Which input color space & gamma to use for these Sony files?

In front of me I have number of files that came out of Sony A6600 using various picture profiles and I need to figure out how to correctly select their input color space / gamma in Resolve. My questions are:

- For set of files taken with ‘picture profiles’ using BT2020 + HLG/HLG1/HLG2/HLG3 Sony’s metadata XML says they are:

CaptureGammaEquation = rec2100-hlg
CaptureColorPrimaries = rec2020
CodingEquations = rec2020ncl

and MediaInfo for them says:

Color range : Limited
Color primaries : BT.2020
Transfer characteristics : HLG
Matrix coefficients : BT.2020 non-constant

So input space for one recorded with BT2020 HLG should be set to Rec2020, but what should be input gamma, Rec2100 HLG or Rec2100HLG (Scene)? I am guessing Scene one but I’m afraid I could be incorrect due to my lack of knowledge.

Also, what input gamma should be used for ones recorded with Sony’s HLG1/HLG2/HLG3 settings?

- For set of files taken with ‘picture profiles’ using SGamut3.Cine + Cine1/Cine2/Cine3/Cine4 Sony’s metadata XML says they are (in case of Cine2 gamma):

CaptureGammaEquation = ex-cine2
CaptureColorPrimaries = s-gamut3-cine
CodingEquations = rec709

and MediaInfo for them says:

Color range : Limited
Transfer characteristics : BT.709

and nothing else.

So input color space should be, obviously (unless I’m missing something), SGamut3.Cine but what should be input gamma? Resolve doesn’t offer Cine1/2/3/4 gammas and Rec709 value for CodingEquations confuses me, does it mean I should be using Rec709 as input gamma?

- For set of files taken with ‘Pro color mode’ and Cine2 gamma Sony’s metadata XML says they are:

CaptureGammaEquation = ex-cine2
CaptureColorPrimaries = rec709
CodingEquations = rec709

and MediaInfo for them says:

Color range : Limited
Color primaries : BT.709
Transfer characteristics : BT.709
Matrix coefficients : BT.709

Does this confirm I should use Rec709 as input gamma when file has been created using Sony’s Cine2 gamma?

Also, what in the world is ‘Pro’ color mode? I am not aware of such color space. Does metadata that says file has BT.709 primaries mean this is just look Sony has baked in within Rec709 color space even though Sony in A6600’s manual bundles it together with Rec709, Rec2020 and other color spaces under ‘Color Modes’ section?

- And last, but not least, for set of files taken using Sony’s ‘creative styles’ (which look off to my eyes once imported into Resolve or played with other players) Sony’s metadata XML says they are:

CaptureGammaEquation = rec709-xvycc
CaptureColorPrimaries = rec709
CodingEquations = rec709

and MediaInfo for them says:

Color range : Limited
Color primaries : BT.709
Transfer characteristics : xvYCC
Matrix coefficients : BT.709

I’ve never heard of xvYCC before. Google says it is Sony’s “extension” of Rec.709. Could that be why they look off? If yes what is the correct way of making them look natural?

I need to figure out how to correctly select their input color space

I'm not sure that's something that can be "figured out". I think that's something you need to know.

Which means if you didn't shoot them, you need to ask whoever did.

HLG1,2,3 variants all have the same gamma curve (or are supposed to?), so the setting for gamma should just be HLG.

Rec 2100 HLG (Scene) would (I think) be the correct input setting. Rec 2100 HLG is meant for output (only). You may have to change this manually in Resolve, if it selects Rec 2100 HLG automatically as an input gamma.

For the slog footage, I would try selecting the correct slog version (1,2,3) as the input color space and not specify a gamma setting (you're not forced to set both, it's an option), and trust to the system to make the correct transform. Maybe someone who works with this footage can give you more information.

Don't know what "pro color mode" is, can't advise.

I need to figure out how to correctly select their input color space

I'm not sure that's something that can be "figured out". I think that's something you need to know.

Which means if you didn't shoot them, you need to ask whoever did.

It seems you either misread and/or misinterpreted what I wrote or I miscommunicated it. I'm the one that shot them and I have listed settings they have been taken with when saying "for set of files taken with ...". Based on that info and metadata I need to figure out which input color spaces and gammas to use because for some of them Sony doesn't seem to provide all the information.

