DISCLAMIER: Let me start this post by saying for the most part, I am camera agnostic when it comes to paid work. In other words, if a client has certain constraints, usually budgetary, I’ll make it work with most cameras. If I have a choice, I will choose the camera I think is right for the job. It’s not my intention to start a ‘this cam is the best’ debate. I’ve seen amazing stuff shot on every camera. I use the a7s II for many projects.
Oriana Layendecker, a friend and who I consider a genius when it comes to beauty and fashion photography, invited me out to shoot some footage of an editorial she put together. If you’re not familiar, editorial work is a big thing in the fashion and beauty world. If you’re a film person, it’s essentially a spec project. Amazing people come together to create a dope story in the form of images, but focused on fashion or beauty and a story told through the model’s expressions/clothing/styling/makeup/hair/etc. If it stands out, the project gets picked up and published by a major or minor fashion/beauty publication. We shot with a model I’ve been running into recently on jobs, Hazel Graye @ Fusion, so it was a fun shoot!
Why the a7s II?
I could only come for a portion of the shoot due to another job I was prepping for, so I knew I would not be able to shoot every single look that the team was doing. For me this shoot was to have fun, so I decided to bring a friend’s A7s II that I had on loan, with a Canon CN-E 85mm T/1.3 and see how it would hold up to my standards of normally shooting 4K and above for beauty (I usually shoot beauty on RED for the resolution). Initially, I was really disappointed with the raw footage. I shot on slog3, which I’ve since come to recognize is a mistake shooting such a flat image on an 8-bit camera. Too much to get into for this post, but see this link for why.
The Initial Results
When I scrubbed through the footage a few days later, I found it was below my standards. It was noisy and softer than I am used to with shooting beauty on RED. I shot at 120fps at 1080p. I typically shoot beauty and fashion editorials at a minimum of 48fps all the way up to 120fps, so I wasn’t able to shoot in 4K on the a7s II because of the need for higher frame rates. Here’s an image coming straight from cam using slog3:
I was shooting between 4000-8000 ISO to get a higher f-stop as well as shooting at 120fps. The a7s II is usually clean at those ISO values from my experience. The lighting was pretty bright so we definitely weren’t underexposed on the face, though I suspect the high contrast between the face and background had something to do with the noise in the shadows. However, I do dabble in color grading and was able to get back to a working image. Not all was lost!
The Final Results
Here’s one segment I put together:
I think it came out decently for what it was, but it answered my main question I went into the shoot with: Would I shoot beauty with the a7s II? For a large campaign that focuses strictly on beauty – faces, hair, lips, etc. I would not.
For social media – Instagram, YouTube, etc where it’s more about content (a beauty project that’s more reality or tutorial focused) vs strictly commercial, I have used the a7s II and would again.
I’ve found with the a7s II and other cameras, if you have controlled situations you can definitely make great images. I’ve just become accustomed to the highlight rolloff with RED and Alexa that make shooting high key and high contrast beauty much easier, especially in post.
At the end of the day – doable with the a7s II, but not my first choice.
Art direction by Oriana Layendecker
Makeup and Hair by Mary Irwin
Film by Arthur Woo
When RED announced the Epic-W with Helium sensor this year, I knew I had to make the jump to the DSMC2 line. My reasons for making the switch from Epic Dragon to Epic-W are for another post, but my biggest concern lay with the frame rates and the cropping of the sensor.
The Epic-W shoots 8K resolution, but only up to 30fps at full field of view. For narrative work, this is perfectly fine as I’m usually never shooting beyond 23.98 or 24 fps. However, I often shoot at 48fps and 60fps for commercial and music video projects. In order to get to those frame rates, one needs to jump down to 6K resolution. According to a handy tool that RED/Phil Holland developed to calculate the crop factor for different resolutions, I could easily see what the crop factor is if going down from full FOV 8K down to 6K and lower.
Who can tell me?
I asked many colleagues how the crop factor would affect the compression ratio of the lenses, and if a “50 is a 50 is a 50” would hold true – that is, does only the FOV change? Or does lens compression ratio and depth of field change when utilizing a smaller portion of the sensor? I got varying answers, so we (my colleague Brian and I) decided to do some real world tests of our own to see how the crop factor affected our lens choices.
