Over the past few days, I’ve been circulating some of the first footage shot with the RED Scarlet-W. There really aren’t many of these out in the wild, and as far as I know; I was one of the first to receive the camera (brain only) after switching my pre-order from the Red Raven to the Scarlet-W. As most would expect, the image is very similar to the other Red cameras, the Epic and Scarlet. Red’s Dragon sensor remains the same, so in theory the image should stay the same, right?
Red’s latest generation of camera bodies, which Red refers to as DSMC² cameras, include the Weapon, Scarlet-W, and Raven. Retired is the surname “Dragon” as all the new DSMC2 cameras from Red now employ the Dragon sensor as opposed to their older Mysterium-X (MX) sensor. From this point forward, saying Weapon Dragon or Scarlet-W Dragon is simply redundant, though it will be hard to drop as the industry has been stuck on “Red Dragon” for a couple years now. You’d be surprised how many calls I hear of people asking for a “Red Dragon camera,” not realizing that “Red Dragon” can either mean the Epic Dragon or Scarlet Dragon… but that’s a whole other topic.
DSMC² Features and Improvements
According to Jarred Land, CEO of Red, there are 7 main highlights that the DSMC2 system brings to the new cameras. This will just be a quick overview, but I’ll be going more in detail discussing each of these features as they come up.
- Internal and simultaneous ProRes recording (with DNxHR/DNxHD coming soon)
- Lower compression ratios for REDCODE (Red’s RAW recording system)
- Auto Sensor Calibration (Automatic black shading)
- 3D LUT support (32x32x32 for the Weapon, 17x17x17 for the Scarlet-W, but NO LUT support for the Raven)
- Lighter weight and new body layout
- Built in WiFi
- 3rd Party Module Ecosystem
Ultimately, it’s up to you whether or not these improvements will be worth the upgrade for you and if these are features that you find valuable. I’ll be going into some of these features more in-depth as I discuss my shooting experience and what I found different in shooting with the Scarlet-W.
Compared to the original Scarlet Dragon, there are certainly noticeable improvements to image quality. With the new brain comes faster processing power allowing for lower compression. But it’s not just lower compression that gives a better image, but the collective whole of the better components and body design that allows for more efficient processing. The main thing I noticed when shooting with the Scarlet-W (after shooting with the Scarlet Dragon several times) is that it’s much less noisy than it’s DSMC1 counterpart. It’s hard to say what in the image has improved without directly compared it to it’s precursor so I will be doing comparison videos in the future between the original Scarlet Dragon and the Scarlet-W. But for now, it’s simply a matter of a much cleaner image, which trickles down to post work and gives you more room to play; you have more freedom to push the R3D files before falling apart. It’s that latitude and such large amounts of data in a small container that truly makes REDCODE such a wonderful system to work with.
The Scarlet-W simply felt like the lightest, most compact camera rig I’ve ever used. I spent the day chasing a little puppy around and never once felt like it was too heavy. It really reminded me back when I used to shoot with DSLRs, I felt like I could put the camera anywhere, swing it around, and really put it in any position I wanted. I couldn’t get over how small it was, it’s really impressive once you get your hands on it.
The build was fairly simply and contained all the necessary components to get it running:
- Scarlet-W Body
- Lemo Adapter A
- Red 5″ touch
- Red Run/Stop top handle
- Red Brick
- Wooden Camera Quickback 2.0
- IDX V-mount plate
- Leica R cine-modded lenses
- Red 120 GB Mini-Mag
- Bright Tangerine Misfit clip-on mattebox with Schneider IRND 1.5 for the beach shots
I’ve since added a Wooden Camera Easy Riser and Easy Top plate (with Pogo-Lemo ports to eliminate the Lemo Adapter) and I’ll be highlighting different build configurations in the coming weeks. For my first time using the camera though, I was very happy with my build and could see it being used on real productions with that exact configuration. I didn’t choose to add a rail mount system to add a follow focus as I just pulled off the barrel, but I could easily see one fitting nice and neatly.
The Shooting and Footage
The whole point of this shoot was to really see how the camera performed in less than ideal conditions. I wanted to push the sensor and see how well it would handle the extreme contrast of daytime conditions at the beach. For the most part, it really exceeded my expectations; there was an enormous amount of information and detail I was able to recover in the Red R3D files.
I chose to shoot with a multitude of framerates and compressions to see how they performed and cut together. I would be shooting 5K for one shot, a few seconds later I would drop to 2K 240fps, back to 4K, and back to 2K. It was all over the place and really dictated by what Oliver was doing. Constantly switching around resolutions and framerates was quite fun and I think it was assisted by the auto black shade feature of the DSMC2 system. Previously, you would want to black shade whenever you wanted to change framerates, but with the new DSMC2 bodies it’s built-in to a range of calibrated presets that are changed automatically.
