The risks are high you will have scene clipping if you don't use log.
10 always beats 8.
FS7: The Great S-Log-2 vs S-Log 3 Debate
- Thread starter karma17
- Start date
The risks are high you will have scene clipping if you don't use log.
10 always beats 8.
morgan_moore
Major Contributor
You always have 1024.
not 10 8 or 14
Spreading it over 10 when you have an 8 scene will not beat it - it will be worse.
There seems to be a consistent ignoration of the concept of recording more stops means less data per stop.. which means less tonality per stop.
I know Im creating an argument here for baked codecs..In theory yes but in reality I like some wiggle room and dont actually like the shape of those curves.
not 10 8 or 14
Spreading it over 10 when you have an 8 scene will not beat it - it will be worse.
There seems to be a consistent ignoration of the concept of recording more stops means less data per stop.. which means less tonality per stop.
I know Im creating an argument here for baked codecs..In theory yes but in reality I like some wiggle room and dont actually like the shape of those curves.
Correct. Which is what I was saying - or trying to say anyway.
Many things in nature don't present us with a scene of a DR of 14 stops or less. That sometimes might mean exposing for the dark shadows thereby blowing out the sky - and reversely could be crushing the shadows into black when we want all the hues in the sky while shooting in a low-lit interior toward a window showing daylight. So in those extremes we're allocating bits in the most important details we want shown in final delivery. And we're doing that shooting 8 bit or 10 bit with or without using log. But don't we need to know how many of those stops might be "seen" in the delivery method before we decide where to allocate the most bits? And then we might know more if there's an advantage to using one of the slog's over the other - especially if the chart Sam presented several pages back is correct?
That's right. When I shot at night I would get the cleanest image with STD gamma in custom mode at 2500 ISO with NR=High. About 99% of the scene was I don't know exactly, let's say 3-5 stops or less. But in targeting that range with STD gamma, specular highlights, were clipped. Maybe they would have been anyway. We sometimes don't realize how much DR is really there in nighttime scenes when considering peak to peak instead of average values. So to finish, 2500 ISO Custom w/high NR would concentrate the most data over the fewest average stops...as long as you weren't counting the specular peaks from LEDs, neon lights, tungsten etc. But I had to give up something in the bargain, the color saturation of those specular highlights. You cannot underestimate what they contribute to the perception. So I was torn between capturing more of the dynamic range and thus color volume of specular highlights with log gamma versus cleaner shadow areas over broader swaths which comprised the greater part of the frame. So yes, I have been in the mix when it comes to decisions about data per stop. But I have also realized that some of those assumptions about a scene being a low contrast, low DR scene are not always true, particularly at night where there is the potential for a very wide DR any time there is a point source of light, or in drama.
I think it's useful, and by now well understood; there are two log curves with the same DR, one with more data focused in the midrange and below, the other toward the extreme highlights. I love your work btw. :thumbup:
morgan_moore
Major Contributor
Yep trade offs. And my personal opinion (back to topic) is that slog2 is the best trade off I know in the fs7 without majorly painting the camera on a scene by scene basis.. which I dont have the time patience, kit or skills to do. Its just a bit more crunchy than Slog3, with no perceivable loss in useful DR, IMO
I should just add one comment, since I've shot a bit of that, streams under treed canopy with sun shine overhead, that using SLog3 in that situation; while allocating more data to the low key increases shadow detail, it's also capturing the sensor noise in greater detail. It could be a consideration using SLog2 to crush it out, and target the most detail within white waters and beamed sunlight.
Thanks. :thumbup:
allenfacemire
Well-known member
Concerning log 2 vs log 3, I pretty much agree with using log2. With log 3 you pretty much need to correct scene by scene and in my world that just isn't practical. Plus I hate to loose my ISO and balance options. Now that I say that, perhaps that's only when used in conjunction with with cine ei, which again, is something I'm not fond of. Don't like applying luts to VF or monitors to see what the image has the potential to look like.
Again, my work environment may differ from many in that I do network news and long form docs as well as TV shows that really don't warrant that type of video scheme. I might reconsider if I were doing a commercial or feature but...I do TV! In magazine world the editors go ballistic if you give them that kind of work flow. They want baked in all day,every day. With Log2 you can kind of batch correct. A bit more forgiving IMO!
But again, opinions are like orifices. Everybody has them.
Doug Jensen
Carbonite Member
If someone wants to bake-in a look they should stay away from any gamma that uses LOG and any kind of MLUT. You can get a much nicer, refined, beautiful picture simply using the one of the Hypergammas, a matrix, setting the master black level, and making other changes to the traditional Sony paint menus that have proven themselves on dozens of cameras already. Anything that can be done by baking in an MLUT can be done better with the regular paint menus. I guarantee it.
Run&Gun
Veteran
This is how I've spent most of my career working, because like a lot of others, including Doug, I started when there was no such thing as "log" or "RAW" and "fix it in post" was a tongue-in-cheek joke.
I'd much rather shoot with the look baked in for the vast majority of what I do. There's something very satisfying and actually motivating about looking at a great looking image while you're shooting and not having to just "Imagine" what it's gonna be like after post. One of the reasons, too, that I really like our color VF's over our old B&W's...
I have a few friends that, for some reason, like shooting in CUSTOM with s-log 2 gamma applied. It doesn't make the image anymore "film-like". Just flatter and more boring to me. It's still gonna have to be graded(they say it doesn't have to be graded "as much", but to me graded is graded. In for a penny, in for a pound...). My opinion, either bake-in a look of how you want it to air or if you're gonna have to grade it, shoot it in "real" s-log.
morgan_moore
Major Contributor
I very much admire people who have the skills (and gear) to shoot baked.
