Sony Venice Review

When I saw a post on facebook from Cinematographer Leland Krane about asking for a DIT for a camera test of Sony’s new flagship camera, the Venice, I was instantly intrigued.  I enjoy diving in with these new cameras to get a sort of headstart on the problems and issues, because normally these things take a while to sort out all the production bugs (engineers sitting in a room coming up with design ideas tends to not work so well on set), and so I wanted to see what all the fuss was about.  I was not able to make the test, I was working on The Good Cop season 1, so I sent Michael Maiatico in my place instead. 

Mike and Leland really put the camera through it’s tests.  They tested different lenses, filters, exposure limits, all sorts of things that one would expect.  For me, I wasn’t too keen on the 709TypeA in camera delog and so I wanted to create my own if we were going to use this camera.

Side note: I don’t like any in camera 709 look, I think they all look like shit.  My SOP is to take a LOG image down the SDI and make my own 709 that I think fits for each type of show.  It’s mainly only on commercials or short form content that I let the in camera 709 play.  This is true with Arri (which is more yellow) and Red (which is more green), so my thoughts on Sony’s 709 isn’t any different.

Here’s the 709TypeA.Here’s what I came up with.

Mine is tweaked to make the skintones pop as well as take out that green hue that’s in the shadows.  The TypeA is okay, but I prefer my version.  To each their own.

Here’s the link to the in-camera 709, in case anyone needs it:

http://bit.ly/2HFpalf

I had worked with Sony’s F55 for 6 months on HBO’s Vinyl and thus I knew what to expect with that camera.

For Vinyl we shot with the F55 in 4K RAW with limited usage of the internal XAVC when we used Movi. I feel that 4K RAW is the only way to shoot on the F55; XAVC does not do that camera justice whatsoever.  It’s only 422 10bit for the internal codec, and it artifacts like crazy when you start pushing it in post.

Seeing that I was unable to go to the checkout at AbelCine Brookyn, I was a bit worried when I heard we were unable to use the R7 recorder on the Movi for our project. The R7 was just too long and we weren’t able to distribute the weight properly with the longer Thalia lenses, so we were forced to shoot with the internal XAVC480 @ 4K.  After spending some time the F55s internal xavc I was NOT happy about this.

I was extremely surprised.

Take a look for yourself, I think you’ll be surprised to see that there isn’t any artifacting, nor is there any banding in the highlights.  The skin tones look pretty spot on and the color representation from slog3 really hits the nail on the head here.

The images out of this camera with the 4K XAVC are stunning. What’s even more exciting is the FF 6K 3:2 option, which I haven’t seen in a while.

Here are a few full 4K & 6K 3:2 frames from the shoot, left in S-Gamut3.Cine & Slog3 so you can apply whatever delog you want

http://bit.ly/2HgTOSM

Working on set with it however, is a whole different thing.  For our shoot I had my protégé Michael Maiatico be our DIT (Michael is a fantastic DIT in his own right, I taught him everything I know) as I knew he’d be interested in seeing how this camera works in the field (He was our Tandem DIT for Vinyl so he also knew the F55 inside & out).

Here’s what he has to say about his experiences.

Michael Maiatico:

Gone are the days of the F55. Something every camera assistant and DIT have been not so secretly crossing their fingers for since its introduction. Mired by a buggy release that even after years into its lifespan had several issues left unresolved the camera featured unintuitive menu systems, a resource dependent processor with woefully lacking communication skills, and an inconsistent inertia based jog dial to navigate with as icing on the cake. The list doesn’t even end there, but this write up isn’t to bash questionable choices of the past. I’m proud to say, much like an idealized life partner, Sony actively sought out constructive criticism and incorporated the feedback into their latest flagship cinema camera, the Sony Venice.

With a camera test and two shoot days with the camera I was able to get a good feel for the camera but bear in mind we’re working with beta firmware so there’s a chance certain features missing may or may not be available with its release.

First up, the pretty. The camera’s image is gorgeous. During our test at Abel Sony reps provided a Sony X300 4K HDR monitor to see what the potential of this sensor in a BT.2020 color space will look like and for the first time in years I had an image truly wow me. I’m genuinely excited to see the integration of HDR monitoring on set as the technology becomes more accessible and supported via hardware. My feeling is that there’s still going to be some time on this though as the X300 is an unwieldy behemoth and belongs fixed in a coloring suite. Also considering the cost of this monster, as well as the cost of having to upgrade most of my hardware, I’m okay leaving this eye candy to the mastering stage for the time being.

Menu structure, UI and overall UX on the Venice is leaps and bounds better than the F55. While still not quite to the simplistic and succinct nature of its German competitor a lot of inspiration has been taken. Most notably there are less options buried in confusing locations and more is accessible from the side menu. Additionally, there is an operator side menu as well which maintains the majority of control the AC side has. While I don’t see many operators making use of this on the jobs I do, having the ability to control the camera on either side is a greatly welcomed addition for the body in tight spaces.

