Ever since I took ownership of my C500mkII I’ve been asked over and over how it stacks up against the Mini LF. I do have an opinion on that, but I decided it’d be a better use of everyone’s time to go ahead and compare all four of the current Full Frame Cinema Cameras available with my friends over at Stray Angel. Alexa Mini LF vs C500mkII vs Sony Venice vs Sony FX9. All at once. It was also “take your girlfriend to work day” so Dani got to sit in as our model. You can see that test footage below:
We shot each camera with the Sigma Cine Prime 50mm and 85mm lenses set to f5.6 and lit the scene in such a way that one could get a good idea of how each camera handles their respective dynamic ranges from F19 down to F2, when aided by an exposure ramp up to f16 and down to f2 on the lens. We had the cameras set to 4300K as we were using mixed lighting. While the look is important, and we’ll get to that, the first thing we wanted to focus on were the features as that’s what you’ll be considering most heavily when deciding to buy or rent. These cameras are actually quite similar in that regard so to save everyone some time, I’ve made a chart that goes through the main specs of each.
| MINI LF | C500mkII | FX9 | VENICE | |
| Sens. Size | 44.7mm ∅ | 43.1mm ∅ | 40.4mm ∅ | 43.5mm ∅ |
| Max Res. | 4448 x 3096 (4.5K OG) | 5952 x 3140 (5.9K DCI) | 3840 x 2160 (4K UHD) | 6048 x 4032 (6K 3:2) |
| Raw | 12 bit | 12 bit | N/A | 16 bit* |
| Compressed | 12 bit | 10 bit | 10 bit | 12 bit |
| Sensitivity | 800 ISO | 800 ISO | 800/4000 ISO | 500/2500 ISO |
| Codecs | Arriraw, ProRes | Rawlite, XF-AVC | X-AVC, MPEG2 | X-OCN, X-AVC, ProRes |
| Proxies | N/A | 2K, XF-AVC, 35Mbps | HD, X-AVC, 9 Mbps | HD, ProRes 422 |
| ND Filters | 6 Stops, 3 Stages | 10 Stops, 5 Stages | 7 Stops, Variable | 8 Stops, 8 Stages |
| HDR | HLG/PQ | HLG/PQ | N/A | HLG |
| Anamorphic | 1.3x & 2.0x | 1.3x & 2.0x | N/A | 1.3x & 2.0x |
| Reported DR | 14+ Stops | 15+ Stops | 15+ Stops | 15+ Stops |
| Autofocus | N/A | DPAF, Face | Face | N/A |
| Media | Codex Cards | 2 CFexpress, SD | 2 XQD, SD | 2 SXS, AXS |
| Video IO | 2 SDI | SDI, Mon, HDMI | 2 SDI, HDMI | 4 SDI, Mon, HDMI |
| Gen/TC | Timecode | Timecode | Genlock, Timecode | Genlock, Timecode |
| Audio IO | 6-Pin LEMO | 2 XLR, Mic | 2 XLR, Mic | 1 Five-Pin XLR |
| Stabilization | N/A | EIS | Metadata-based | N/A |
| Mount | LPL, PL | EF, PL | E | E, PL |
| SOOC “Look” | Balanced Contrast | Low Contrast | High Contrast | Balanced Contrast |
| Accurate Skintones | Accurate Skintones | Green All Over | Olive Skintones | |
| Green Mids, Color Acc. | Magenta Cast, Bright | Darker than average | Balanced Color, Greenish | |
| Strengths | In-Camera Raw | In-Camera Raw | Variable ND | Rialto Extension |
| Arri Look | Most ND, Smallest | Great AF | Most Versatility | |
| Industry Standard | Best Price/Performance | Cheapest | Robust Image | |
| Weight | 5.7 lbs | 3.8 lbs | 4.4 lbs | 8.6 lbs |
| Rental | $2500/Day | $695/Day | $495/Day | $1150/Day |
There can be a bit of confusion around Full Frame cinematography, but after reading this article from Art Adams and the stuff Steve Yedlin has put out you should be in good shape. Essentially, there are absolutely advantages to shooting on bigger sensors (higher physical resolution, less noise, better in low light), but there’s no inherent “look” to full frame/large format cameras that can’t generally be replicated on smaller sensors or in cropped modes. All four of these cameras are “Netflix Certified”, they’ve all got built-in NDs, they’ve got roughly the same dynamic range, the same basic I/O, they’re even roughly similar sizes. These cameras are the best of the best right now, and as time goes on cameras will only become more and more similarly powerful. This isn’t 2008 or even 2012 anymore, that’s for sure.
