Head to head: Arri Alexa, Panasonic EVA1, RED Dragon, Canon C300 mark II, Blackmagic Ursa Mini and Sony FS5
Intro
There is a big chance you have a preference in terms of the "look" of a certain camera system as a camera operator, DOP or director. Some cameras and brands like Arri, RED and Canon are known for their particular characteristics.
Most of these results are based upon the untouched footage that comes straight out of the camera. Tests rarely get graded by a dedicated colourist. We wanted to find out how different professional cameras from a wide range of price points compare in a test setup.
The test setup
We decided to light it quite flat with a big diffuser. We included a bare lightbulb in the background. While this isn't as scientific as a dynamic range chart it still gives a rough idea of the dynamic range of the cameras. To check the noise characteristics in shadow areas we put a black frame in the bottom right corner.
We used the following equipment:
Cameras
Arri Alexa Mini - 3.2k Prores 4444
Red Dragon - 6k R3D 8:1
Panasonic EVA1 - 4k 422 400mbps
Blackmagic URSA Mini Pro - 4.6k Prores 444
Canon C300 mark II - 4k 422 320mbps
Sony FS5 - UHD 420 100mbps
Lens, filter and mattebox
Arri MMB-2
Sigma 35mm f1.4 Art (EF mount)
Formatt Firecrest ND 0.3
Formatt Firecrest ND 0.6
Lighting and chart
Arri Skypanel S30 (2x)
Arri Skypanel S60
Chimera 72x72" Pro Panel
X-Rite ColorChecker Classic
Grading
DaVinci Resolve 15 Studio
Dolby Professional Reference Monitor PRM-4220
Laser projector
We used the SmallHD to check focus - Jesper Wahlqvist
Shoot
We shot tests at the highest possible internal quality (4k on most cameras) at the native ISO. We then went on and shot at iso 3200 (or equivalent on the FS5, max ISO on the Ursa was 1600). Finally, we shot some slow-motion footage to check out resolution and artefacts.
We were aware that the EVA1 and FS5 can get a significant increase in quality when recorded externally, but we decided we wanted to compare the practical internal capabilities of the systems instead of relying too much on external recorders. This would be a test on its own since there are a lot of variables in terms of recording formats and recorders (click here to see our FS5 ProRes RAW test).
The three Skypanels combined gave a good output for the tests - Hugo Nilsson
We also are aware of the fantastic RAW capabilities of the Ursa and Alexa, but the lack of good compression makes it clunky and difficult to data wrangle on anything longer than a few hours without a dedicated DIT. Since we wanted to compare the cameras as they usually get used in productions we ended up only shooting RAW on the RED.
The final setup - Minco van der Weide
One of the most important factors of professional cameras is ease of use. Cameras in higher price-ranges are mainly rentals, so usually the less quirky the better. We found the Alexa to be the easiest to use, but this opinion might be slightly biassed since we are used to the Arri system.
The EVA1, Canon, RED and Ursa are still all quite user-friendly in terms of buttons and menus, and while every camera has its quirks it seemed to be all totally doable. One camera that stood out was the Canon. The C300 mark II becomes incredibly tall when rigging up.
The most tricky camera to use and mainly expose was the FS5. Sony menus have a bad reputation and this holds up in this, and previous tests we did. It is, however, something that you get used to over time, but the Canon and EVA1 are definitely easier to use for a first-time operator.
Grading
To get a good sense of what these cameras are capable of we decided to get our footage graded at Tint, a leading postproduction company in Stockholm. It has fantastic facilities and the industry standard Dolby PRM-4220 Reference Monitor.
Our footage got graded by the talented Oskar Larsson (i.a. known for his work on Lexus, H&M, Volvo and Dior commercials). Post-production supervisor Vincent Larsson (i.a. known for The Square) was also present during the grading.
The look and feel of cameras tend to be quite a personal and subjective matter. However, the overall conclusion of this test was that all the tested cameras can perform surprisingly similar. When used correctly all the tested cameras were easy to match and ended up looking completely comparable to an untrained eye.
Grading the footage at Tint - Oskar Larsson & Vincent Larsson
As expected the Alexa and RED still had the edge in terms of dynamic range, the rest came however very close. We were pleasantly surprised by how well the EVA1 matched with the Alexa, it had great colour and dynamic range.
The FS5 seemed to lack in the colour department and overall depth of the image. The dynamic range was also less then optimal, but this also has to do with the fact that it's recommended to overexpose by 2 stops when shooting S-LOG2. We still were surprised by how well the 8-bit 4:2:0 footage compared to the rest, even on the reference monitor.
The area where you saw quite a difference was the vector-scope when zooming in on the colour chart. When we compared the 8-bit 4:2:0 footage out of the FS5 with the 10-bit 4:4:4:4 footage out of the Alexa we saw a lot more spikes on the Sony side.
The wider spread on the Alexa shows more colour variances and less patchy colour tones. This difference was however minimal between the other cameras since the Sony was the only one limited to 8-bit recording in 4K.