But it is not the way (apart from for live)
Shooting flat.. which gives you 'headroom' (recording a range of DR beyond 709) is much more sensible.
'flat' could be raw, log or anything really.. but it is head (and foot) room that counts.
Ill try and outline my reasons..
-in the future you may have to re-deliver the footage for a wider delivery space.
-unless you have a black easy up and calibrated 4k 24in monitor on set you probably cant actually see the finer nuances of what you are doing - tweaking the nuances in a comfy post room under little time presure is better than tweaking with a large costly crew on downtime.
-the world has more than six stop 709 range (unless you have a truck of HMI to control ratios). If you want to shoot baked you will need to repaint your camera scene to scene to best match the camera to the scene DR.
-if you traditionally control highlights (like a sky) by using ND grad.. well you only have the shape of the grad. Consider a landscape scene with a church tower.. I bet you dont have a grad with a little slit for the church tower. Pulling down a sky using not a physical grad but a grad in post you can shape around the church tower.. and track the digital grad to match any camera moves.
--
Shooting log and using a basic 709 LUT should have no impact on your 'traditional' on set experience, that basic 709 lut should be applied in post with one click.. so it should minimal impact on a traditional post experience (no grading)
So if you shoot log you can operate entirely traditionally if you have one LUT in place.
But you can also sleep better knowing that if you need to swing the colours, bring down a sky or deliver to a future wide DR spec you have in the background captured all the data you need.
If on set you choose to toggle that LUT on and off you then can apply imagination to where that footage could go beyone the basic application of the LUT should that be required.
--
Many DPs who go for a look on set (Deakins et al.) Work in that fine environment where they do indeed have the easy up, the calibrated monitor, and a truck (or three) of HMI to control thier ratios and maybe even time to wait for the sun to move.
The low end DPs who then try and 'do it like Deakins' but without the resources fall over. Of course deakins has been shooting raw since leaving film so actually he did go home with a load of headroom.. he just rareley needed to use it.
--
Fixing it in post may be looked down upon but 'finishing in post' is an art that goes right back to the ealiest capture of images with Ansel Adams et al. Adams work shot on 10.8 plates still makes an Arri 65 look like small chip pixelated mush- 100 years on!
https://www.alanrossphotography.com/ansel-adams-zone-system/
https://petapixel.com/2013/09/12/marked-photographs-show-iconic-prints-edited-darkroom/
Last edited:
Run&Gun
Veteran
And it also depends on the market segment you’re working in at the moment, too. The majority of what I’m doing is for network TV, thus a quick turn. Sometimes there is no time for any type of correction, even a one-click basic LUT. I know it sounds ridiculous, but those are the facts. There are times that even sit-down interviews for pieces not intended to run until several days later have sound bites pulled from them to run shortly after being sent in, because the subject may have said something “important”.
Don’t get me wrong, I hate the philosophy/mantra of “good enough” and I’m always an advocate of getting the best image and archive possible, but for a lot(not all)of what I do(lit interviews with a shelf life of 3-7 days), log will net nothing(or almost nothing) in the end.
Don’t get me wrong, I hate the philosophy/mantra of “good enough” and I’m always an advocate of getting the best image and archive possible, but for a lot(not all)of what I do(lit interviews with a shelf life of 3-7 days), log will net nothing(or almost nothing) in the end.
allenfacemire
Well-known member
Run&Gun is spot on. That's my environment as well. Bakey Bakey. That said, log2 left untouched, actually looks pretty good, particularly when you realize that shooting for networks news/magazine/long form,etc., that by the time it gets pumped and re-pumped through the various feeds, up/down links, transmission, etc., the picture can sometimes build contrast, which for untouched log2 can be an asset. But I have also found a general overall correction works just fine; and I'm not qualifying that with "good enough for TV", which many times it is.
allenfacemire
Well-known member
Not sure what your issue is with my personal assessment Mr. Roper. Pretty condescending at best. Agree or disagree. Fine. Posting brussel sprouts is not helpful to the dialog.
Not too interested in engaging but I started shooting film professionally in 1964. Only thing we worried about was 16 or 35. Color or B&W. Corrections were done at the engineering level. I learned a lot from the engineering dude,who show me the scopes and how they, the engineers, can manipulate (correct) an image that may have been over or underexposed. You learn quick enough which end of the spectrum is more salvageable...over or under.
All of that said, I still stand behind my brussel sprouts.
Not too interested in engaging but I started shooting film professionally in 1964. Only thing we worried about was 16 or 35. Color or B&W. Corrections were done at the engineering level. I learned a lot from the engineering dude,who show me the scopes and how they, the engineers, can manipulate (correct) an image that may have been over or underexposed. You learn quick enough which end of the spectrum is more salvageable...over or under.
All of that said, I still stand behind my brussel sprouts.
Doug Jensen
Carbonite Member
Holy crap! I thought I was the old man on the block and I didn't get started until 1980. My hat is off to you.
I did a 2 camera interview with the FS7 and my Lumix G7 And I used slog3 in custom. Whitened both camera's and they matched perfectly.
For me slog3 seems easier, but maybe that is because I am used to it.
I tested with slog3 and slog2 (outside) and after color-correcting it in Davinci Resolve I couldn't see any difference.
For me slog3 seems easier, but maybe that is because I am used to it.
I tested with slog3 and slog2 (outside) and after color-correcting it in Davinci Resolve I couldn't see any difference.
That was Rodney Strong 2013 speaking. I'll have a talk with him, sorry.