Depending on your level of production, you may or may not need certain things. For some people, the FX9’s autofocus and variable ND might be a big selling point. You might want the price/performance ratio or the small form factor of the C500mkII. Maybe you need all 4 SDIs on the VENICE and want to lean on that 16bit raw. Maybe you’ve worked with ARRI for the past 10 years and don’t want to change workflows. It’s entirely up to you in regards to what’s most important.
It’s likely I don’t know you, so I’ll just go through and share my experiences with each camera, what stood out to me, and hopefully that’ll help you make a decision.
As I said, I own this camera so I’ll start here (sorry about the missed focus up there).
Overall I think the C500 is the best “bang for the buck” out of the four. Possibly on the market today overall. It’s their toppest-tier sensor (the C700FF) squeezed into a C200-esque body for half the price. I’m not a Canon fanboy by any means, but they’ve done a fantastic job here seemingly addressing the many criticisms lobbed their way over the past 10 years.
To start, it’s got 6K 12bit Rawlite internally. It’s not full “raw” as it were, but it’s still incredibly impressive and the data savings are welcome. It’s like extremely malleable compressed footage except you can choose your color space/gamma/wb/etc after the fact. The colors in Rawlite compared to the XF-AVC are definitely more defined but even compressed it looks good. It’s also got an easily swappable PL mount that uses Cooke /i metadata.
When shooting Rawlite you can also generate really beautiful 2K proxies that are saved to the SD card. They’re surprisingly robust and I actually found that if you pull the CFexpress card, the C500 will still record the proxies so, if you really wanted to, you could just use those. If you were shooting something for web, or you needed a TON of footage, or using SD cards made sense for workflow reasons, this could be a great option. They’re 4:2:0 8bit but it’s 35Mpbs so it already beats out my C100mkII in data rate, and they’re just proxies! A 6K image supersampled to 2K ain’t half bad.
If I had to guess, I’d assume that most people are going to use the 10bit supersampled (full-readout) 4K in XF-AVC. It does have less color depth but you’d be hard pressed to tell just by looking at it. You’ll also have to shoot compressed if you want to use the EIS, but the obvious advantage is the extra room on your memory cards.
On the body you’ve got 12G SDI, 10 Stops of ND, two XLRs, a 12v XLR power port, two CFexpress cards and one utility SD slot, a host of programmable buttons, and a place to put your lens. It’s got a lot in its tool kit that allows you to be incredibly flexible in tons of situations. I’d say the C500mkII is best suited for doc or indie film work with the option to be run as a b-cam to the Alexa or even the VENICE, surprisingly.
Arri Alexa Mini LF
It’s got the Arri magic, it’s part of one of the most widly-used workflows around, and it’s in a lil’ carbon fiber Mini body. The only difference is the sensor is bigger but it’s the same ALEV3 sauce they put in the original Mini, the SXT, and the 65. It was also the most robust image to work with in Resolve, with less noise than the competitors in the extremes and held its color very well. It’s got the look we’re all used to, which many of us love, and if you’ve got the budget it’s likely your first choice. Without question, Arri is king of the cinematic mountain.
That being said, since it’s the Mini it’s missing a lot of “full body” features. The Minis are generally intended to be used as B-Cameras to the full sized Alexas but let’s be real, most people are renting Alexa Minis not Amiras (I love the Amira). Here, the Mini LF gives you two SDIs, a 6-Pin LEMO port for Audio, some built-in NDs, and that’s about it. You can’t shoot S35 either, but I suppose if you were going to do that you might as well just get the original Mini as it’s the same sensor. In any case, like the original Mini you’l likely be building it out, hiring a separate audio crew (the LF does have a scratch mic!), and have all the usual accessories ready to go.
