Last week RED announced that they would start shipping their Epic-W and Weapon cameras – both of which are fitted with the new Helium sensor. As many of you already know, Helium is an 8K Super 35mm sensor that is capable of capturing up to 75fps in full resolution on the RED Weapon.
While I haven’t yet had a chance to shoot with the new Weapon or Epic-W, I did download RED’s sample footage from their website and spent some time experimenting with it in post. At some point in the future, I might do a full review of either the Epic-W or the Weapon, but for now I wanted to leave you with my first impressions after analyzing the sample shots.
I’ve always applauded RED for their ability to push technology forward and deliver cameras that exceed the status quo in many respects. They have been a disruptive company since they first hit the scene with their RED One, and their new Helium sensor certainly furthers that legacy. There’s a lot to love about the new sensor – namely it’s handling of colors – but there also some caveats that filmmakers should be aware of when working with it.
THE SAMPLE FOOTAGE
I loaded the Helium sample shots into RED-CINEX PRO to see how they would play back on my system (a 12 core Mac Pro), and how grade-able they would be. Based on what I had heard previously, I was expecting there to be a lot of shadow detail in the footage, as well as improved color depth overall.
Ultimately I did feel that the Helium sensor offered a noticeable difference in color handling when compared to footage from RED’s previous cameras. The skin tones on both clips felt very natural, and the images were extremely flexible… Especially with the first sample clip, which looked equally great when underexposing and overexposing the image in post.
Surprisingly, even after grading the clips in RED-CINEX PRO, I was able to play them back smoothly at a 1/4 or 1/8 resolution, which I wasn’t expecting since the 8K material is extremely high resolution. 8192 x 3456 to be exact.
As different as the Helium sensor is from most other digital cinema sensors, it does seem to share one common trait (at least based on this test footage) – slightly overexposed footage seems to yield the best results when graded. While both clips were extremely flexible when making color temperature/balancing adjustments, the second clip (which was shot at ISO 800), didn’t perform as well when I bumped up the ISO setting to 2000.
When I pushed the ISO on the second sample shot, it felt like the some color information was lost, while some noise started to appear in the shadow areas. This is completely acceptable, and certainly not a knock against the sensor (as any RAW digital footage would likely behave in the same way)… But it’s simply worth pointing out that this camera, like most other digital cinema cameras, may perform best when slightly exposing to the right.
The first image below is ungraded. The second is bumped up to ISO 2000.
The detail present in the images is also pretty astounding. When you punch in to 100% on a 1080p monitor, you can really appreciate just how much resolution you are working with. The image below is a 1:1 crop of the original shot, exported as a PNG file.
I found it interesting that the shot of the eye looked a bit soft when blown up to 100%. This was likely a lens issue (more than a sensor issue), as I’m not sure how many lenses exist that can truly do justice to an 8K resolution sensor. In reality, this will never matter as most people for at least the next 10 years will be watching content in 4K resolution or below. However, I’m sure in time more and more 8K-capable glass will be released that may allow this sensor to resolve even more detail.
While my first impressions of the footage have been very positive so far, there is one big caveat to consider – the post production workflow. Unless you own a post house, or are running an extremely beefy system with access to a massive RAID/SAN solution, you are going to find that the achilles heel of 8K footage is the data management and overall post-workflow.
I’ll re-iterate that these 8K files were surprisingly smooth to play back at a reduced resolution. That said, playing back footage with real time proxies is one thing, and rendering out footage for a final master is another.
In order to experiment with some different workflow techniques, I converted the first test clip to ProRes 4444 in full 8K resolution. The export ran very slowly, even on my Mac Pro, and it ultimately resulted in a 3.1GB file for the 6 second video clip.
You might be thinking – That doesn’t matter because I can just edit my footage natively with real time proxies… And that is true. You can certainly get away with avoiding transcoding for the majority of your post-production process. But eventually you will have to render out a master file. Whether you are mastering to HD or 8K, it’s going to take a long time to transcode that footage, and depending on what system you’re working on, there’s a chance it might be extremely difficult to export your data master.
I learned this the hard way when working with 4K RED One footage on my film Footsteps in 2011. I edited the film in Premiere Pro using the native .R3D files, and didn’t realize until the last second that I was going to have a very hard time exporting my master. It took me literally dozens of attempts to finally create my data master, and ultimately I had to export the film in small chunks in order to make it work. This was especially frustrating at the time, seeing as I was cutting it close to the submission deadline of a festival that I nearly missed because it took so long to create the file. In 2016, exporting a 4K master is a piece of cake. In 2011, it was not – at least on my system.
Now this may or may not be an issue for you, depending on what kind of system and software you’re running. But it’s certainly worth taking into account that the file sizes, and render time associated with 8K footage are going to be massive, and may cause you some serious workflow hurdles down the line. Not to mention you are going to need a huge amount of storage space, especially if you plan to create multiple back up copies of your RAW source footage.
I truly do believe the Helium sensor is a massive achievement for RED, and I very much respect the fact that they as a company continue to push the technological side of this industry forward.
For me personally, this is not a camera I would invest in (at any price), simply because the workflow at the moment is going to be cumbersome, and I have absolutely no need for 8K footage. That’s not to say that this camera isn’t going to be very valuable for certain types of productions (large scale features, green screen work, visual effects plates, etc.). However, for an independent filmmaker that needs to find the right balance between image quality and efficiency, cameras like the Arri Amira are still going to take the cake.