm a translight". So it was a struggle between making the camera look good, and trying to make the image / scene look believable. Darkening down the window with ND gel would have just made it feel more artificial, and really blowing out the window would mean death by public execution on blogs and forums everywhere. But if this had been a set where I was shooting an actual story / scene, I would have pushed it even further with the over exposure to help hide the fact we were on a set and the backdrop was so close to the window. Or, I would have had the set redressed with blinds or curtains to obscure the translight. So while the choices I made on this test were in service to making the camera look good, they are not necessarily the creative choices that I would employ in service to a specific story.

***Another real world example of this is The God Father shot by Gordon Willis. He made some very bold choices with that film, none of which were trying to make the camera look good, but all of which were in service to the story. Yet he ended up creating some beautiful imagery that supported the narrative.

- My Own Great Camera Shoot Out 2012
I think that Steve Weiss and his documentary on the test have hit the nail on the head: we have reached a point where technology can allow anyone who knows what they are doing to use tools at any price range to create compelling imagery. And I think that my reel for 2012 is a great example of that point. In my reel below, I have lit everything in service to the story being told, using the tools I had to their fullest potential. It covers a wide range of camera systems: Alexa, Epic, Red One, Red One MX, C300, and the Nikon D7000. Can you tell which is which? (No cheating if you've seen these clips before, or go looking for them elsewhere on my site).

Now, I am not saying that I would want to shoot a feature film with all of these cameras. Nor would I necessarily want to see all of them projected on a 40' screen. But all of these cameras were well suited for the projects that they were employed for, and they all hold up well at 1080p on the web. The camera system is only one small part of the equation of creating compelling and cinematic imagery.

- The "Real" Cost
These days as every penny is pinched in production, I think it is important to run some basic figures to see what the "real" cost would be for each of the cameras. To simplify this process I am going to make the following broad assumptions:

Accessories & Support Equipment-
The first assumption I am going to make is that all of these cameras carry with them the same accessories and support gear. The one exception to this will be the iPhone, as it did not use a cinema lens. For that camera I'll deduct the rental cost of the lens. So the major variable here will only be the camera body itself. I am basing these numbers off of my local camera rental house, Koerner Camera. And I am basing the lens rental off of the published rate of Angenieux 17mm - 80mm, as they do not have the Fuji lens listed. I also realize that some of these cameras require speciality gear that does impact the cost. I've noted the larger items when appropriate. (Like an external recorder for the FS-100, F3, & Alexa). Yes, these are generalities and assumptions, I do not have a complete gear list for every camera, so I cannot account for every penny. These are just broad strokes here.

Time On Set-
The next assumption that I'll make will be that the time on set was costing $1,000 every ten minutes. This is just a random number I've made up to base comparisons on. I have no clue what it was actually costing Zacuto. The time it takes on set to light for a particular camera is affected by two factors: the limits of the camera and the experience of the cinematographer. I'd be fooling myself If I thought I could light as fast as Rodney Charters, ASC, even if I was using the same camera system. His experience means he will get there more quickly than I would as he has more experience in lighting to draw from. Personal taste is a part of this also as it directly affects the time it takes to light the scene. Some cinematographers like to work with more light sources, while others like to work with less. So I acknowledge that there are more variables at work here than just purely what it takes to light for a specific camera system. While I haven't accounted for experience, or personal preference in the numbers below, it is worth keeping the factors in the back of your mind.

Grip & Electric-
I am going to assume that everyone had the same grip and electric package, and crew. I'll only be adding the cost of the items that are noted in red the technical documentation. (As if each light was a Pay to Play Item.)  Different camera systems, and lighting styles will ultimately affect the crew size and the G&E order, but I'm trying to keep this as simple as possible. I have used the rental prices of one of my local G&E houses, Gear Head Grip to come up with these numbers.

The short run time of the clips used for this shoot is not very meaningful number in my opinion. So instead of using the data from one short clip, I have calculated out the storage space required for 1 hour of storage for each system. This number will then be multiplied by three, as that is the minimum number of backups that should be made for digital acquisition. Again to keep life simple, I am going to use the hard drive cost for G-Raid drives. I am not suggesting that this is the system for how data should be handled, copied, or backed up. (At a cost of $229 per 1TB, the cost per GB is $0.229)

Post Production-
I am going to say that the cost of the color grading session was $1,000 per hour. Again, this is a number I am just making up, but it is not too far from what it costs to grade a project using the latest equipment with a talented colorist. The only catch here is that Sony took the F65 material and processed it themselves before handing it over for the grade, so the post production time for this camera does not reflect how long it actually took.

With these assumptions acknowledged, let's take a look at what the "real" cost is for each camera system based on a one day shoot:

These Are All VERY Rough Numbers - NOT meant for actual budgeting purposes

Apple iPhone - TOTAL COST: $5,212.56
Camera: - $350. (16GB iPhone $200 w/contract - $550 lens rental)
Time On Set: $3,200
Grip & Electric: $855
Storage Space: 33GB (Roughly 11GB / hour, or 180MB / Min)
Storage Cost: $7.56
Post Production: $1,500

Panasonic GH2 - TOTAL COST: $7.167.06
Camera: $40 (Rate taken from Cleveland Camera Rentals)
Time On Set: $5,220
Grip & Electric: $360
Storage Space: 205.5GB (Roughly 68.5GB / hour or 1,140MB / Min @ 150 Mbps)
Storage Cost: $47.06
Post Production: $1,500