Differences between the Mini LF and the regular Mini are: much nicer viewfinder*, a 6-pin LEMO audio port, a sync port, 12v and 24v accessory power outlets, Codex Compact drives (on the operator side instead of the back), MXF ProRes instead of Apple ProRes, more assignable buttons (awesome), and the new LPL Mount (which is easily adapted to PL). If you’re used to using the Mini’s and have the budget, I’m sure you’ll love the Mini LF. Unless you’re shooting Open Gate you’ll be working with a 4K image, but there is no S35 mode so you’ll want to make sure you also budget for full frame glass. That means something like the Signature Primes, something like Fujinon’s Premista zooms, or the Sigma Cine Primes which are really nice. Since it’s PL or LPL only, you won’t be able to use traditional stills lenses.
*The new viewfinder is sick. It’s got a much nicer screen, a new button layout and menu, it’s hot-swappable, you can use it to control the camera from far away with an extension cable, the cable rotates in place so you don’t have to align a tab to get it in and it can’t get twisted, it’s even got a headphone jack. It even has a heater built-in so the glass doesn’t fog up! Crazy stuff.
The Alexa Mini LF is likely best suited for larger budget productions who have all the necessary workflow and production considerations to make it a smooth experience. Or maybe something like a music video where all you need the camera to do is make an image and all you need is a lens, a battery, and recording media.
Sony FX9
To be perfectly honest, we were all kind of disappointed with the FX9. It’s a great camera on its own, and the image is solid, but when compared to the other three cameras it looks less exciting. All things considered, you’re probably going to shoot 4K DCI or UHD on any of these cameras, but Sony calling it a “6K Camera” is kind of cheeky. If that’s the case, my aforementioned C100mkII is a 4K camera (someone tell Netflix!). It also came out, seemingly, before it was done cooking as there’s a ton of features that are expected to come to the camera in future firmware updates. One look at their product page and you’ll see tons of asterisks everywhere.
It seems like the FX9 is more like the Full Frame update to the FS7 instead of a new camera overall. For many people, that’s perfect. The FS7 has to be the most rented camera available right now, probably followed by the C300s and the Alexa Mini, and that means there’s thousands of workflows out there not needing to change very much. But, where the C500mkII has the C700FF sensor in it, the FX9 does not have the VENICE sensor in it.
The FX9 came out incredibly green in comparison to the other cameras, but if you’re used to editing with Sony cameras you’re probably used to that. That doesn’t mean it’s a bad look but it’s definitely something to watch out for and it really is only noticeable when compared to other cameras, I can’t say you’d see it in a vacuum. We shot in S-Cinetone, which is awesome to see come over from the VENICE as it really is a step up from the “classic” Sony look. I actually found if you over exposed it a bit it would even out the greenness, and it recovers highlights pretty well so I’d recommend aiming higher than you’d think when exposing this sensor.
The highlight features of this camera are the excellent hybrid auto focus, dual-native ISO, and variable ND filter making it a good choice for documentary style shoots where getting the shot supersedes artistic intent. It records to XQD cards, and has a utility SD slot for proxy recording and updates and the like.
The FX9 is probably best for folks who are looking to upgrade their existing FS7 workflow, primarily documentary, corporate, and ENG type stuff.
Sony VENICE
If the C500 is flexiblity on a budget, the VENICE is flexibility with a budget. Even when I first saw it a few years ago I knew it was going to be a hit. It’s got an image that stacks up with the best of them in a camera body that gives you almost every tool you could need. It’s a little heavy, sure, but it’s smaller than the full-bodied Alexa offerings and offers a ton of features that we dig. Unlike the LF, you’ve got a bunch of different crop types so if you want to shoot S35, anamorphic, 3:2, 17:9, 2.39:1, go right ahead. It can even shoot 110fps in 4K DCI. The menus are new and relatively easy to understand, and you’ve got operator and assistant-side screens (the assistant side is, let’s say, “familiar”). I love the operator-side screen I won’t lie.
