GH5 V-Log LUT

OK, I found a couple of issues. First, Resolve wasn't automatically applying the correct Input Color Space to the media. I had to manually change that in the bin. Second, I don't know if this is a bug with 15 or a glitch on my system, but Resolve frequently stops updating the display with my changes, so it looks like nothing is happening. A restart seems to solve that (for a while).

Between those two issues, RCM is looking much better than at first glance.

Having said that, there is still another hurdle that I don't know how to overcome, and the primary reason I don't use RCM - the Max Timeline Luminance. I have noticed that this setting not only changes how the media looks in the scopes, but how it exports and looks on a TV. I have NO idea what my TV is capable of, and no way of knowing what my client's TV is capable of, and so no way of knowing what this setting should be. In fact the whole idea of this setting makes no sense to me at all, given the vagaries of display capabilities.

Input Color Space: Panasonic V-Gamut/V-Log
Timeline Color Space: Rec.2020 Gamma 2.4
Output Color Space: Rec.2020 Gamma 2.4
Timeline to Output Tone Mapping: Luminance Mapping
Max. Timeline Luminance: 300 nits
Timeline to Output Gamma Mapping: Saturation Mapping
Saturation Knee: 0.9
Saturation Max: 1.0

Yes Resolve 15 is buggy in RCM - I'm on contact with technical support over several issues and have now reverted back to using a color space transform FX node on every clip ( which is slow). Also you have your timeline in Rec.2020 but do you have a Rec 2020 monitor to see the footage on as it won't look right on a REC709 screen at all. Ideally you need a monitor on a decklink so it can get a 10bit signal out unless you have a Quadro GPU. Lastly 300nits is way too low for your luminance mapping. Hope this helps as I'm now out of ideas and out of my depth!
 
OK, I found a couple of issues. First, Resolve wasn't automatically applying the correct Input Color Space to the media. I had to manually change that in the bin. Second, I don't know if this is a bug with 15 or a glitch on my system, but Resolve frequently stops updating the display with my changes, so it looks like nothing is happening. A restart seems to solve that (for a while).

Between those two issues, RCM is looking much better than at first glance.

Having said that, there is still another hurdle that I don't know how to overcome, and the primary reason I don't use RCM - the Max Timeline Luminance. I have noticed that this setting not only changes how the media looks in the scopes, but how it exports and looks on a TV. I have NO idea what my TV is capable of, and no way of knowing what my client's TV is capable of, and so no way of knowing what this setting should be. In fact the whole idea of this setting makes no sense to me at all, given the vagaries of display capabilities.

Let's get it working in 709 first. Do exactly the following:

From the media page, right click on the clip and select 'Attribute'. For Data Level choose 'Full'.
From the media page, right click on the clip and select 'Input Color Space'. Choose 'Bypass'.


From the Project Settings Page:

Input Color Space: Panasonic V-Gamut/V-Log Change to 'Bypass'
Timeline Color Space: Rec.2020 Gamma 2.4 Change to 'V-Log/V-Gamut'
Output Color Space: Rec.2020 Gamma 2.4 Change to REc.709 Gamma 2.4
Timeline to Output Tone Mapping: Luminance Mapping Change to 'Simple'
Max. Timeline Luminance: 300 nits Don't need this setting with 'Simple' (above)
Timeline to Output Gamma Mapping: Saturation Mapping OK
Saturation Knee: 0.9 OK
Saturation Max: 1.0 OK
 
OK, I found a couple of issues.
...
...
Max. Timeline Luminance: 300 nits
It's one thing that people do not want to learn color management or understand how HDR works and use options they apparently do not understand.
But then blaming the bad results on the tool makes absolutely no sense.

On the one hand people want to use raw, log, HDR, color spaces etc while at the same time expecting they can use an instant magic do-it-alll-filmic-super LUT that makes everything look perfect.
That's not how it works. If you do not want to spend time color correcting/grading use Rec.709 and be done with it.
 
That's what I tried to do, but the Panny 709 LUT just wasn't looking as good with the 5s media as it did with the GH4 media.
No, I was suggesting to record in Rec.709.

