Using an FS5 to shoot in low light – what can I do?

The PXW-FS5 is a pretty good camera overall. Compared to cameras from 6 or 7 years ago it’s actually pretty sensitive. The exposure rating of 800 ISO for the standard rec-709 picture profile tells us that it is a little more than twice as sensitive as most old school shoulder cams. But it also suggests that it is only around half as sensitive as the king of low light, the Sony A7S. The A7S is so sensitive because it’s sensor is 1.5x bigger than the sensor in the FS5 and as a result the pixels in the A7S are almost twice the size, so are able to capture more light.

So what can you do if shooting in low light? 

The most important thing to do is to make the optical system as efficient as possible. You want to capture as much of the available light as you can and squeeze it down onto the FS5’s sensor. If you take a fast full frame lens and use it in conjunction with a Speed Booster type adapter you will end up with similar performance to using the same lens, without the speed booster on an A7S.

This is because the lens has a fixed light gathering capability. Use it on an A7S and all of the captured light is passed to all of those big pixels on the sensor. The light is split evenly across 4K’s worth of pixels.

Use it on an FS5 with a speedbooster and the same thing happens, all of the light is compressed down, which makes it brighter and all of this now brighter light falls on 4K’s worth of pixels. The smaller pixels are about half as sensitive, but now the light is twice as bright, so the end result is similar.

metabones_mb_spom_m43_bm3_speed_booster_ultra_0_71x_1259766-1024x1024 Using an FS5 to shoot in low light - what can I do?
A speedbooster adapter such as the Metabones makes a huge difference in low light. But you MUST use a full frame lens.

The biggest performance gains are to be had from using a very fast lens and then making sure all the light from the lens is used, none wasted. Anything slower than f2.8 will be a waste. If you are thinking of using the Sony f4 lens for very low light… well frankly you may as well not bother. The lens is THE most important factor in low light. When I go up to Norway to shoot the Aurora I use f1.4 and f1.8 lenses.

What about Picture Profiles?

The standard picture profile isn’t a bad choice for low light but you might want to look at using cinegamma 3. Although with a low light, low contrast scene none of the picture profiles will be significantly different from the others with the exception of PP2, PP7, PP8 or PP9. None of the profiles make the camera more sensitive, the sensitivity is governed by the sensor itself and all the profiles do is alter the way gain is distributed across the image.

PP2 will crush your shadows giving less to work with in post. The log curves in PP7,8,9 will roll off the darkest parts of the image, again giving you less in post. So I would probably avoid these.

For color I suggest using the Pro colour matrix. This works well for most situations and it will help limit the noise levels as it keeps the saturation fairly low keeping the noisy blue channel in check.

Gain or ISO?

I recommend you set the camera to gain rather than ISO as the ISO’s for each each gamma curve are different, so it can be difficult to understand how much gain is being added, especially if you are switching between gamma curves. Use gain and you will have a good idea of the noise levels as every time you add +6dB the image becomes one stop brighter and you double the noise in the image, +12 dB is 2 stops brighter and 4x noisier than 0dB etc. ISO is an exposure rating, it is not a sensitivity measurement. But don’t use too much gain or too high an ISO as this will affect you ability to use some of the very good post production noise reduction tools that are available.

Noise and Noise Reduction.

If shooting in very low light then you are quite probably going to want to use some noise reduction tools in post production. “Neat Video” works very well at cleaning up a noisy image as do the various NR tools in the paid versions of DaVinci Resolve.  These post production  tools work best when the noise is clean. By that I mean well defined. When using any of the 709 or Cinegamma curves a bit of gain can be used, but I wouldn’t go above 12dB as above this the NR starts to introduce a lot of smear and this than makes it hard for any post production NR processes like Neat Video to do a decent job without the image turning into a blurry mess. So don’t go crazy with the gain or use very high ISO’s as the post production NR won’t work as well on footage that already has a lot of in camera NR applied.

And if you can add a little light-

If you are adding any light use a daylight balanced light where you can. Video cameras are least sensitive in the blue channel. If you use a tungsten light which is predominantly warm/red to get a good white balance you have to increase the gain of the cameras least sensitive and as a result most noisy blue channel. This will add more noise than if you use a daylight balance light as for daylight you need less gain in the noisy blue channel.

There is no miracle cure for shooting in very low light levels. But with the right lens and a speedbooster the FS5 can do a very good job. But just in case yo haven’t worked it out already, I’ll say it one more time: The lens is the most important bit! Beyond this your next step would be adding an image intensifier for that green night vision look.

Advertisements

6 thoughts on “Using an FS5 to shoot in low light – what can I do?”

  1. This is perfect timing as I’m shooting a short horror film in the fall with the FS5. Here’s my problem that I haven’t really found and answer for yet I hope I can get your expert advice on this.

    At the end of the film we are going to have a scene where the main character is in the forest with a camping lantern heading towards a mysterious bright glowing light deep into the forest.

    I plan on shooting the whole thing in CDNG RAW for as much bit rate and color depth as possible but I’m really afraid of having noise in the shadows. The area of his face will be expose properly with and in post bringing everything down 2 stops with an input lut but I’m afraid of shadow noise on areas that are almost completely black.

    I’ve never shot raw but my understanding is that on the FS5 the best settings are through pp7 Slog2 and.cine color profile.

    Is there any way of getting the benefits of raw CDNG but without the shadow issues? I’d much rather have loss in information in the highlights than the shadows for this but I don’t want to have less flexibility in post with the standard 50mbps files that we get from standard profiles.

    Thank you again and I read every single one of your blog posts 🙂

    1. I’m sure you are already aware of this, but for the benefit of others that might read this:

      1: Raw won’t make the camera more sensitive or improve it’s ability to shoot in low light.
      2: 12 bit linear Raw has some limitations in the amount of data in the deepest shadows.
      3: You must NEVER use gain or anything other than the base ISO when shooting raw.
      4: PP7 is the recommended profile to use when shooting raw, however it is also possible to use PP8 and PP9.

      So to answer your question: If you really, really are shooting at plus 2 stops and then printing down the entire scene in post, background and shadows as well, then you won’t have a problem. But you will have a problem if you start trying to lift the shadows to pull more detail out of them. So light it properly and print down and you will be OK.

      1. So get the biggest, brightest, most powerful lamp you can for the mysterious light coming through the trees. It is the contrast between the bright light and dark shadows that will make it look like dark night time. Take a look at the opening scene of “The Assassination of Jessie James By The Coward Robert Ford” for an idea of how to do it. The lamp on the train is an 18K HMI and the oil lamps contain 60W bulbs if I remember right.

      2. So for my particular example, since I’m shooting 12 bit RAW, I know you said Sony recommends going through PP7 SLOG2. Does this bake the SLOG2 gamma into the raw? I know you’ve mentioned before that SLOG3 has a slight edge in data in the shadows because of a slight knee in the shadow curve on SLOG2. Would you recommend I do raw through SLOG3? Or is possible to get the same benefits going through SLOG2 and (if I’m not mistaken) be able to choose between SLOG2 and 3 in post production during debayer and get the same benefits.

  2. True! Unfortunately you need Canons ff leses to use the speedbooster and most of them aren’t made for Video like the sony 28135 on a ff camera. And there are a Lot of People complaining about these speedboosters with photo lenses when filming.. do you think sony will make a full frame fs7 in the near future? Venice is way to heavy for documentary work and quite expensive..

Leave a Reply

Your email address will not be published. Required fields are marked *

*

This site uses Akismet to reduce spam. Learn how your comment data is processed.