Also, 8 single stops of internal ND. Finally, this has become a reality on a camera and its as useful as we’ve always dreamed. What’s better is that there was no noticeable color shift even up to ND 2.4. Imagining the potential time save this will be on fast paced TV or commercial sets as well as crane and underwater work brings a smile to my face. I’m always looking for ways to reduce my imposition on my AC brothers and sisters, so not having to bug them with frequent ND swaps when working with an indecisive DP is a much welcomed feature that benefits the whole department.

It seems that Sony has also removed the SDI split signal ‘feature’ that made on set to off set consistency a commutative disaster. With the F55 on set monitoring was never truly accurate since the signal was both extended and legal. I still don’t know Sony’s justification in that decision, but in order to get the closest match on set to post our LUT boxes were set to Extended in Legal out. This is still true for the Venice in that it is sending an extended signal that will need to be changed to a legal out at the LUT Box, but from my admittedly less than scientific tests it didn’t have the same inconsistency in highlight representation that the F55’s split signal displayed. I could be wrong, but I hope I’m not.

We captured our footage via an R7 recorder into Sony’s X-OCN ST codec, which if you haven’t had any jobs with an F55 since the R7 came out is a much welcomed compressed Sony RAW codec that lowers your total data load somewhere around 33%. In my personal stress tests, I struggled to find any tangible loss in push/pull or color manipulation in direct comparison to uncompressed Sony RAW. As mentioned above, however, with the improvements in image quality with the internal XAVC codec this is a less critical feature than it was with the F55, but with HDR delivery increasing in demand it’s best to utilize X-OCN. There’s also an X-OCN LT codec available which is even lighter than the ST (about 60% smaller than uncompressed RAW), but I haven’t tested it thoroughly enough to figure out its ideal application or where it falls short of its heavier sibling.

In our working build of the camera we were limited to FF 6k 3:2 (6048 x 4032) mode at 23.98fps. From Sony’s documentation it seems that that 23.98fps limit is here to stay at max resolution but should allow for 30fps in Full Frame 17:9 and 16:9 modes and up to 60fps in 4K modes. I find it interesting that Sony chose to have such a low max FPS on this camera and not include higher speeds at 2k like the F55. A selfish part of me is relieved that I wont have directors or clients in commercial environments throwing off data calculations with last minute high speed capture changes when they inevitably ask “How slow can we go?” but the technophile in me still would love more frame rate options to have in the camera’s proverbial tool set.

Features that weren’t included during our test, that I’m excited to see, included in the official release are most notably the dual native ISO and interchangeable sensor. Since the Panasonic Varicam first debuted its’ dual native ISO, I’ve heralded it as one of the most useful advancements in imaging technology that I hope other manufacturers would take a page from. The Varicam 35 and LT tout a native 800 and 5000 ISO while the Sony Venice claims a 500 and 2500 and I can’t wait to get my hands on the two together to test in a controlled environment. My trick with the Varicam was to pull down the native 5000 to 2500 in order to shift my stop distribution to retain more information in the shadows. The only gripe I had with the Varicam at 5000 native was that its look was noticeably unique and up until now never had another camera to compare if this is inherently due to the high sensitivity or just Panasonic’s implementation. I hope Sony brings the heat here so Arri and RED will be competitively pushed into adopting the technology.

The interchangeable sensor brings a very palpable ‘futureproof’ quality to the camera as well. If the resolution wars continue and deliverables start calling for more than 4K, rental houses will be able to throw in something like an 8k sensor without having to reinvest in an entirely new body. Personally what I’m more interested in, than needlessly increasing pixel count, will be more courageous choices in sensor options. The notion of interchangeable sensors acting like different film stocks always felt like an untapped concept in the digital world. Unfortunately for that dream, and fortunately for my career, the power of 3D LUTs is incredibly extensive and can easily act as those different stocks so I guess the point is moot. That being said I’m pretty positive we’ll at least see a monochromatic version of the sensor, likely at an even higher base sensitivity.

Sony has a lot of work ahead of them, if they intend to reclaim prominence in the world of digital cinema, mostly in overcoming negative associations with their last generation of cameras. Despite the uphill battle, I believe with the right amount of exposure, the Venice provides enough competitive features and course corrections from real world feedback to re-instill hope in the company and the future of their cinema offerings.

Everyone on this project kicked ass.  We had some trouble out of the gate, like the camera truck breaking down and showing up 3 hours late, but in the end we got everything we needed to.  Music videos are sort of regarded in our industry as crazy long hours, not managed properly, and poorly executed.

It was my goal to not have any of that and I hope that my crew enjoyed being a part of this project as much as I did.  I’m very thankful to everyone.

Here’s the link to the music video, feel free to check it out on FB or YT:

Here are some BTS photos from the shoot.  If you have any questions feel free to post a comment below and I’ll respond as I can!

Advertisement