HLG1,2,3 variants all have the same gamma curve (or are supposed to?), so the setting for gamma should just be HLG.

Rec 2100 HLG (Scene) would (I think) be the correct input setting. Rec 2100 HLG is meant for output (only). You may have to change this manually in Resolve, if it selects Rec 2100 HLG automatically as an input gamma.

Thank you for your reply! Yes, Resolve automatically setting it to Rec2100 HLG, not HLG (Scene), is what threw me off to start wondering (due to my lack of knowledge) should I use HLG, not HLG Scene like I used to.

Moments ago I came across this article https://xtremestuff.net/sony-and-hybrid ... -hlg/about HLG by Asim Siddiqui (aka @RikshaDriver on this forum) that provides lots of value and confirms (if I am interpreting it correctly) one should use HLG curve for all HLG variants.

For the slog footage, I would try selecting the correct slog version (1,2,3) as the input color space and not specify a gamma setting (you're not forced to set both, it's an option), and trust to the system to make the correct transform.

I haven’t used S-Log gammas, I have used Cine1/Cine2/Cine3/Cine4 gammas which are different and ones I am trying to figure out.

I've never used a Sony camera, so that may be my disconnect here. But, you don't really 'figure this out'. You choose the correct Input based on how it was shot.

There are only four options for Sony.

Sony Input Color.png (46.68 KiB) Viewed 32306 times

I've never used a Sony camera, so that may be my disconnect here. But, you don't really 'figure this out'. You choose the correct Input based on how it was shot.

You keep focusing on choice of few words I used instead of what was meant and later clarified. If that will help replace "figure out" with "what are correct inputs based on these shot settings".

There are only four options for Sony.

Sony Input Color.png

Yes, I know, but shots haven't been taken in those color space + gamma combos. My question is for shots that were taken with settings I listed.

As far as I can tell, cine1,2 and 3, are picture profiles which conform to rec. 709.

However, SGamut3.Cine would appear to be a log profile. So input color space would slog3 and gamma sgamut3.cine. But I'm not a Sony shooter.... Somebody here probably is.

However, SGamut3.Cine would appear to be a log profile. So input color space would slog3 and gamma sgamut3.cine.

Isn't it the other way around, SGamut3.Cine is color space and SLog3 is gamma?

https://www.sony-asia.com/electronics/s ... s/00145905

Yes, you're right that SGamut3.Cine is a color space. Slog3 is a log format, but also used as an input for color space. I can't tell you how to handle this footage, unless the options in Resolve lead the way..

Yes, you're right that SGamut3.Cine is a color space. Slog3 is a log format, but also used as an input for color space. I can't tell you how to handle this footage, unless the options in Resolve lead the way..

Thank you, hopefully those with more experience with Sony files will pitch in ...

Regarding Sony HLG1, 2 & 3, I will quote directly from the write-up:

Whilst the HLG3 profile is practically the same as HLG BT.2100 (which means that a standard HLG BT.2100 profile can be used for footage normalization), the HLG2 and HLG1 profiles are not standard as they’re limited to 95% and 87% of the full signal level respectively. When these HLG profiles are ingested and normalized based on the HLG BT.2100 specification, the footage is incorrectly normalized at lower levels resulting in an overall darkened scene – the gamma curves are simply not the same.

For conforming HLG BT.2100 footage in Resolve, please have a read:

https://xtremestuff.net/recording-editi ... ma-part-2/

Note: The issues with HLG & Input DRT still apply now as they did when I wrote this 2 years ago.

For other footage which has custom Cine transfer functions using S-Gamut3.Cine color space,

You can separate the "color space" and "gamma" in the project settings to give you Rec.709 (Scene) Gamma and S-Gamut3.Cine color space.


Since there is no data available regarding the Cine transfer functions, you have to assume with Rec.709 (Scene) and approximate the curves for correction.

Second that, there are no settings for those 'Cine' profiles, which are a Sony thing. I consider them Rec709 color with specific gamma curves, some stretching the shadows, others improving the highlights. A simple curve should get them under control.

