For a Low Light shoot with FX6 / FX9 - best ISO and Color profile settings?

JimmyMcV

JimmyMcV

New member
Phillip Bloom pointed out that S-Cinetone (on the FX6 at least) will have lower noise than Slog3 at higher ISOs. Yes?

I'm shooting a haunted house performance that will have some very low light scenes. I'm questioning whether my default choice should be Slog. The FX6 has dual base iso -- so you have to shoot at 800 or 12,800 when in log mode I believe. If I shoot in S-Cinetone, however, I can go to a higher ISO and make my HIGH BASE ISO higher than 12,800 and still get some decent results in terms of noise? (I realize that not shooting slog will decrease dynamic range).

EDIT: I just found out my second camera, the FX9 has a high base iso of only 4000. Oof. So now I wonder what to do about that.

I'm very new at this camera so I'm unsure what the optimal settings are for shooting low light.

I'm also thinking of foregoing my stock zooms lens (f4) and shooting with the 24mm (f1.2) to help get more light in there as well.
 
Last edited:
Just how dark do you anticipate it to be?

Remember that with S-Cinetone your HI base ISO on the FX6 drops to 5000.

I've done some low-light shooting in both modes (e.g., in the 20-30 minutes after sunset, with relatively slow lenses) and I think I found s-log3 to be preferable. Normally I might say to avoid adding extensive gain in post, but in this case it seemed to work better than going with the in-camera gain of S-Cinetone.

Having said that, YMMV, and I'm keen to see what others have found.
 
The FX6 at 12,800 will see all kinds of things your eyes can't. If you can see it, the camera can too.

Personally, I hate mixing workflows, there's just too many unknowns to worry about - so if it were me. I'd just shoot in regular Slog3 Sgamut3.cine, put some fast glass on the cameras, and go from there.

Also, overexposure is your friend in these low light situations. So if you can get away with rating the camera at 3200 or 6400 EI when shooting 12,800. That'll net you cleaner results.

The FX9 is very clean at 4000 ISO, but will still benefite from a bit of overexposure if you have the headroom.
 
What is this headroom? In the dark clearly you are shooting wide open* and even considering 360 shutter. You at at the head of the headroom already.

Leavaing only selection of profile and iso which I cant really comment on as I dont own these cams.

My modern canons certainly do better when ISO is ramped in camera not in post.

Deoending on the scene/situation I think a bit of rim lighting can do wonders, even $20 amazon panels -even if it is a ghost train such additions would probably work for the viewing public as well as the camera team

edit: one other thing. 'phillip bloom says' forget that. unless it is a rental camera a test in a cupboard and playing with resultnt files in post is not beyond most of us. in post be bold.





*some lenses dont really work wide open as lack of dof or too moozy (veiling flare?)- they should be as open as possible
 
Last edited:
Generally I agree, it doesn't matter what Phillip Bloom (or me or anyone else) says, it's usually better to find your own best. I did some low light tests the other day, only brief ones, but I found I agreed with PB on this occasion (I don't always!). I found up to about 14db in cinetone the results were better than slog at 12,800. In my brief tests I also found that hi sensitivity in cinetone at 6db was still superior to slog at 12,800. But I agree with Sam, you need to find what's best for you. And you also need to watch the contrast/rolloff balance in cinetone, but that's another story.
 
I think I'm going to do some tests. Since the high base on the FX9 is 4000, should I set that on the FX6 to match? But since 4000 is not the FX6's "high base" will that be noisier than the FX9 at 4000?

The trouble with this shoot is there are going to be parts where there's plenty of light. But I will have any stoppage time. So, I guess I need to set it to the ISO that works for the low light and then simply close the aperture (or add ND?) when I get to the well-lit parts. Yeah?
 
I'd probably set them both to their respective high bases in whatever profile I was shooting and work from there. If in S-Cinetone, I would add further gain if necessary. 24mm 1.2 sounds like a good idea. Yes, ND for the brighter parts.

Will you have time for a quick run through before the shoot? You might find that the things that are supposed to be seen by the audience are bright while the darkness is supposed to be dark. No one wants to see the door to the staff toilet.
 