Coming from the DSLR world, I had a good feeling for how crop factor affected lens choice. However, for the first couple of years I was focused more on learning lighting and camera movement, and didn’t “see” the effects of lens compression ratio yet, unless it was a long telephoto or ultra wide angle. I wasn’t shooting enough to see the differences between 28-75mm by just looking at an image.
Working with directors who know what they want, you add an incredible amount of value as a DP by being able to give a recommendation on a lens for a shot or to have a conversation about why the director should consider using a specific focal length rather than the one they were thinking. You have to know what each lens ‘feels’ like, both in the compression of space as well as FOV. This is why figuring out which lenses I should use while cropping is important to me – it would affect my recommendations.
I don’t want to get too technical on this post because quite frankly, I was looking for which lenses I needed to use if we changed the distance from the subject to retain the same field of view. Here are various frames with accompanying explanation.
24mm CN-E at 8K, 6K, 4K
I think this test is the most telling. We often use the 24mm close up to talent to make a space seem larger than it is (but try to avoid edge and face distortion). We wanted to see how the cropping of the sensor down to 6K and 4K would look to the camera.
8K, 24mm, f/2.1, 2′ 1″ from subject. This frame just screams 24mm just by looking at it. Most experienced DPs know this look and can tell you the approx focal length just by looking at it. I love this look for certain projects. Slight distortion and a smaller background makes the subject seem a little large than life. We normally try to avoid the facial distortion you see here from being too close, but sometimes you want to exaggerate this look.
6K, 24mm, f/2.1, 3′ from subject. With a 1.33x crop from 8K to 6K, our lens effectively becomes a 32mm focal length. While it retains the property of a 24mm lens (because the lens doesn’t change, only our FOV and crop factor), to achieve a similar frame above, we moved the camera back. Notice that because it’s now a 32mm focal length to the camera, moving the camera back to achieve a similar frame has the crop factor increasing the compression ratio in camera (again, the lens itself does not change properties). You’ll notice the building and the trees are much closer to subject now that previously, even though we haven’t changed the lens.
4K, 24mm, f/2, 5′ 8″ from subject. We then changed to a 4K frame, which is a 2x crop, making our 24mm lens at 4K effectively a 48mm lens. The jump from 8K to 4K is pretty obvious with the compression ratio of the background to foreground elements now looking much closer together. Also, we begin to lose the distortion of the subject while entering the ‘normal’ lens field of view.
The following images display the same behavior with the 24mm at 8K, 6K, and 4K respectively, with the backpack as our reference point. You can see that the lens compression ratio changes based on the crop and where we moved the camera in relation to the backpack to match the frame for each shot. The lamp post creeps closer and closer in the more we’re cropped on the 24, even though the backpack is in the same general size to camera.
Stationary camera – “a 50 is a 50 is a 50”
Here we compared 8K, 50mm, f/2, 11′ 2″ from subject to 6K, 35mm, f/2, 11′ 1″ from subject (Backpack). The 6K 35mm is equivalent to 46.55mm. We can see here that when we don’t move the camera, there is a difference in the lens view and space compression even though our FOV is approximately the same. We wanted to do this test to see what a lens would behave like if we didn’t move the camera (they’re about an inch difference from one frame to the next in terms of distance from the subject). This effectively shows that a “50 is a 50 is a 50” – if you’re not changing the distance of the camera from the subject, the lens properties don’t change, lens compression ratio does not change, only FOV changes. As a sidenote, the depth of field is greater on the 50mm as well*. It’s when you move the camera to match your frame where we see changes in the compression of space.
Conclusion and Caveats
The takeaway here is that when there is a crop factor, on the sensor, lenses can be somewhat interchanged when working with ‘normal’ focal lengths, it’s okay to use an equivalent focal length to compensate for the crop factor. I’m comfortable swapping a 35 for a 24, a 50 for a 35, and a 85 for a 5o on a cropped sensor. We did not test what happens when we shot at 24mm 8K, jumped down to 2K, to mimic a telephoto lens – first of all this was not a technical test, and secondly we would not do that for the type of work we do. This was a practical test to see how we can compensate for the loss of FOV when cropping by using different focal length lenses.
See anything that you think needs to be corrected or looked over again? By all means, reach out to me at email@example.com – I would love to hear your thoughts!