Regarding Sony HLG1, 2 & 3, I will quote directly from the write-up:

Whilst the HLG3 profile is practically the same as HLG BT.2100 (which means that a standard HLG BT.2100 profile can be used for footage normalization), the HLG2 and HLG1 profiles are not standard as they’re limited to 95% and 87% of the full signal level respectively. When these HLG profiles are ingested and normalized based on the HLG BT.2100 specification, the footage is incorrectly normalized at lower levels resulting in an overall darkened scene – the gamma curves are simply not the same.

For conforming HLG BT.2100 footage in Resolve, please have a read:

https://xtremestuff.net/recording-editi ... ma-part-2/

Note: The issues with HLG & Input DRT still apply now as they did when I wrote this 2 years ago.

Hi Asim, thank you for helping again, hope you’ve been doing well

I will definitely have to read & digest your articles as I’m trying to use HLG when shooting (not to make HDR deliverables, at least not right now, but to get more info to work with for SDR work).

In the meantime, so if I stick to using just HLG or HLG3 on Sony I can trust they will match same specification Resolve is using? And if Sony’s HLG and HLG3 are both same do you happen to know what would be the reasons for Sony to have them both, are there any pros/cons between them?

For other footage which has custom Cine transfer functions using S-Gamut3.Cine color space,

You can separate the "color space" and "gamma" in the project settings to give you Rec.709 (Scene) Gamma and S-Gamut3.Cine color space.

Since there is no data available regarding the Cine transfer functions, you have to assume with Rec.709 (Scene) and approximate the curves for correction.

So Cine transfer functions are not Rec709 even though Sony’s metadata XML says CodingEquations = rec709? So I should experiment which one of Cine gammas used for recording brings me closest to desired look easiest after I’m done grading it?

Also,

1) Do you happen to know do Sony files with one of Cine gammas use full or video data levels?

2) If I manage to get recordings with full data levels (say by using external recorder, like Atomos Ninja V+) do you happen to know would that result in visible increase in quality?

Thank you again!

Sony is using neither strictly full range nor video range for those 'Cine' profiles, instead they record some 'super whites', which you can easily pull back when grading. These can give you about 2/3 of a stop.

And then, to get the most out of your sources for grading if your output is Rec709, you don't need to use any kind of HDR profile when shooting. Slog-3 will give you plenty of room for grading and it's available as a standardised setting in color management. The most important point is recoding in a 10 bit codec.

Sony is using neither strictly full range nor video range for those 'Cine' profiles, instead they record some 'super whites', which you can easily pull back when grading. These can give you about 2/3 of a stop.

That would explain what I was seeing. When I would pull down/up file that I have ETTR, and attributes are set to auto or video data levels, I would see that there is more data, making me think file might have full data levels after all, but when I would change attribute to full data levels I could never make it to look natural, it always looked off, no matter what I tried.

And then, to get the most out of your sources for grading if your output is Rec709, you don't need to use any kind of HDR profile when shooting. Slog-3 will give you plenty of room for grading and it's available as a standardised setting in color management.

I’m staying away from SLog3 as far as I can. I tried it and it is a pain in the back for me, things go horribly wrong way too easily and it doesn’t seem to borrow itself well to “run and gun” situations. So I figure I need to learn how to walk … walk fast … run fast well first before I try running with the wolves.

The most important point is recoding in a 10 bit codec.

Not having 10-bit is one of the things I suspect is impacting me but I am trying very hard to find out can I get by with 8-bit footage from A6600 cause spending more money is the last thing I need right now.

Thank you again for your continued help!

In the meantime, so if I stick to using just HLG or HLG3 on Sony I can trust they will match same specification Resolve is using? And if Sony’s HLG and HLG3 are both same do you happen to know what would be the reasons for Sony to have them both, are there any pros/cons between them?

HLG3 has slight tweaks such as lifted shadows but you can treat it as bog standard HLG for correction purposes.

The only thing to consider here is that you're exposing correctly.

The production practice is to expose at 38% IRE, but when you use the BT.2100 (Scene) input preset, this will result in an over-exposed clip. You can confirm this by recording a Mid-Gray chart and verifying the linear signal output value in Resolve.

If you're using BT.2100 (Scene) for input and you intend on working in a scene referred color managed workflow, then your camera's exposure should be targeted at 21% IRE.

