In this corner, weighing 11 lbs. and a height of 7.5 inches and boasting 5K imagery, The Red Epic!!!!!
In the other corner, weighing 6 lbs. and a height of 7 inches and displaying 4K capture, The Canon C500!!!!
When we started prep on Need for Speed, the director Scotty Waugh wanted me to do an extensive camera test of all the possible players on the market so that we could find the visual voice of Need for Speed. This was a daunting task to pull off. Zacuto had done the Great Camera shootout and I loved the in-depth tests that the experienced cinematographers had put them through. When you are going for a specific look, these tests need to be done internally so that you can personally see, with the specific way that you expose, which camera sensor responds the best.
I compare digital camera sensors to film emulsions. The reason a cinematographer selects a specific camera for their film is the look and feel of the sensor and their lenses.
You are about to embark on a journey of my discovery that is as unbiased as I can be. I will describe the subtle details in narration so that it is a much more immersive experience for all of you. I know that these images are not in 5K and 4K for your review. We are limited by the internet but I think the details that I am pointing out still shine in HD. These extensive tests between the Red Epic and the Canon C500 were conducted over three days with seven other cameras. We will be keeping up the comparison tests once a month.
The first test is our ISO range in the daytime. I like to do these both in the day and night. With the daytime ISO test, you can really see the noise that would not be as apparent during the night because it can be masked in the blacks. Look at the grey card, the shadow in the trees behind them and the shadow areas on their faces for the digital noise as we crank up the ISOs on each camera. Also look at the color chart to see how it holds its color as you increase your ISO levels and the contrast level increase on their faces.
I found that as we increase the Epic’s ISO, it started to lose color depth and had a very magenta cast. One thing that has always been apparent with Red footage is its magenta, pink nature. The C500 did not exhibit any of these qualities. Once the Epic got up to 3000 ISO, the image looks very thin. It loses a lot of its color information. You can see it in the background as well with the greens; everything looks very de-saturated and dead.
We noticed that the bit depth and its color on the C500 seemed really thin once we got over 4000 ISO, which is still pretty impressive. We made this our top end, if we had to go there. But the highest we really wanted to go on the C500 with night exteriors and night interior cars would be at 1600 to 2000 ISO. All the lights on the street seemed to respond well to these levels and they felt balanced.
This test is to find the exact level of Neutral Density where a camera starts to become polluted with IR (Infrared). When you use high levels of ND to take your exposure down to an f stop that will deliver more cinematic depth of field, the sensor pics up on IR and that IR contaminates your image with a reddish brown hue. That is very difficult to get out of your image even with the most powerful of color correction devices.
Each sensor has an IR filter, but finding that IR filter’s breaking point is what this test is about. None of the camera manufacturers will tell you at what level their IR filter is so I compare this test to Indiana Jones’s expedition to find the Holy Grail.
We systematically did this test with Tiffen WW Straight ND filtration and we quickly realized the limits of the RED Epic’s IR Filtration. We wanted to see where it started to happen. After a .9 ND, you could see the IR pollution start to enter the blacks. What you are looking for is this reddish brown contamination. Look to the trees behind our model, over each shoulder.
On the flip side, Canon has really mastered this IR filter technology with the Still Photography side of their company. So even with eight stops of ND added, there is no IR pollution evident in any of the blacks.
I love this test because it is all about finding your digital sensor’s voice. I broke it into two parts, one being Over Exposing and the other Under Exposing.
We start with the Epic with the key light matching the exposure on the camera. If the light was at a 4.0, then the camera’s exposure was set to a 4.0. We slowly open up the lens in third stop increments. The reason for this test is to see how the camera handles over exposure and whether it looks like video when it blows out or if it looks like film. The Epic started to clip at about +2 and 1/3 stops. It seemed to react completely differently than the C500 in its over-exposure. The Red Epic felt very hot overall and the C500 was just clipping the male model’s forehead. So it feels like the C500 held its highlights about +2/3 of a stop more. However, the Epic overexposed in a more filmic way, softer than the C500, which looked more like video clipping.
With the C500, this overexposing did not look organic or feel like film. At +2 and 1/3 stops over exposed, the faces of our models started to clip in a very digital way. It was not creamy. I felt this was one of the biggest limitations of this camera. You can see the model’s forehead clip at +21/3 stops and continue to feel very video looking all the way up to +4 stops.
Let’s conclude on this part of the test. The RED Epic totes 15 stops of latitude and the C500 totes 12. We are about to witness where the RED holds its latitude in the under and it is impressive.