Canon 7D - TOTAL COST: $5,290.74
Camera: $160
Time On Set: $3,500
Grip & Electric: $117
Storage Space: 60GB (Roughly 20GB / hour, or 330MB / Min) 
Storage Cost: $13.74
Post Production: $1,500

Sony FS-100 - TOTAL COST: $5,089.27 [$4,792.58 internal recording]
Camera: $550 ($300 for camera + $250 for AJA KiPro)
Time On Set: $2,723
Grip & Electric: $262
Storage Space: 237GB (Roughly 79GB / hour , or 1,320MB / Min @ 176 Mbps for ProRes 422 (HQ))
[Internal Recording: 33GB (Roughly 11GB / hour, or 180MB / Min @ 24 Mbps)]
Storage Cost: $54.27
[Internal Recording: $7.58]
Post Production: $1,500

Canon C300 - TOTAL COST: $10,550.80
Camera: $450
Time On Set: $8,400
Grip & Electric: $185
Storage Space: 69GB (Roughly 23GB / hour , or 375MB / Min @ 50 Mbps)
Storage Cost: $15.80
Post Production: $1,500

Sony F3 - TOTAL COST: $6,100.75 [$5,529.99 internal recording]
Camera: $950 ($450 for camera + $500 for Gemini 444)
Time On Set: $3,630
Grip & Electric: $189
Storage Space: 357GB (Roughly 119GB / hour , or 1980MB / Min @ 264 Mbps for ProRes 4444)
[Internal Recording: 48GB (Roughly 16GB / hour, or 263MB / Min @ 35 Mbps)]
Storage Cost: $81.75
[Internal Recording: $10.99]
Post Production: $1,250

Red Epic - TOTAL COST: $8,037.14
Camera: $1,000
Time On Set: $5,046
Grip & Electric: $171
Storage Space: 1,398GB (Roughly 466GB / hour , or 7.76GB / Min @ 129 MB/s for 5k FF, RC 7:1, w/HDRx)
(Red can record in a wide range of compression ratios, with and without HDRx. All effect the total file size.)
Storage Cost: $320.14
Post Production: $1,500

Arri Alexa - TOTAL COST: $5,993.75 [$7,099 for ArriRAW]
Camera: $1,500 (+ $500 to record Arri RAW via Gemini 444)
Time On Set: $3,368
Grip & Electric: $44
Storage Space: 357GB (Roughly 119GB / hour , or 1980MB / Min @ 264 Mbps for ProRes 4444)
[Arri RAW: 3TB (Roughly 1TB / hour , or 18GB / Min @ 300 MB/s)]
Storage Cost: $81.75
[Arri RAW: $687]
Post Production: $1,000

F65 - TOTAL COST: $3,687
Camera: $2,500
Time On Set: $0
Grip & Electric: $0
Storage Space: 3TB (Roughly 1TB / hour , for 4k RAW)
Storage Cost: $687
Post Production: $500

After looking at the numbers it is interesting to see how closely priced all of the camera systems are in the end. It is worth noting that as the camera systems go up in quality, the G&E cost goes down, but the storage cost goes up. With any camera system there is a balance of cost of the actual camera, the crew it takes to run it, the crew it takes to light for it, and how it affects the cost of post pipeline. All of this needs to be balanced with the needs and budget of the rest of production.

- Reality Check For Indie filmmakers
With the release of the C300, Scarlet, and now the FS700, I get the feeling from what I have read online that a lot of indie filmmakers see these "affordable" higher end cameras as the solution to their problems. Well, it is time for a reality check. The Zacuto Shoot Out, and the opinions of many of those interviewed in the documentary all point towards the fact that there is a plethora of good choices in camera systems today. And there is more to creating a compelling image than just having the latest and greatest camera. (And an unscientific poll taken on No Film School, supports this, as the winner by a land slide was "No Favorite.")

If what you are shooting on the GH2, 7D, or 5D isn't satisfying you, your images are not going to get any better on the C300, Scarlet, or FS700. They are only going to get you more expensive imagery. That flatly lit white wall in the empty office room with no production design isn't going to magically look better now that you are shooting in 4k. And the fact that you can now shoot at 960 frames a second isn't going to save your story if it is weak to begin with - it is just going to take longer to get through. I don't mean this as a discouragement, but rather as an encouragement. Quit worrying about what you are shooting your project on, and start focusing on your craft. As your craft improves, people will take notice, and the money will come, and you will eventually be able to move up to better equipment. But it all begins with  focusing on your craft. Is 4k great? Sure, it's nice. But lets get realistic here. There is a 99.5% chance that your project is only going to end up on the web, or only viewed on small screens. It is not going to be projected in a theater, much less at 4k. So don't sweat the numbers. A powerful story and a smaller screen will hide a number of camera imperfections.

With all things being equal (and they rarely are), on any project I work on, I would much rather have a solid story, with solid production design, solid acting, solid directing, solid crew, and a lesser camera, then a great camera with everything else being weak to nonexistent. So instead of putting your money into a better camera system, put it into the story, what is in front of the lens, and the crew behind the lens - you'll get a much better return on investment in the end.

Wrapping It Up:
I appreciate you hanging in there with me until the end. I know that this has been a rather lengthy write up. But I hope that it has clarified some issues, as well as given you some things to think about and consider as you approach your next project.

Until Next Time - Get Out There And Shoot!
Ryan E. Walters, Cinematographer
See Older Posts...