FS7: Fair To Say? S-LOG 2 for day and S-LOG 3 for night

K

karma17

Well-known member
Hope I'm not beating a dead horse, but as someone new to the FS7, I'm trying to get my mind around S-Log and trying to use the "best" setting for the given situation. Quite honestly, it has been a little frustrating.

When I am looking at the options for Hypergamma curves, it seems Sony offers two flavors of each.

For instance, for the FS7, they have HG 3 with middle gray at 40 and HG 4 with mg at 33.
Also, HG 7 with gray at 40 and HG 8 at 33.

My understanding is that for daylight shooting, it would be better to go with HG4 or HG 8 with the lower middle gray
and HG3 or HG 7 for night with the boosted gray points. I also understood that the gamma curves are applied before
any compression occurs and therefore, does not degrade the image quality in that sense.

This leads me to S-LOG 2 and S-LOG 3, which seem to follow the same pattern, that is, S2 has a lower gray point than S3,
so it would be better suited for handling lights and brights, whereas S3 has a boosted gray point, which would seem to be
better for low light.

But, after seeing Mr. Jensen's video showing that S-LOG 2 and 3 can be graded with the appropriate LUT to look virtually identical,
perhaps it doesn't even matter?

It just seems like Sony is providing two versions of each curve, and perhaps there is some reasoning for that.

So far, I'm primarily shooting in CINE EI, S-LOG 3, but can't say I'm that happy with the way it handles highlights, which is making
me want to use S-LOG 2. You can say I might have had too much of the S-LOG 3 Kool Aid. My
gut sense is to use S-LOG 2 for day and S-LOG 3 for low light, realizing both will need correction in post.

Thanks.
 
Is switching worth it for the extra complications in your post workflow? For me its not and I find compression to be a much bigger issue than what flavour of slog I shoot.
 
I guess at this point I will just shoot in CINE EI Slog 2 and not worry about switching in low light situations. It does seems like S-LOG 2 does a better job of masking noise in the shadows than S-LOG 3, at least straight out of the camera.

Thanks.
 
My preference is to use SLog 2 or 3 for daylight but std gamma at night. Both logs seem more noisy in the dark.
 
Why do people ask forums for information that is simple to draw ones own conclusions.

Get a camera a dark situation - available every night - and shoot some tests.

--

I say this because what I will now say will confuse - there is dark there are low and wide dr scenes and there is under exposure.

Under exposure does not need to happen at night - one can have a long shutter speed and or a fast lens.

Some night time scenes have low dr, some high.

I would say..
Low dr scene well exposed - slog 2
Wide dr scene well exposed slog 3
Slightly thin exposure slog3
Very thin exposure - avoid if possible - engage custom - up the ISO - I'm not sure if custom slog is worse than a hyper gamma.
 
I have shot test footage and my conclusions were:
HG 3 ISO 800 : the least noise, darkest image
HG 7 ISO 1600 : minimal noise, increased brightness
CINE S2 ISO 2000: noticeable noise, increasing brightness
CINE S3 ISO 2000: the most visible noise, brightest image

However, upon applying a LUT and grading, the S3 could be made to look like the HG3, HG7, and S2. This conclusion was also reinforced
by seeing others who have shot in S2 and S3 and made the footage look virtually identical, that is, with the exception of S3 blowing out the highlights.

And so, my conclusion was (and is) that it doesn't really seem to matter much whether you shoot in S2 or S3 because they can be
graded to look virtually identical. And in low light, the same appears true. However, it does seem that S2 handles
highlights better (based on what I've shot so far) and so what I don't understand is how SLOG 3 can be said to have a wider dynamic range when it clips at 93 IRE?
Is this what Art Adams meant by gravy stops in one of his articles? Nice if they're there, but don't count on them.

I am still learning about all this and do not always presume my conclusions are correct.

And I do appreciate the responses.
 
If you're doing such tests and grading afterwards to match images then you know as much if not more than people here.
-

Slog v Slog.

DR is how many stops below clip you get usable data.

So if your data becomes unusable 6 stop below clip with slog2 but unusable 8 stops below with Slog3 then has more DR irrelevant of clip point number.

This should be a simple test - open wfm shoot a white card open iris until clip incrementally stop down. Examine conclude.

Don't be fooled by fools that are on this board.
 
Last edited:
I said he should consider it for his tests - I did not present a conclusion.

My personal observations are thus.

-"dark" is a non ideal shooting scenario. You are going to have to make some trades with the devil.

In camera ISO boost and "wrong" colour space will give you results that are better than "correct" colour space mixed with huge under exposure.

Would it not seem logical that processing the image before compression is better than post processing especially when the compression algorithm is weak in the shadows.

Logically the compression algorithm is weak in the shadows because (in baked world) we concentrate on skintone not shadow detail - so when designing compression one trims data in the shadows.