Comparison
With decent grading skills, all the tested cameras can perform adequately for most well-lit work. The slow motion and lowlight capabilities are where the differences became more distinct. We do however recommend to check these differences out by downloading the footage since compression would remove most of the noise when uploading to YouTube.
If we had to rank the cameras in terms of overall quality of footage in a well-lit scenario we would end up with the following order. Please keep in mind that most of this is open for interpretation and preference, so we encourage you to download the high-resolution files below. By testing it you will hopefully come up with your own conclusion.
Overall image quality
Arri Alexa Mini / RED Dragon 6k
Blackmagic Ursa Mini Pro
Panasonic EVA1
Canon C300 mark II
Sony FS5
The Arri Alexa Mini and RED both have distinct characters in terms of image quality. Even though the Alexa tends to be used more often in high-end productions, there is no denial about the fantastic image quality coming out of the RED. After showing it to many industry people we ended up with a tie in the first place. On a professional set with good light you can't really go wrong with either of these two cameras.
However, when shooting in low-light the order changes dramatically. The overall usability of the cameras around 3200ISO would be the following.
Lowlight performance
Panasonic EVA1
Canon C300 mark II
Sony FS5
Arri Alexa Mini
RED Dragon 6k
Blackmagic Ursa Mini Pro
There are however big differences in terms of internal noise-reduction on most cameras, especially when shooting RAW, so this is not a totally fair test without applying noise-reduction in post-production. But since the amount of noise reduction in post would be totally up to preference, we ended up comparing the noise coming directly out of the cameras.
Slow-motion was an area where the cameras differed most. It's difficult to list them since they all differ in crop factors, resolutions, codecs and frame-rates when shooting slow-motion. Adding it all up we would end up however with the following order.
Slow motion performance
RED Dragon 6k
Arri Alexa
Panasonic EVA1
Sony FS5
Canon C300 mark II
Blackmagic Ursa Mini Pro
Often overlooked is the media storage part of shooting with pro cameras. When file-size increases the on-set and postproduction workflow can change drastically.
While it might sound like a great idea to shoot your production in RAW, the backup and media transfer process can quickly become so confusing that the extra quality doesn't overweigh the extra costs and time RAW takes to handle.
File-size to quality ratio
Panasonic EVA1
Canon C300 mark II
Sony FS5
RED Dragon 6k
Arri Alexa & Blackmagic Ursa Mini Pro
The EVA1 delivers fantastic 10bit 4:2:2 files in 150mbps and 400mbps. The image is downscaled from 5.7k to 4k resulting in sharp but still organic looking images.
The C300 mark II delivers files very similar in gradability and sizes. However, apart from the magenta colour cast, the visual sharpness seems to be less. This obviously has to do with the lack of a downsample.
Sony cameras are generally seen as workhorses in the broadcast and corporate sector. It is however not possible to shoot 10bit 4:2:2 with the FS5 in 4k, and S-LOG is generally considered as tricky to expose and grade optimally.
RED was hard to place on the list of file-size to quality ratio since it's practically the only camera that is mainly used in RAW mode on real-world productions. The amount of compression that's optimal depends a lot on the shoot. Resolution, movement and colour all play a role. However, 8:1 is generally seen as a good balance between quality and file-size. You end up with fantastic flexibility in post, but file-sizes are still a lot bigger than the cameras above. Since the difference will be hard to notice in a well-lit scenario the RED ended up on place four.
The Alexa and Ursa ended up sharing the last place in terms of file-sizes. Their ProRes codecs are edit-friendly and widely regarded as the industry standard codec. However, the compression is inefficient and file-sizes tend to be rather large without much improvement in visual quality. This lack of compression does mean that playback in most NLE's is very smooth.
It will be interesting to do a test once ProRes RAW gets supported on the Alexa and Davinci Resolve, but this probably will take over a year since Atomos has exclusive rights on the codec at the time of writing.
Verdict
We are living in an age where cameras aren't the limit to creativity anymore. Not that many years ago it would be impossible to argue that a €4.000,- camera could be directly compared to a €40.000,- camera. Nowadays the differences become less and less noticeable.
We'd argue that most professional cinema cameras can better be compared in terms of ease of use on set, rather than just the output.
A special thanks to Falsett and Tint for making this test possible!
Download
To see the complete differences in quality you can download the high-resolution files by clicking here.
You are free to use, modify and publish it under a Creative Commons Attribution licence, make sure to mention Filmkompaniet and provide a link to our website (www.filmkompaniet.com).
We would love to hear your thoughts, share your results in the comments below.
Disclaimer
No camera manufacturers were involved in this test, the verdict is our honest and unbiased opinion. We at Filmkompaniet own an EVA1, but since the post-production is done by an independent post house it did not affect any of the conclusions. We saw it as a great learning experience and hope this test is helpful for people interested in any of the tested cameras.