The settings on the camera were:
- ISO 800 (native)
- 180° shutter angle
- 23.976 project base
- Various framerates, resolutions, and compression ratios
- 5K – 23.98 fps – 4:1, 50 fps – 9:1
- 4K 120 fps – 13:1
- 2K 240 fps – 8:1
ALL the beach footage was shot with a Schneider Platinum IRND 1.5 with a Leica R 35mm. The Schneider ND gives a distinct blue/green tint to the image. I would ride the iris to set exposure, though obviously it may not be perfect since I’m chasing around a little puppy 🙂
All clips were color corrected and graded with DaVinci Resolve. No noise reduction or post sharpening was done.
To someone who has never worked with RED’s massive resolutions and R3D files, the post workflow for RED cameras might seem very daunting and time consuming. I personally find it very easy once you understand how to transcode your footage and edit using proxies; it’s really the only way I can get around to editing large 5K files without spending thousands on a powerful system. My PC is significantly outdated — it’s over 5 years old and I’m still running an i7-2600K processor but that didn’t stop me one bit in editing 5K footage.
I transcoded, edited, and colored all the footage in DaVinci Resolve. Redcine-X didn’t seem to work, it would constantly crash on my Windows based PC but that was fine, I’m much more comfortable with Resolve.
My workflow with all the 5K, 4K, and 2K footage is simple:
- Transcode all the R3D’s to 1080p or 720p Quicktime Proxies (took me 40 minutes to transcode 100 GB of R3D files)
- Edit using the 1080p or 720p proxies
- Replace the proxies with the original, full resolution R3D files
- Export for delivery
It’s that simple.
I have made available some RAW R3D files for download:
4K 120 FPS shot of Oliver in the car: http://www.mediafire.com/download/8zo13uhcmzida3q/A001_C036_0312WQ_001.R3D
5K beach shot of Oliver sitting there: http://www.mediafire.com/download/lez4b4qog9abp3z/A001_C057_03122L_001.R3D
Click for full resolution. All clips downscaled to 1920×1080.
I think the most impressive clip that really shows of the dynamic range of the Dragon sensor in the Scarlet-W is the middle before and after with Oliver in the car. He was underexposed inside car and he was back light by the half open car window of the outside, which was overxposed. I was able to recover a significant amount of shadow detail and bring him up without the use of any power windows or anything else. There’s very minimal impact to noise even bringing up the exposure in Resolve.
The last clip of the video was a mistake with grading on my part. I went too heavy on the contrast and shifted the highlights to an awkward pink. It was 2 AM at that point and I wanted to sleep and gave it a quick 1 minute coloring job. In retrospect, I should’ve spent more time to fix it, but I included it to show that there was detail in those blown out highlights after I went up with the contrast.
Slow-motion and 2K 240 fps – Is it usable?
Obviously a significant amount of the video was in slow motion, it was the primary feature I wanted to test in this video. There was a mix of 5K 50 fps, 4k 120 fps, and 2K 240 fps (at 8:1 compression). To be honest, I could not tell the difference between the 5K 50 fps and the 4K 120 fps. Even though the 4K 120 fps was 13:1 compression, it was hard to distinguish which was which — it looked that good.
It’s very clear which are the 2K 240 fps shots. They are noisy, soft, and have dancing artifacts (1:05 in the video). The question is, is it usable?
To me, 2K is not a viable resolution to shoot when you’re shooting in RAW. The reason for this is because at 2K RAW, you’re cropping the sensor to such a small size that the noise and grain is amplified significantly. At 2K, you’re effectively shooting smaller than Super 16, you really can’t expect sharp and amazing performance at such a small sensor size. It didn’t come to a surprise to me, nor should it come to a surprise to you if you decide to shoot with such little sensor size. I guess one thing that was surprising was how the noise and artifacts would flicker, which was not expected.
Phil Holland creates a fantastic sensor size comparison graphic. You can see that 2K is really a fraction of the Scarlet-W’s 5K sensor.
However, if shooting ProRes, 2K is most certainly usable. Shooting ProRes downscales the 5K sensor to the ProRes resolution you are shooting, as opposed to how it crops the sensor when shooting RAW.
Photo courtesy of Red’s Crop Factor tool
Photo courtesy of Phil Holland from the RedUser forum
If you got this far, thank you for taking the time to read this long post. I plan on writing blog posts more frequently with more videos, comparisons, and all that fun stuff that I love talking about. This is just an outlet for me to share my thoughts on gear, and as a gearhead it’s nice to have that outlet when no one else will listen.
Stay posted and please feel free to subscribe to my YouTube channel where I’ll be posting my content: https://www.youtube.com/channel/UCPKnsz2m0uI9JQJJXBKib3w