Are we close? - S-LOG3 to S-Cinetone LUT

[cyvideo]

But still wondering if I understand you to say that the basic Cine -4 curve is still different from Cinegamma-4.? The Sony manual says that Cine 1 & 2 are equivilant to Hypergammas but for Cine 3 & 4 its just got that totally confusing gobblygook about contrast in the shadows and highlights that I've never been able to make any sense of. Maybe it made better sense in the original Japanese?

Yes, you are correct when you refer to Sony's published data re certain Hypergammas relationship to Cine gammas. BUT, and this is a big but, that relates to the older Sony PDW and EX series cameras. This is what Sony said about those relationships:

PDW-700 has HyperGamma (Ver1.2 or higher) such as,

HG1 : 325% D-range is compressed to 100%
HG2 : 460% D-range is compressed to 100%
HG3 : 325% D-range is compressed to 109%
HG4 : 460% D-range is compressed to 109%

These HyperGammas are the same as HDW-F900.

PMW-EX1/EX3 has the same CineGamma as PDW-F335/355 such as,

CINE1 : Same as the HyperGamma HG4
CINE2 : Same as the HyperGamma HG2
CINE3 : Independent setting for EX1/EX3
CINE4 : Independent setting for EX1/EX3


Therefore, please use CINE1 or CINE2 if you want to use the same gamma curve as PDW-700 or HDW-F900.

The thing is the Cine1, 2, 3, 4 that we see in the A7 series measures up quite differently to the Cine1, 2, 3, 4 that we see in the earlier EX series cameras. Which doesn't surprise me, as the dynamic range of the later A7s greatly exceeds the dynamic range of those earlier EX series cameras. And compressing that greater range into our max 7 stop Rec 709 space is going to be different. Sure, in both camera gamma cases there are ranges of 100% and 109% Sure certain Cine gammas are similar in as much as Cine2 in the A7s is limited to the 100% broadcast level as is Cine2 on an EX and Hypergamma 2 on the PDWs. All of these are rated at 325% dynamic range. So yes, there is equivalence there. But using the same gamma curve to compress a 10-11 stop camera and a 13-14 stop camera into a 7 stop 709 space is not going to be the same.

I've been rubbished in the past when I say one of my favorite gammas is S-LOG2 used in Custom mode with an appropriately rolled matrix if you want to deliver nice looking highlights and maintain detail in the shadows. Rubbished with statements like "You can't use S-LOG2 in Custom?". You only have to look to Sony's own FS7 Reference Guide to see why they suggest it in certain lighting situations.

"E2.4 S-Log2
You can use the S-log2 gamma curve in Custom Mode
to capture an extremely large scene brightness range
(14 stops or 1300%). This will result in a very flat on screen
image. The PXW-FS7 is primarily designed to use S-Log
recording in the Cine EI mode, but the option is also
offered in Custom Mode to help deal with particularly
difficult lighting situations."

S-LOG2 tops out at 14 stops. If you have a camera that delivers less than 14 stops, the best data distribution of the camera's dynamic range is delivered by S-LOG2 not S-LOG3. S-LOG3 is well capable of 16 stops. But that 16 stop bucket has wasted space if you are only going to fill it with less than 14 stops. Seeing the highest measured dynamic range under test of any Sony camera is 13.2 stops (according to CineD) out of the Venice 2 one can rest assured that S-LOG2 on the cameras that still offer it is a good option when used in Custom mode to capture and compress anything up to 14 stops into a 709 7 stop bucket. Alister Chapman produced a good explanation on the merits of the different LOG curves some years back: Some years ago when I tested out Sony's 18-110 f/4.0 zoom in my test street I used my own ready rolled S-LOG2 Custom matrix and have used it often on jobs where generally it's had a happy reception on the look of the images. A UHD example of that can be seen below, performing in bright sunny shadow conditions. A big plus is that Custom S-LOG2 allows white balancing.

BTW is there a good LUT for Slog3 to from the FX9 to match s-cinetone?

Not that I'm aware of. But you are welcome to try out the one I have made for jobs where I need to match S-LOG3 to S-Cinetone. How it will go with the FX9, I can't say. How a particular S-LOG3 image has been exposed will have a baring on that. I designed this to work with S-LOG3 with peaks at 75 IRE. Which is about 1.5 stops above Sony's recommended 51 IRE peaks. And this is to match S-Cinetone images with their peaks hitting 90 IRE. Adjustment nodes added prior to the LUT node can obviously be used to massage delivery levels if needed.

Chris Young

https://www.xdcam-user.com/2016/12/the-great-s-log2-or-s-log3-debate/

 

[AmbiSonde]

When someone suggested using HLG instead of SLog3 it caused a meltdown, there was one person who even claimed HLG can't be graded, LOL. There's a reason why Sony doesn't include SLog2 in their cameras anymore, can you guess why? See chart...

For those who missed it the first time, HLG is significantly cleaner than SLog3 even when SLog3 is overexposed by 2 stops yet it retains at least stop more dynamic range. HLG has a better distribution of samples than SLog3, it is more contrasty and therefore needs less transforming than SLog3 so it is much more robust in grading.

S-Cinetone gamma with the REC709 gamut is by far the best direct to REC709 Profile Sony has made possible. The S-Cinetone colours are verging on the clichéd but the gamma curve is excellent which has hybrid log characteristics, more contrasty than HLG but with excellent log highlight preservation.

 

[cyvideo]

S-Cinetone gamma with the REC709 gamut is by far the best direct to REC709 Profile Sony has made possible.

Very much agree. I've been using Resolve Color Space Transform with HLG 3 to 709 n a number of cases and am more than happy with the extended range we can squeeze into the 709 bucket. Being able to use low ISO 125 along with the clean low noise floor of HLG 3 produces a very acceptable image. As for the "HLG wasn't made to be graded." comment. Well, neither was Rec 709. No baked in color space was designed for grading, but it's been done for countless years ever since Rec 601 and then 709. Absolutely no way you wouldn't grade HLG to improve an image as long as you know its limitations and just how far you can go. As we've been doing with 709 grading.

Chris Young

 

[AmbiSonde]

Very much agree. I've been using Resolve Color Space Transform with HLG 3 to 709 n a number of cases and am more than happy with the extended range we can squeeze into the 709 bucket. Being able to use low ISO 125 along with the clean low noise floor of HLG 3 produces a very acceptable image. As for the "HLG wasn't made to be graded." comment. Well, neither was Rec 709. No baked in color space was designed for grading, but it's been done for countless years ever since Rec 601 and then 709. Absolutely no way you wouldn't grade HLG to improve an image as long as you know its limitations and just how far you can go. As we've been doing with 709 grading.

Chris Young

Pleased to see I was preaching to the converted all this time.

Is HLG3 clean because of the low ISO or do you think Sony has employed its chroma noise reduction on that profile? It is very clean, suspiciously so, compared to SLog3.

I use two profiles, Cinetone with REC709 Gamut and HLG3 with REC709 gamut, and match them to the dynamic range of the scene I'm shooting. They both play well together on the same timeline. I've compared HLG3 with BT2020 gamut and have yet to see any benefit, there maybe some super saturated vibrant edge case but I have yet to find it. **Tip, if anyone does use HLG3 with REC709 gamut Resolve will think it's REC709 so you have to tell it is HLG with REC709 in the input colour space settings for the clip.

I read so much hype about S-Cinetone and couldn't have been more unimpressed by its look, even graded it just looks like 'Yet another film look LUT' but the gamma curve, yes, now you're talking Mr Sony! It's a bit of a gem.