You should have given us that information from the start! Yes, if you have only 8 bit, neither use log nor HLG. In this case those Cine profiles are a good choice. Before I got a better camera, I was using Cine 4 in most cases.

shots haven't been taken in those color space + gamma combos.

Ahhh, that was my misunderstanding.

I had assumed that Resolve would include all the possible Sony options. I didn't realize the camera had others options not yet included in Resolve.

You should have given us that information from the start! Yes, if you have only 8 bit, neither use log nor HLG. In this case those Cine profiles are a good choice. Before I got a better camera, I was using Cine 4 in most cases.

What a coincidence, I’ve been playing with various profiles over last few days and so far I like results from Cine3 and Cine4 gammas in Pro color mode most. Which color space/mode you used Cine4 with?

Any tips when exposing with Cine4, please? What you recommend for zebra settings? What I should expose skin for?

Also, if you don’t mind, which camera you use now and which settings you recommend with it?

P.S. I wasn’t mentioning specifics of the camera because I was trying to figure out which color spaces and gammas to use in Resolve for certain Sony settings, not which profiles would be best for camera I have. I would have same challenge and question what to use in Resolve if one day I get Sony camera that supports 10-bit and I used same profiles so I still need to figure that out.

In the meantime, so if I stick to using just HLG or HLG3 on Sony I can trust they will match same specification Resolve is using? And if Sony’s HLG and HLG3 are both same do you happen to know what would be the reasons for Sony to have them both, are there any pros/cons between them?

HLG3 has slight tweaks such as lifted shadows but you can treat it as bog standard HLG for correction purposes.

The only thing to consider here is that you're exposing correctly.

The production practice is to expose at 38% IRE, but when you use the BT.2100 (Scene) input preset, this will result in an over-exposed clip. You can confirm this by recording a Mid-Gray chart and verifying the linear signal output value in Resolve.

If you're using BT.2100 (Scene) for input and you intend on working in a scene referred color managed workflow, then your camera's exposure should be targeted at 21% IRE.

Thank you again for very valuable guidance!

shots haven't been taken in those color space + gamma combos.

Ahhh, that was my misunderstanding.

I had assumed that Resolve would include all the possible Sony options. I didn't realize the camera had others options not yet included in Resolve.

Yes, unfortunately they don't which is a headscratcher because with all the people out there using Sony for videos one would think it is in Sony's and BMD's best interest to get together and make sure user can easily get correct translation for any of Sony's color space/mode & gamma settings, not just ones that are notoriously hard to use even for non-beginners and thus many stay away from.

There are far too many presets in consumer cameras to accommodate in applications like Resolve. And most of those choices are some flavor or other of rec. 709 anyway, which don't require transforms because they're expected to be presented more or less "as is".

Your own work might be simplified by selecting one of the Sony slog options and just sticking with it, using color management to normalize the footage. My few encounters with slog weren't encouraging, but not much time went into it, and and if that's what your shooting, learn to deal with it.

There are far too many presets in consumer cameras to accommodate in applications like Resolve. And most of those choices are some flavor or other of rec. 709 anyway, which don't require transforms because they're expected to be presented more or less "as is".

Personally I would argue that while one shouldn't chase every single possible thing out there one should cover what majority of market out there is likely to use, not just ones that are hardest to use and thus most likely to result in failure. And quite a few people out there use Sony, avoid SLog, and Cine1-4 can be found in A7SIII, A7IV, FX30, FX3 ...

When you make effort to accomplish things too high you lose not just current but also future customers. But if you make it easy/easier and offer steps in between you are more likely to build satisfaction and thus loyal base. Some of them might move onto SLog but some of them never will and they will still be happy. Can you imagine what would happen to Resolve's customer base if one could say "Here is a quick and easy way to get great and correct results out of your Sony without having to pull your hair out getting a PhD in SLog shooting and grading"?

reading things as "avoid slog" Isn't working in these log profiles easier than ever with all the cst options? We work with multiple sony cameras almost every day, and i always ask for log for almost all kinds of projects. As long as i know which exact flavour (which is mostly specified in the xml file). Also when i need to match them with drones, actioncams, RED, whatever, it's a breeze to have it all in log.

reading things as "avoid slog" Isn't working in these log profiles easier than ever with all the cst options? We work with multiple sony cameras almost every day, and i always ask for log for almost all kinds of projects. As long as i know which exact flavour (which is mostly specified in the xml file). Also when i need to match them with drones, actioncams, RED, whatever, it's a breeze to have it all in log.