The VENICE also has a removable sensor block, which allows you to use the RIALTO extension, which enables you to use the sensor block and lens mount separately from the camera body. This way you can get it into small areas like the dashboard of a vehicle with relative ease. I’ve also seen folks wearing the main camera body as a backpack and hand-holding the RIALTO, which is a great idea that essentially gives you VENICE performance in a Mirrorless form-factor. Very cool.
Its top gear is 6K, a bit bigger than the C500’s 5.9K, and it can do so in linear RAW or XOCN which Sony says is indistinguishable from RAW (and I’d argue is more or less true, save some extreme grading). I’d like to take this time to point out that I shot the VENICE on XOCN-XT thinking that was what they called raw, and that’s why we didn’t shoot raw. Whoops.
Notes
In general we noticed that these cameras look really close after grading. Each sensor is kicking out an image that is at least 10bit, and the codecs are all up to the task. While the FX9 was the hardest to match to the other cameras it still didn’t have any issues with any particular color alone and on its own still produces a really pleasing picture. Each camera had a few sticking points where I’d move, let’s say, orange one way and red or brown would inevitably move out of place, or green would want to be darker or lighter and that would move near-colors in directions I didn’t want, but that’s inevitable. Theoretically you can adjust the picture profile in-camera to try and tackle that. Overall I found purple to be the weirdest, no camera seems to record purple too well and all of them objected to aggressive adjustments of that swatch. I’d love to know more about why that was an issue.
In terms of noise, we weren’t testing the ISO range or even the dual-native ISOs on the Sony’s, just what a well-exposed image looks like on each camera and how that image can be pushed around. Overall we noticed that the cameras handled really well when over-exposed (not clipped obviously) and the Arri did the best when under-exposed (although it was the only one shooting “true” raw). Another interesting thing is that the C500 and FX9 seem to “switch” characteristics when over and under exposed; the C500 goes more green when under whereas the FX9 leans magenta when over. Also you’ll have to forgive me missing the focus on the exposure ramps for the C500 and VENICE, when it was at 5.6 everything was fine but f2 highlighted that I wasn’t dead on. My mistake.
The C500 definitely had the “milkiest” look SOOC. I don’t yet know what that means scientifically (ie why they’ve chosen to make it so flat) but I noticed that the C500 is relatively low-contrast and skews magenta (at least compared to the other three). Skin tones are dead-on, but when coloring my footage I find myself leaning on the shadow log wheel for some added “bite”. It’s an easy image to grade but you’ll be doing more luma adjustments than color most likely.
The Alexa seems to be green-ish in the mids/low-mids but skin tones are left alone. I found that interesting. If you don’t like that you’re mostly making creative adjustments. Along with the C500, the Alexa had the Brown/Orange/Tan colors dead on, whereas the VENICE (and FX9) were off a bit there, which would explain the difference in skin tone response.
The VENICE image, every time it popped up, would elicit surprised reactions from the folks we showed the footage too. People pretty universally agree it looks great. The FX9 seems to have the greenish channels dialed WAY down, which would explain the “depth” of the image and also the tint. I also found I wasn’t able to nail down the reds/pinks in a way I liked on the FX9. Oddly enough with the Sony’s I was best served leaving the red areas alone. Overall they have less luminace and are more green/yellow than the Alexa/C500’s tendency towards red/orange/magenta as the favored colors.
I found that all of the camera bodies were relatively easy to work with once you got used to where their buttons were and how the menus worked, only taking a few minutes to get acquainted. That being said, Stray and I are going to work on a new series tentatively titled “Where the #*%$ is the On Switch!?” where we give you the quickest lowdown possible on how to run not only these cameras but the other commonly-rented ones as well. No one wants to get stuck on set lookin’ a fool.
I found that offloading the footage, I was hitting speeds of around 300MBps on every camera so clearly I was just bottoming out my USB3 connection which is better to see than any particular card being slower than the other (it’d get up in to the 325/350 range sometimes though!)
This test was not meant to be exhaustive, but just a look at what one could reasonably expect from any one of these cameras. If anyone is interested in more specifics we’d be happy to gather those notes and knock them out in a follow-up test. If you read all of this and want to see me say it out loud, the video I made with Stray can be found below:
OWL BOT — YouTube // Vimeo // Instagram // Twitter
Stray Angel Films — YouTube // Instagram // Twitter