FS7: Sony PXW-FS7 are the colours correct?

[Liam Hall]

I'm not wrong. I just wrote what I found out that'll work for me. Either use custom mode and have correct red or use Cine EI and get correct red by using Sonys S-log3 LUT. Is that wrong?

If your methodology is wrong, which it is, how can your conclusions be anything other than 'opinion'?

 

[Lenilenapi]

Alister chapman says that slog 3 is only designed for sgamut cine color space which you only have in Cine EI . Otherwise it will map out incorrectly and problems are possible . He also suggests shooting CineEI w 800%709 LUT on in VF so you can use ordinary Zebras and your eyes for exposure and focus won't be hard . Sounds like good advice for doc shooting .
Just. Be aware you have much more DR than you see .

 

[morgan_moore]

Alister chapman says that slog 3 is only designed for sgamut cine color space which you only have in Cine EI . Otherwise it will map out incorrectly and problems are possible . He also suggests shooting CineEI w 800p9 LUT on in VF so you can use ordinary Zebras and your eyes for exposure and focus won't be hard . Sounds like good advice for doc shooting .
Just. Be aware you have much more DR than you see .

This is exactly as I have done and IMO it works perfectly. Put your zebras on detail you want to retain and you still have some space for specular highlights.

Two gotchas (or more)

-no zebras in 4k
-no LUT or zebras in HFR

The only exposure I have fouled up so far is that my LUT mystically turned off

Edit..

To add
If you do this at EI800 you will get a clean file
If you use the sony lut the reds will be right
If you use CatBrowse the EI is passed to the 'colour tab' and you dont need to adjust exposure in post
If you use Davinci the EI is lost and you will have to adjust your exposure in post (unless you shot at 2000)

 

[Alpha Beta Omega]

If your methodology is wrong, which it is, how can your conclusions be anything other than 'opinion'?

You can't say it's wrong without explaining what's wrong with it. Then it's only your opinion.

 

[Liam Hall]

You can't say it's wrong without explaining what's wrong with it. Then it's only your opinion.

Fair enough. Tomorrow I'll write up where I believe your methodology falls short.

 

[Alpha Beta Omega]

Thanks, you learn as long as you live.

 

[Liam Hall]

Ok. As promised, here are some of the areas where I feel your method of analysis fell short;

1. You write:

First you have to have a point of reference and I picked my Canon 5D mk3 as the reference.

The objective of your test is to analyse the colour the FS7 produces, specifically red, not to match it to a low bit-rate camera that has a different colour space. The only reference you need is a good colour chart.

2. You write:

I used only daylight as light source to avoid anything that could influence colours. Had I used fluorescent or LED chances are colours could have been off by them. This is the image above and to me it looks perfectly fine.

What colour temperature was the daylight? Daylight can be anything from 2K to 20K. It looks to me that you used ambient light, so you'll have no way of verifying the colour on the chart. You need a constant light source with a verifiable and repeatable colour temperature.

3. The colour chart needs even lighting. Yours is far from even.

4. You need to measure the colour produced by each different gamma set-up using a vector-scope.

5. The chart needs to be exposed correctly - on your chart the exposure varies due to the lighting.

6. For cineEI you need to apply a LUT before analysis.

7. You write:

It’s only when the camera is in base mode of Cine EI red hue is skewed. When you are in Custom mode red looks fine.

I'd say the reverse is true, but I haven't tested that theory yet. Will do soon...

8. You write:

If you want colours to end up correct or at least to the same fidelity as the Canon 5D mk3 you can use the option to shoot in Custom mode..

The 5D3 can be set-up in any number of ways using picture profiles. I own two 5D3s, and they produce different colours. 5D3 has less colour fidelity than the FS7. You'd need to tweak each individual camera using a waveform to match them properly. But yes, you'd use custom mode for this.


Sorry to diss your methods Osslund. Accurate testing is difficult (one of the reasons I rarely share mine!) In the past few days I've seen quite a few tests on the FS7 that fall short of accurate and quite a lot of FUD on the internet because of it.

 

[Alpha Beta Omega]

Thanks for breaking it down. I did the test for myself in the first place but decided to share my findings so it's not a be-it-all-correct-test. Its subjective and not objective as everything else I write on my blog.

My primary objective was to check if the colour of red had another hue than considered correct. By looking at the red ring on the L-lenses in my video I concluded that yes they did look off. I soon found out that it was the Sgamut3.cine that made the raw footage have a different red hue. But as I learned in the process by another Dvxuser there's the Sony provided LUTs for S-log3 footage and with that LUT red looks just as good as my reference the Canon 5Dmk3.

I picked to use the Canon 5Dmk3 in stills mode might I add. Just because it's the camera I have with the best quality and colour science. At least that's the way I look upon my 5Dmk3 in stills mode. I would never use the poor video recording for such a test.

Again it's all subjective and you are right I could have done it much more scientific and correct. In the end as soon as I narrowed it down I found little reason to make it more elaborate.

 

[Alpha Beta Omega]

BTW, I simplified my article since I did such an unscientific test!