I'm sure that is true for experienced professionals like you. But for beginners like me ... I've heard many say shooting and grading SLog3 keeps kicking their behind too, especially with 8-bit footage.

Second that. Slog3 is particularly problematic with 8 bit, since it’s very flat. In many situations without extreme contrast you will be using only 4 or 5 bits of the video range. Yikes!
With enough bit depth it’s a good choice, better than Slog2, which is more difficult to get right when grading. Please remember that 10 bit is not 20% better, but 300%.
Now to those profiles. Every manufacturer makes their own, and regularly adds a few more or tweaks them. Impossible to chase. And then, they are not intended for grading, they are rather like Rec709 with a ‘look’ added. Simply choose what you like. I like Cine4 with Pro Color. You can choose to set Zebras to 100% and ETTR. Watch out, though, not all profiles have Superwhites. Or set zebras to around 70 for skintones.

I like Cine4 with Pro Color.

Thank you for confirming I wasn’t wrong when I felt Pro + Cine4 are giving me best result, at least I got something right

You can choose to set Zebras to 100% and ETTR … or et zebras to around 70 for skintones.

I am already using zebras at 100+ but I’ve seen people recommending 109+ (some 107+) for Cine4 because, they say, Cine4 gives 109% of video levels. Which one of the values would be technically correct, please?

Watch out, though, not all profiles have Superwhites.

Hmmm … {scratching my head} … I’m confused now. I was under impression there is no difference between profiles when all settings are same. What exactly do you mean by this please? How can I find out if the profile does / doesn’t have superwhites?

Don't overthink it. Just go out and shoot!

That said, Superwhites just mean that you can pull back a bit of detail from seemingly blown out areas.
109% is about right for Cine4, but I tend to use 100% zebras to keep some safety margin – if I use zebras at all. I don't like the flickering, it's distracting. I rather use an exposure meter, which I always carry around with me (it's called iPhone, and the app is Cine Meter II).

109% is about right for Cine4, but I tend to use 100% zebras to keep some safety margin – if I use zebras at all. I don't like the flickering, it's distracting. I rather use an exposure meter, which I always carry around with me (it's called iPhone, and the app is Cine Meter II).

I too find zebras annoying so I mapped their on/off to one of the slots on function key to get them out of the way once I am done evaluating exposure.

I'm trying to figure out how high correct number for zebra is because my playing around with them makes me feel there is a lot of usable space left on the plate when using 100 instead of 109 and once I see zebras kick in I back off third of the stop anyway.

Would you mind clarifying, please, what exactly you meant with "not all profiles have superwhites"?

Not all the Cine profiles have the same level of tolerance to mild overexposure.
This is Cine4, mildly overexposed.

Clipping.jpg (270.34 KiB) Viewed 31708 times

And now pulled back about one stop. In the upper left I got definition of the seemingly blown-out sky and could even discern the ridge of a mountain.

Pulled down.jpg (277.3 KiB) Viewed 31708 times

But since I have a different camera now and Sony is not only improving the sensors, but optimising the profiles for them, I can't do the homework for you. Anyway, you need to know your own camera by heart.

Find a contrasty situation with constant lighting and try for yourself. Set zebra to 100+, increase exposure to the point where zebra sets in, then do some up and down bracketing in thirds of an aperture. Repeat this for the different Cine profiles. Don't forget to take notes of all settings!

Now import them all into DR and check where you really get a flat line for clipping in the highlights, while softly pulling down gain or playing with the curve.

If you own a test chart with a grayscale, you can even observe how the different Cine profiles treat shadows and light, which is their main purpose within the limits of Rec709.

Not all the Cine profiles have the same level of tolerance to mild overexposure.

Oh, so when you said “not all profiles have superwhites” you meant “not all gammas have superwhites”?

But since I have a different camera now and Sony is not only improving the sensors, but optimising the profiles for them, I can't do the homework for you. Anyway, you need to know your own camera by heart.