Ok, thanks. I just shot some Cinetone and Slog3 in my living room and will compare tomorrow. Luckily, yes, I will be able to watch the whole run-through the night before. The FX6 is so lightweight I think I might be able to hand hold this thing for 90 minutes. The only problem with the 24mm is that there is no stabilization with that lens. So, if I do use that instead of the 24-105mm zoom then I'm definitely shooting 4K so I can stabilize select shots in post. (Output is 1080).

So, it sounds like, if I shoot Slog3 I should just put the FX9 in high base and FX6 in high base and hope that they have similar amounts of noise/grain?
 
I agree with Grug that you want to overexpose when shooting at high base ISO. I usually just expose to the right and also check false colors to see that most or all of my detail is above minimum exposure. And, yknow, expose as hot as possible without blowing stuff out

I think sam Morgan Moore is right that you have a ceiling to how brightly you can expose. You can't always get an overexposure even at high base ISO. But the point is that you should aim high, not for a standard exposure

And i agree its probably best not to match ISO but to set each camera to its respective high base ISO
 
Well, now I’m not sure. Did another set of tests today and found that S log at 12800 had slightly less noise and better control than cinetone at 14db In really low light. Guess I need to tighten my test methods! The moral is: do your own tests and don’t listen to people like me!
 
ahalpert said:
I agree with Grug that you want to overexpose when shooting at high base ISO. I
Click to expand...

That would not be my advice. Low and High Base should be exposed to the same target level -- whatever someone chooses that target to be. Purposely over-exposing High Base at a target level higher than what you would use for Low is unnecessary and it won't help. If someone feels that over-exposing High provides the best results, then I would say they should also be over-exposing the Low Base as well. The two should be treated equally.
 
Last edited:
My feeling is that it helps you escape the increased noise floor by overexposing at 12800 and pulling it in post
 
If that was true, then why would it not be the case all of the time? Why just do it for High Base? Why not over-expose Low Base as well?

LOG is a curve. It is not linear, and over-exposing can cause problems with grading. if you could just pull things down in post with no ill effects, there'd be no reason not to push every exposure right to the very brink of clipping all of the time.
 
I think you can get a suboptimal result by exposing too hot at high base ISO. I don't know about rating 12800 at 3200 as Grug suggests (although I'm sure he's done more methodical testing than I have)

But generally I observe relatively more shadow detail apparent at low base iso and relatively more highlight detail apparent at high base iso. And i thought that was common if not universal among digital cameras

And there are people who ETTR low base iso to reduce noise. I don't personally do that but I feel like it's more of an issue at high base ISO and I've observed that if I expose it the same way then I'm not pleased with my shadows
 
I'm just saying that I expose High and Low to the exact same target level and I see no advantage to going beyond that target level for High -- thus over-exposing. If there is an advantage to the technique you mention in post #12, I'd like to see some video that shows it, and also some explanation as to why that technique would work for High but not for Low.

I think threads like this really make exposure seem a heck of a lot more complicated than it needs to be. Just choose a target level . . . and if you can't hit your target with Low then switch to High, but the target stays the same. Simple as that.
 
If it is a true low light situation then overexposure is not a possibility as t2 at 360 won’t get above 60 on the wfm - that’s my idea of low light- if this is the case then the top roll of the curve does not come into play.
if 5.6 is hitting 100 you don’t indeed want to open to 2
 
Agreed. In a true low-light situation much of this conversation is moot. But some people might consider that anytime you must jump from LOW to HIGH to get a proper exposure in S-LOG, that right there constitutes a low-light situation. In other words, if there is not enough light to use LOW, then it could easily be called low-light situation. And the jump from LOW to HIGH is a very significant change in sensitivity that will often allow over-exposure if someone chooses to ETTR. My advice is not to over-expose in that situation. I aim for the same exposure target as I would normally try to achieve -- by stopping down a little bit or adding some ND when possible. I don't over-expose in High (or Low) even when I can.
 
My current plan for tomorrow's big shoot: Since we will have on-camera lights and many of the scenes will have enough light, we plan to shoot in SLog3 and start at LOW BASE and for those scenes where it gets crazy dark we'll press a button to switch to HIGH BASE. I think we'll have the Exposure Index for both bases rated one stop lower. So, half the number (e.g. High base 12800, with EI at 6400). Hopefully this will reduce the noise in the shadows a bit .