*This article has been reviewed and critiqued by various cinematographers – thank you for your help in keeping me honest with the technicalities.
I’ve worked with fashion and commercial director Brian Choy quite a bit in 2016, and he brought me onto a shoot with a long time client of his, the GRAMMYs. They’ve been shooting a series called “Reimagined”, where up and coming artists perform covers of chart topping songs. Most of the series has been shot in LA, but there has been a need for NYC crew for East Coast talent. Brian was selected to direct a couple of pieces for Devvon, an NYC based performer with a huge social media/YouTube following.
With his fashion background, Brian drew inspiration from fashion shows with small hanging light fixtures (also reminiscent of indie music vids). He also had in mind a long camera move that would open the piece, and asked me the best way to do it within the budget that was given to us. I recommended using a dolly (originally wanted a Fisher, but due to space and budget we had to do Dana) move across the stage, with an operator on a fluid head tracking the talent (who would be standing in one spot), and using an operator controlled zoom on a cine zoom lens. I could tell the team was ambivalent about my recommendation. What do you do as a DP when your verbal recommendations aren’t met with full confidence?
This was our final outcome, but how did we get here? Photo by Tom @ Mastershot Films
This is where the power of PreVis comes into play. This was the second project this month I was able to plan for using Cine Designer in Cinema 4D, and convince the director/producers how we were going to pull it off, all before ordering any equipment or me having to jump on any complicated conference calls to try to explain what I was thinking. I was easily able to get the dimensions from the studio we shot in and mocked up the entire camera move on a dolly with a fluid head and an automated zoom. I was also able to calculate the distances from the dolly to the artist and order the proper zoom with the focal lengths we needed. In this case we ordered a Canon CN-E 30-105. This is the mockup I did in Cinema 4D:
With these calculations there was no doubt that we could pull off the camera movement Brian wanted. Using PreVis as a DP is groundbreaking. A big thank you to Matt Workman and Cinematography DB for creating this amazing tool for DPs to stay relevant and save time and money in pre-pro.
Now that we had the camera A move and cameras B/C placement locked in, we went onto the lighting. My assignment was to get a cooler color on the background and a softer key. We ordered some Lagoon Blue gels and had some 2ks with 216/250. The look still needed to look like a live performance and not like a music video, so we had to balance the look of stage lighting with the inspiration from the brief.
We worked with set designer Espen Øydvin, a young, up and coming designer hailing from Norway and currently in NYC. He was able to string up all the bulbs manually (which I think took him a couple of days!). He also gave me the information I needed to calculate the footcandles all the bulbs would be giving off in a given area, which was helpful in determining the brightness of the lights we would need for proper exposure on talent without completely losing the bulbs.
Gaffer Thomas Chaves from Mastershot films also did a great job, given that he had dual Gaffer/Grip roles. It was challenging but he got the job done. He even recommended the Lagoon Blue to us when we were originally thinking of a colder blue, like steel blue, but the lagoon was really nice, still cool but a little more inviting that steel blue.
I had top notch AC Kyle Forbes on this project and he helped me get situated with the Microforce zoom to get me the automated zoom I was looking to achieve.
We had a hazer on order but we found out the day of that the ventilation system in the studio wasn’t able to be turned off, so the haze kept leaving the room. This actually worked out; even though we didn’t get the stream of lights from the lekos that we wanted, it created a live atmospheric haze that the client actually really liked, so we rolled with it.
We had a great team that was able to get it done despite a limited amount of time/space that we had to shoot in. Here’s the final outcome:
If I could do it over again with a slight bump in the budget, I would definitely have used a Fisher vs Dana dolly. Danas are great for small moves, but it was challenging to repeatedly push the dolly across 20 feet of speedrail, operate the pan/tilt to follow the talent, and control the zoom all at once on my own. It was doable, but not preferable for repeated shots over time. Luckily Devvon was great at getting his songs done in a few takes so we didn’t have to roll all day and possibly get less accurate on the camera moves.
Directed by Brian Choy
DP: Arthur Woo
Production Design: Espen Øydvin
Key Gaffer/Grip: Thomas Chaves
1st AC: Kyle Forbes
Camera Operator: DiDi Lin
Audio Engineer: Marc Smith
Production Assistant: Ryan Mitchell