Find a contrasty situation with constant lighting and try for yourself. Set zebra to 100+, increase exposure to the point where zebra sets in, then do some up and down bracketing in thirds of an aperture. Repeat this for the different Cine profiles.

Now import them all into DR and check where you really get a flat line for clipping in the highlights, while softly pulling down gain or playing with the curve.

I do believe I have to do my own homework with my own gear (hence me doing all these experiments and asking all these questions) and your steering in the right direction is very valuable and appreciated

Doing what you described above is now on my list of things to do. In the meantime I’ve found this video by Gerald Undone that seems to confirm Cine4 has highest latitude for overexposure of all Cine gammas and that zebras for it should be set at 109+:

It will be interesting to see what I find.

If you own a test chart with a grayscale, you can even observe how the different Cine profiles treat shadows and light, which is their main purpose within the limits of Rec709.

I own ColorChecker Passport Video, if you can recommend something with better grayscale that is inexpensive I will appreciate it.

For a pure grayscale, there are relatively cheap charts by Kodak for photographers.

reading things as "avoid slog" Isn't working in these log profiles easier than ever with all the cst options? We work with multiple sony cameras almost every day, and i always ask for log for almost all kinds of projects. As long as i know which exact flavour (which is mostly specified in the xml file). Also when i need to match them with drones, actioncams, RED, whatever, it's a breeze to have it all in log.

I'm sure that is true for experienced professionals like you. But for beginners like me ... I've heard many say shooting and grading SLog3 keeps kicking their behind too, especially with 8-bit footage.

Hmm well i do have some experience, but i don't consider myself a real, professional color grader. I learned a lot from people on youtube, for example Cullen Kelly. He has some video's where he explains the color management part in a a way which is really easy to understand. Now that i checked again, he specifically has a video about slog. Didn't watch it yet, but here's a link

Which is all very good advice, like I expect from Cullen Kelly.
But serious Slog is coming out of a camera with at least 10 bit.
Anything less is just marketing, like "Look, Mom, my camera has log like the big guys!".

But serious Slog is coming out of a camera with at least 10 bit.
Anything less is just marketing, like "Look, Mom, my camera has log like the big guys!".

that's very true

reading things as "avoid slog" Isn't working in these log profiles easier than ever with all the cst options? We work with multiple sony cameras almost every day, and i always ask for log for almost all kinds of projects. As long as i know which exact flavour (which is mostly specified in the xml file). Also when i need to match them with drones, actioncams, RED, whatever, it's a breeze to have it all in log.

I'm sure that is true for experienced professionals like you. But for beginners like me ... I've heard many say shooting and grading SLog3 keeps kicking their behind too, especially with 8-bit footage.

Hmm well i do have some experience, but i don't consider myself a real, professional color grader. I learned a lot from people on youtube, for example Cullen Kelly. He has some video's where he explains the color management part in a a way which is really easy to understand. Now that i checked again, he specifically has a video about slog. Didn't watch it yet, but here's a link

Thank you for the pointer! I've seen some of Cullen Kelly's videos on other topics and I liked them a lot but I haven't seen this one, I will have to add it to my watchlist.

For a pure grayscale, there are relatively cheap charts by Kodak for photographers.

Is this the one you have in mind?

https://www.bhphotovideo.com/c/product/ ... e_and.html

Yep.

Yep.

Cool, that is very affordable! Thank you!

I’ve found this video by Gerald Undone that seems to confirm Cine4 has highest latitude for overexposure of all Cine gammas and that zebras for it should be set at 109+:

It will be interesting to see what I find.

In the case somebody one day ends up reading posts above when searching for information on how to determine what zebra setting should be used with certain gamma and what might be best practice when exposing with that zebra setting, I’ve done some research and experimenting over last few days with my Sony A6600 and found that:

1. Zebras will not work (show up on screen) if setting for zebra is higher than what it should be for current gamma. So to determine what zebra setting value is maximum for gamma you are interested in set zebra to highest possible value (say 109+) and point camera to surface that is overexposed well beyond what camera can handle. If zebra is not showing up start rolling back values until zebra shows up. That is maximum zebra setting for that gamma.