I do the underexposing part of this test because I feel that digital sensors look better when they are starved of light. This is a technique I used with film. I under exposed most of my Kodak stocks -2/3 of a stop. When you start underexposing the image, it feels more filmic, to my taste. This is where the Epic cleaned up. Its underexposing was so amazing. I did something a little different in these tests when we color corrected them. The under exposing was so impressive with the Epic that after -2 and 1/3 stops, I started to bring back the image to look normal and to see how far I could push it. We did not do this with the C500. When you see the side by sides, the C500 will continue to get darker but the Epic stays the same. The incredible thing about this was when I tried to push the C500 to do the same, it fell apart and the digital noise became very apparent. With the Epic, even at -5 stops down, no digital noise was injected and the color remained true to form.
With the C500, I found -2/3 to -1 stops under exposed was the right recipe, but I felt with the Epic, the underexposure recipe was more like -3 stops.
This is a test I like to do because it gives you a sense of what fill levels your camera sensor responds to on a face. I lit the model more with a side light so that we could see the fill level better. I started at -2 stops down from the key light on the fill side, which is the camera left side of her face. Then I took the level down in 1/2 stop increments so that you can see where you would keep your fill if it were a woman and/or a man. I did this using a Sekonic light meter so that it was accurate.
I wanted to see how the Red Epic handled this warm light. I have had problems in the past with using very warm sources. Our model’s key light was 2700 Kelvin and the fill light was 3600 Kelvin.
I felt that both cameras delivered the test well. The Red Epic had a more magenta feel overall. What I loved about the Epic was the way it fell off into the shadow side of her face. It was gradual and elegant. The Epic delivered a creamier skin tone, but on the pink side. The C500 was nice and golden, but fell off more drastically in the shadow areas. The C500 image seemed sharper overall, a little more snap. The background color was more purple than blue on the Epic. The C500 rendered the cool tone in the background, pretty much spot on to my eye.
I knew that we would be doing a ton of shots inside the cars with our actors physically driving. I wanted to make sure that by just adding a little dashboard fill light, I could literally drive down our race route and not have to add a light. We tested the Red Epic in this configuration and right off the bat, the Epic felt noisy. It felt muddy and it felt like I was not able to capture the minimal style of lighting that I wanted to go with in our driving scenes. It was not energizing the light in the background or the light coming into the car. When the C500 came on the screen, it seemed to see all the colors in a much brighter and more vibrant way. It energized the background, as well as the active light in the car. It felt so real, seeing its display of colors out the window as well as traffic lights spilling red into the car. When the traffic light went green, you could see that all play out on our driver’s face.
Director Scott Waugh loved this because he knew the way to have the audience experience speed at night is to actually see something moving by in the background on the sides of your frame. I knew that our schedule and budget was incredibly tight, which meant we needed to take advantage of as much lighting as the city of Macon, Georgia could provide us. We made the C500 our go to night camera based on these tests. It gave Scott the freedom to have our actors drive without any lights out on the hood that would block their vision and enable us to link our actors with other actors driving as well.
Because we did not have lights all around the car, I could have my camera operators looking out the front windshield and pan right back to our actors driving. These are shots that are usually done during the day because you lack this ability to pan because of the lights. We were able to light over three miles of night street racing with about 12 lights. YES!!! It is possible. That was so great for our stunt players because they could really go at it and get the cars cranking through turns and not worry about running out of lighting real estate. This style of shooting action sequences as a play, like we did on Act of Valor of the SEALs hitting a target, was the right recipe to put the audience in the driver seat at speeds over 180mph.
What I really took away from this test was that it is absolutely essential to use the right camera sensor for what it does best. The Red Epic did well with under exposing, but then fell short in the night driving scene where underexposure is everywhere. It had a very hard time with IR pollution and delivered across the board a more magenta image. The C500 delivered great skin tones, a very clean image with ND filtration without IR pollution and owned the night, but fell short in the over exposure area. More tests are on the way. Next month the BMCC goes head to head with the Sony F5. Stay tuned.
How to test your camera’s latitude. This is one of the most important tests that you can perform because...
When Canon approached me to test and shoot a short film for their new 1DC 4K DSLR, I was...
The Canon 5D Mark II opened the floodgates to new and creative filmmaking three years ago and is still...
When Shane told me he wanted to shoot behind the scenes footage of “The Last 3 Minutes” I was...
Speed and production value are intertwined. Production budgets are shrinking and companies are asking for more from cinematographers and...
I treat each new digital camera that comes out as a new film emulsion, one that needs to be...
Find the GoPro Hero 3 shots in the new Need for Speed SuperBowl Trailer and enter to win.
Waiting for Lightning, a documentary about pro skateboarding pioneer Danny Way, began production in 2008. Over the course of...