This guy made a decent video on shooting with an Exposure Index at 1 or 2 stops lower than your base:

https://www.youtube.com/watch?v=YBUaXkFK5gQ
 
Doug Jensen said:
I'm just saying that I expose High and Low to the exact same target level and I see no advantage to going beyond that target level for High -- thus over-exposing.
Click to expand...

I'm not sure I agree with the one-size-fits all approach to exponometry. For example -

I received many interviews to edit from a client that I thought were really noisy. The lighting was pretty (Rembrandt) and usually no color correction was needed or desired beyond the standard Sony LUT. But the seamless backgrounds were dimly lit and often very noisy, which polluted their color. The noise was also exacerbated when they wanted me to crop in for closer shots or if any exposure adjustments were required. I assumed that they raised the gain in-camera and that they had insufficient light levels after diffusing their sources.

But then I shot one of these interviews for them, and the producer kept asking me to bring the brightness lower and lower. And I realized that she was driving the noise train. (This was at base ISO.) So, without telling her, I turned the brightness on my field monitor all the way down and brought up the exposure in camera, essentially showing her a preview of how it would look after pulling it in post. (A -1 EV LUT could have done the same.) She approved the look. And when I brought the exposure down in post, it looked radically different from the other interviews I had received. Clean background with rich, unpolluted hues.

Yes, you risk pushing your highlights up into unfriendly territory with this method. But the brightest part of the image here were highlights on skin/hair/clothes, well below the top of the curve even with overexposure.

However, for a normal scene without such large areas of dark, rich hues and subtle gradients, I would consider no benefit from overexposure.

Also, what's your take on exposing a scene with mixed lighting - shadow and direct sun? Let's say the subject is in shadow or they're in dappled light. Inasmuch as I'm trying to evoke a feeling in the viewer rather than simply record what I see, I feel like there can be multiple acceptable ways of exposing the same scene and the question is what you want to achieve in that specific instance and how to get it
 
ahalpert said:
I'm not sure I agree with the one-size-fits all approach to exponometry. For example -

I received many interviews to edit from a client that I thought were really noisy. The lighting was pretty (Rembrandt) and usually no color correction was needed or desired beyond the standard Sony LUT. But the seamless backgrounds were dimly lit and often very noisy, which polluted their color. The noise was also exacerbated when they wanted me to crop in for closer shots or if any exposure adjustments were required. I assumed that they raised the gain in-camera and that they had insufficient light levels after diffusing their sources.

But then I shot one of these interviews for them, and the producer kept asking me to bring the brightness lower and lower. And I realized that she was driving the noise train. (This was at base ISO.) So, without telling her, I turned the brightness on my field monitor all the way down and brought up the exposure in camera, essentially showing her a preview of how it would look after pulling it in post. (A -1 EV LUT could have done the same.) She approved the look. And when I brought the exposure down in post, it looked radically different from the other interviews I had received. Clean background with rich, unpolluted hues.

Yes, you risk pushing your highlights up into unfriendly territory with this method. But the brightest part of the image here were highlights on skin/hair/clothes, well below the top of the curve even with overexposure.

However, for a normal scene without such large areas of dark, rich hues and subtle gradients, I would consider no benefit from overexposure.

Also, what's your take on exposing a scene with mixed lighting - shadow and direct sun? Let's say the subject is in shadow or they're in dappled light. Inasmuch as I'm trying to evoke a feeling in the viewer rather than simply record what I see, I feel like there can be multiple acceptable ways of exposing the same scene and the question is what you want to achieve in that specific instance and how to get it
Click to expand...

No offense, but this sounds like such a bassackwards way of setting exposure that I wouldn't even know where to begin to offer a counterpoint. I would never, ever, set exposure by how something "looks" on the monitor or viewfinder. That is a recipe for disaster and totally unnecessary with modern cameras. Why do you think light meters, waveforms, zebras, histograms, false color, etc. were invented? And on top of that, I would never let a producer or client pressure me into doing something that I knew was fundamentally wrong.
 
You must log in or register to reply here.
Back
Top