I would guess (and casually observe) that xavc is designed primarily for capturing a baked image.

-
I would also suggest that shooting raw and presenting raw with various post routes is not a good way to extrapolate how capturing xavc will behave.
 
karma17 said:
But, after seeing Mr. Jensen's video showing that S-LOG 2 and 3 can be graded with the appropriate LUT to look virtually identical,
perhaps it doesn't even matter?
Click to expand...

Weren't those clips graded from the same RAW clip on F55 clip though? Makes it meaningless when considering what Gamma curve to use for internal capture on the FS7. Do as Sam says, make your own test and draw your own conclusions. FWIW, which is nothing, I often shoot SLog3 ISO2000 outdoors and ISO1250 inside YMMV.
 
Sometimes I wonder.... Slog was never designed for lowlight. Neither s2 nor s3. In such a situation you would end up with footage that may become ungradable. Use standard for lowlight, use the HG PROFILES as suggested by Dough. You will not need more then the classical 6 stops from rec709 displays, even if you have also more with HG.
 
wschmid said:
Sometimes I wonder.... Slog was never designed for lowlight. Neither s2 nor s3. In such a situation you would end up with footage that may become ungradable. Use standard for lowlight, use the HG PROFILES as suggested by Dough. You will not need more then the classical 6 stops from rec709 displays, even if you have also more with HG.
Click to expand...

Yes surely logic would dictate that the sole use of Slog 2 or 3.. is for max DR.. if there is a large DR or at night time.. bugger all DR.. why use slog at all.. with its tiny amounts of data in the bottom end.. so it can handle a large DR.. ?
 
Something I keep seeing said in these threads is that "SLog2 can be graded to identically match SLog3" which is true (nowadays that's pretty much true with any two cameras / formats within reason) BUT isn't it also true that the log profile you choose plays a larger role than that statement implies? For example: SLog2 will give you more stops / range into the highlights which means if you shot the exact same exterior scene the SLog3 profile will lose information / detail / gradation in the sky before SLog2. So while the shot overall may match in color, etc. there would still be clipped highlights in SLog3.

Is there something wrong with my way of thinking of the two different options?
 
If you go back and read the threads you are referring to you will see that I have pointed out that difference numerous times. Even though there usually isn't much difference between them after grading, I still prefer S-LOG2 for the reasons you have cited.
 
Liam Hall said:
Weren't those clips graded from the same RAW clip on F55 clip though? Makes it meaningless when considering what Gamma curve to use for internal capture on the FS7.
Click to expand...

Why is it meaningless? The camera used to shoot the S-LOG footage should not matter, and the advantage of RAW on the F55 is that I can grade the clips in post both ways -- with no harm done.
 
Last edited:
Doug Jensen said:
If you go back and read the threads you are referring to you will see that I have pointed out that difference numerous times. Even though there usually isn't much difference between them after grading, I still prefer S-LOG2 for the reasons you have cited.
Click to expand...

That's what I took away from your posts as well, Doug so I'm glad I was on the same page with your intended message.

However, I think some are acting as if it makes NO difference in which you choose to shoot with because in the end they can match but I wanted to point out that it DOES affect what you're able capture / DR you have to work with in the grade.
 
Doug Jensen said:
Why is it meaningless? The camera used to shoot the S-LOG footage should not matter, and the advantage of RAW on the F55 is that I can grade the clips in post both ways -- with no harm done.
Click to expand...

It is not meaningless in terms of colour but for DR evaluation it is potentially flawed.

Because raw is not compressing the data..

If you examine this image https://sony.i.lithium.com/t5/image/serverpage/image-id/7925i082E029E2F42EFAD?v=1.0

Slog2 gives less than 20 tones to the stop 10 below clip (compared to 60 odd for Slog3) .. basically at 20 levels per stop you are 'digging the dirt'

Hopefully your 16bit raw files allocate more data to that stop.

So when you perform a couple of transforms (raw - Slog - Slog - 709) you still are drawing from the raw data pool.

If you take your raw file into Resolve aplly the look (slog 2 or 3) , export an XAVC, then import the XAVC back into resolve and LUT that you have a better comparison.
 
Last edited:
Donny 123 said:
Yes surely logic would dictate that the sole use of Slog 2 or 3.. is for max DR.. if there is a large DR or at night time.. bugger all DR.. why use slog at all.. with its tiny amounts of data in the bottom end.. so it can handle a large DR.. ?
Click to expand...

THere is dark and dark.. dark with a (car) headlight is a very wide DR scene.. wider DR than a headlight on in the day time.
 
Doug Jensen said:
Why is it meaningless? The camera used to shoot the S-LOG footage should not matter, and the advantage of RAW on the F55 is that I can grade the clips in post both ways -- with no harm done.
Click to expand...

Point is, depending on your shot, in-camera you might expose Slog2 and Slog 3 differently.
 
You must log in or register to reply here.
Back
Top