If you record in log, raw, HDR I think you have to invest time in understanding things instead of expecting some instant-magic-film-look LUT.

Resolve is a wonderful tool but if you don't want to invest in learning things (the manual has over 2600 pages) it's becomes like a kite-flyer trying to operate a jumbo jet and after the jet crashes saying that the jumbo jet sucks!
 
If you record in log, raw, HDR I think you have to invest time in understanding things instead of expecting some instant-magic-film-look LUT.

I guess my frustration is that I don't have to do that with the Blackmagic Pocket. The BM LUT included with Resolve does work just that magically, and I was hoping for similar with the GH5s.
 
I think part of where I went wrong was in following the advice below.

https://www.youtube.com/watch?v=DGY14ygjpLU

While it's clear that I still have a LOT to learn, Tom's recommended settings are sufficient that I can shoot tomorrow's wedding in 4K with a rented GH5s with confidence.

Thank you everyone.

I think where the video is wrong is where you were right to question it, the call for timeline to output tone mapping>luminance mapping>300 Nits. Maybe Cary has some thoughts but that sounds to me like display referenced EOTF for ST2084 PQ. You are exactly right, you don't have any idea what the client's maximum nits display is, the blogger seemed to be using the tool with a WFM as a means of highlight control which is not correct for conventional display gamma. You are also correct that RCM is not self identifying input color space for V-Log/V-gamut, as it would for 709, 2020 and even STD ARIB B67 HLG. Resolve is only able to use metadata to make this recognition, metadata that is not standardized for V-Log/V-gamut as for the others, so this is not a flaw in Resolve, simply an omission from h.264 VUI protocols.
 
Jim Simom, I guess my frustration is that I don't have to do that with the Blackmagic Pocket. The BM LUT included with Resolve does work just that magically, and I was hoping for similar with the GH5s

Well you could always sell the GH5S and get the new BM Pocket 4K Camera next month, and use thr BM LUTs. :happy:
Cheers
 
I think where the video is wrong is where you were right to question it, the call for timeline to output tone mapping>luminance mapping>300 Nits. Maybe Cary has some thoughts but that sounds to me like display referenced EOTF for ST2084 PQ.
If your target display is Rec709 SDR then you should not use this feature.

For HDR 10 you would have a typical minimum of 1000 nits so if you want to tone map everything above 1000 (or higher) then consider it, although tone mapping is kind of an auto mode better to manually grade up to 1000 nits instead (if that is in fact your max target).
 
I didn't want to be rude but all those workaround nodes and grading steps are not necessary for properly exposed V-LogL. Calm down and learn.

Sorry - trying to work through my own issues with RCM late at night while offering advice to others is not a good idea.
 
One thing I'm not sure about is the V-LOG profile in Resolve but I'm assuming this is the 'real' profile that applies to their pro cameras and not the 'lite' version that ships with the GH5 cameras and thus may not be quite the same in terms of gamut at least? If you capture in HLG it 'should' have a defined gamut and response curve as it's an international standard like REC709. I was toying with going back to V-LOG as a capture profile but now I'm understanding how to deal with HLG in Resolve a bit better (without LUTS) I think I'll stick with HLG capture.
 
One thing I'm not sure about is the V-LOG profile in Resolve but I'm assuming this is the 'real' profile that applies to their pro cameras and not the 'lite' version that ships with the GH5 cameras and thus may not be quite the same in terms of gamut at least? If you capture in HLG it 'should' have a defined gamut and response curve as it's an international standard like REC709. I was toying with going back to V-LOG as a capture profile but now I'm understanding how to deal with HLG in Resolve a bit better (without LUTS) I think I'll stick with HLG capture.
I agree, HLG is the more sound way to go over V-Log-L in my opinion.
 
I agree, HLG is the more sound way to go over V-Log-L in my opinion.

We are all in agreement on that but unlike V-LogL where there is a straightforward transform to 709, I haven't found that to be the case for HLG, other than just leaving in HLG which is already more or less watchable on 709 if not ideal. It's not a problem for my personal use but for handing off footage to a client, it's not optimal and I wouldn't do it.
 
Back
Top