2. If I ETTR Cine4 with 109+ zebra until it shows up and then back off 1/3 of a stop I’m dangerously close to clipping, it wouldn’t take much to clip a channel. That might explain why some advise to set zebras to 107+ for Cine4, they might be trying to give themselves some headroom. Personally I will be backing off another 1/3 of the stop in an effort to not just stay on the safe side but also further minimize chance of losing detail in highlights.

Hope this helps someone one day

Thank's for the report from testing.
Again, don't overthink it, use your judgment by eyes too. In many scenes you may have highlights, like practicals or strong reflections, which can't be protected from clipping. Or you'd underexpose all the rest. OTOH, nobody wants a blown-out highlight on a forehead, (where some makeup goes a long way).

Those grayscales can be quite helpful to understand the different gamma curves. If your image were linear (which it isn't) you should see a straight 'X' from those. The way that X is getting curved in different areas will show you what is enhanced by the specific profile. It'll always be trying to fit the higher DR of modern sensors into the much smaller DR of Rec709, but stretch or compress different parts of the curve in favour of the other one(s).

Of course, all of this is not Sony's monopoly, all 'looks' or 'profiles' you can choose in hybrid cameras are doing something similar.

... use your judgment by eyes too. In many scenes you may have highlights, like practicals or strong reflections, which can't be protected from clipping. Or you'd underexpose all the rest. OTOH, nobody wants a blown-out highlight on a forehead …

Of course

Since my post I’ve discovered dangers of listening to tech videos that are strictly “going by the numbers” and using ETTR (at least when using gammas that Resolve doesn’t know how to handle, like Cine4).

When I took two shots with Cine4 under same lighting using 1/60 of a second for first one (after ETTR) and 1/125th for second one, and tried to match exposure between two of them afterwards (bring down first one or bring up second one) result was a headache/failure.

For start, even though there is a stop of difference between the two using +/- 1 on either Asim Siddiqui’s aka @rikshadriver Exposure Plugin or HDR palette’s global exposure was resulting in huge exposure difference between the two, it would take around 0.3 to make them approximately match in exposure.

Also, even after matching exposures look between the two was noticeably different, starting with where the grays were, and unnatural. It was taking lots of pulling and pushing to make it semi-decent but still not what I would consider acceptable.

However, I didn’t have such problem when working with shots that were taken under same exposure settings with profile that Sony claims matches Rec.709 standard (PP4 on A6600). Exposure, grays, and look would all line up almost perfectly after I used +/1 in either HDR palette’s global exposure or Asim’s Exposure plugin, and natural, things were behaving in natural, consistent and predictable way.

I’m guessing this might be due to Cine4’s gamma curve that Resolve doesn’t know how to handle as I am adjusting exposure, and this article seems to confirm my feeling in section that talks about “un-natural image response area”

https://www.xdcam-user.com/2011/01/unde ... and-s-log/

Searching further after this I found people recommending Cine4 should be actually underexposed by approximately half of a stop, not ETTR, in order to avoid placing values on part of the curve used to protect highlights. I will test that advice but I doubt this is the correct way to approach it, I feel it is just mostly avoiding issue by avoiding to place values on the part of the curve that is not matching Rec709 but that it isn’t avoiding it completely, that one would still have end up with issues in certain areas after exposure adjustments push values into them. Plus this might result in chain reaction by too many values cramped in smaller area resulting in noise and color shifts once they are pushed up, especially with 8-bit color.

So short term wise I will have to try other gammas to see which one brings me closest to goal fastest with least amount of headaches and I’m having a feeling it will end up one of those that are matching Rec709 (closer). Long term wise I’m having a feeling that for anything better I will have to stick with one of the gammas Resolve knows how to handle (whether HLG or SLog variant).

In either case I’m feeling it is becoming painfully obvious 8-bit files are leaving me very very little headroom to work with, even when working with non-log, and that I will end up having to purchase 10-bit capable camera sooner than I hoped. Only question is which one, A7IV or FX30? I keep seeing contradicting reports which one is better image quality wise so if anyone has input it will be appreciated.

Those profiles are not really meant for grading. They should be exposed as close as possible to the final look you are aiming for. In my experience, you can pull-down the highlights a bit, getting some more DR.
Apart from minor color corrections, that’s about all.
You can ask me about the A7IV, but not the FX30.

Those profiles are not really meant for grading. They should be exposed as close as possible to the final look you are aiming for. In my experience, you can pull-down the highlights a bit, getting some more DR. Apart from minor color corrections, that’s about all.

Yes, that is one of the reasons I said I’ve discovered dangers of listening to YouTube tech videos etc., number of them talk about “grading” Cine4, and also talk about ETTR in almost same breath, when I’m now discovering that what I have in front of me are mainly pre-baked looks that leave very little wiggle room. Some of them might be better suited for certain type of scenario than the other but that’s it.

You can ask me about the A7IV, but not the FX30.

Thank you for the offer, I will reach out if I have any questions when time for that decision comes.

Well, of course you can grade them, if you like the results. After all, we are making video for humans, not for scopes. It’s just that the are only 8 bit too.

Well, of course you can grade them, if you like the results.

Crucial word here being IF. With so many of videos I watched that talk about grading cine this or log that having either saturated or washed out orange/pink/teal/whatever look it is impossible for me to say is that because creator liked that look or it is because there will be a problem when following their advice that it is (un)intentionally hidden behind “I like this look”. Considering so many of them accompany their video with “you can buy my LUTs here” and when you look at what they offer it is practically always “creative” LUTs and never technical LUTs I am starting to suspect it is later rather former.

I guess it is starting to make sense to me why Leeming’s LUT for Sony has only Cine2 out of all Cine gammas and why he keeps emphasizing to use 91+ zebra ETTR for it or otherwise result will be unpredictable.

Me, majority of time I like accomplishing real life like look. I believe I should be able to accomplish that before I start applying “artistic” looks, otherwise I will just keep hiding behind “I like this artistic expression of myself” and never learn how to do things right way. Any klutz can slap an "artistic LUT" and say "just call me Picasso and call it a day".

Well, of course you can grade them, if you like the results.

Crucial word here being IF. With so many of videos I watched that talk about grading cine this or log that having either oversaturated or washed out orange/pink/teal/whatever look it is impossible for me to say is that because creator liked that look or it is because there will be a problem when following their advice that it is (un)intentionally hidden behind “I like this look”. Considering so many of them accompany their video with “you can buy my LUTs here” and when you look at what they offer it is practically always “creative” LUTs and never technical LUTs I am starting to suspect it is later rather former.

I guess it is starting to make sense to me why Leeming’s LUT for Sony has only Cine2 out of all Cine gammas and why he keeps emphasizing to use 91+ zebra ETTR for it or otherwise result will be unpredictable.

Me, majority of time I like accomplishing real life like look. I believe I should be able to accomplish that before I start applying “artistic” looks, otherwise I will just keep hiding behind “I like this artistic expression of myself” and never learn how to do things right way. Any klutz can slap an "artistic LUT" and say "just call me Picasso and call it a day".

Me, majority of time I like accomplishing real life like look. I believe I should be able to accomplish that before I start applying “artistic” looks, otherwise I will just keep hiding behind “I like this artistic expression of myself” and never learn how to do things right way. Any klutz can slap an "artistic LUT" and say "just call me Picasso and call it a day".

So true, and good attitude to dig deeper than that. BTW, wasn't it Picasso who is cited as saying: "Art is one percent inspiration and 99% transpiration"?

What Leeming is stating there is true for all LUTs, technical or creative. They will give the right results only at the same exposure they were created for. That's why a real transfer function is always better, but developers only spend the time on doing them right for something standardised and documented, like major manufacturer's log versions.

Now, on to real life look: no camera is seeing colours the same as you do (not counting differences in human color vision). Fun fact: my older iPhone SE comes closest to hit the targets from a professional chart on the vectorscope. Not even an Alexa is that close.
Not that I would seriously recommend using an iPhone for filming. But maybe that's why it's so good as a light meter with CineMeter II, even for color temperature. Not as good as a professional instrument, but close enough for something I always got on me. Asking myself what to do if it ever dies on me.

Finally: I'm so sick of even multi-million dollars productions using that orange/teal look. Don't they have any creativity any more? Or is already an A.I. behind it?