V-Log and 95% Zebras

[PJD]

To get familiar with how the S5iiX records ProRes HQ video using the V-Log LUT, I shot a quick test at ISO 640, with Zebras set at 95%. I shot 5776 x 3056 resolution, but of course resolution doesn't affect exposure or Zebras. I did a custom in-camera white balance on the grayscale chip chart.

I noticed in V-Log you can't set Zebras higher than 95%, which you can do with many of the other built-in LUTs. I'm not sure why that is, but it's not a problem. This test showed me that with V-Log, anything past 95% has little or no detail in it. Now I know anything white/shiny that shows 95% Zebras will clip >100% if you raise exposure even just 1/3 stop more. Of course, sometimes you want part(s) of an image to clip past 100%; it’s a matter of taste. But the cam’s Zebras are accurate & predictable.

When I brought the video clips into Davinci Resolve Studio, I used its Color page to do a one-click color correction using the “OneShot” color chart in the frame. No other adjustments.

In the 1st grab, notice the white “chips" in the upper-right (and lower-left) corners of the grayscale chart.

On the S5iiX LCD, the white chips displayed 95% Zebras, because that’s what I'd set the camera’s zebras to, and I adjusted the camera’s exposure so Zebras would _just_ barely appear there. The camera’s Waveform monitor indicated the white chips were just a hair below clipping 100%. If I adjusted the exposure down “one click” (1/3 of a stop darker), the Zebras disappeared.

Davinci’s Waveform monitor also confirmed the white chips were at 95%, just a hair below clipping. So, the camera’s WF and Davinci’s WF agree. Hooray!

In the 2nd grab below, the cam’s LCD displayed 95% Zebras in both of the TWO upper-right (and lower-left) white chips, because I’d adjusted the cam’s exposure UP one click (1/3 stop brighter). For example, this caused the upper-right chip to over-expose, and the chip to its left to be at 95%. As expected, the clipped whites >100% were unrecoverable in Davinci.

 

[inspiredtimothy]

Is 95% the clipping point in other picture modes or just V Log?

 

[CharlesH]

For me a big advantage of Panasonic video cameras is the good waveform monitor so you don't need to use zebras. This gives a constant reading for the whole scene so you know the the overall exposure situation. My Sony cameras don't have waveform monitors so for video I use Zebras, and it always seems like a step backward in exposure control.

 

[CharlesH]

Is 95% the clipping point in other picture modes or just V Log?

They all vary a bit. Use the luminance spot meter to determine this.

 

[inspiredtimothy]

They all vary a bit. Use the luminance spot meter to determine this.

I'm sorry, I'm not really sure how the luminance spot meter works. How would I go about doing this?

 

[Michael_too]

Is 95% the clipping point in other picture modes or just V Log?

Just Vlog. The others all clip at the maximum (109 I believe)

 

[PJD]

For me a big advantage of Panasonic video cameras is the good waveform monitor so you don't need to use zebras. This gives a constant reading for the whole scene so you know the the overall exposure situation. My Sony cameras don't have waveform monitors so for video I use Zebras, and it always seems like a step backward in exposure control.

Agreed. I'm just doing these simple tests to see what's what since the camera is new to me.

The S5iiX waveform display is really nice to have. The attached snapshots show how the 95% Zebras change with a small difference in exposure (in this example, f3.5 and f3.2). The 2nd bar from the upper right in the grayscale goes from showing no zebras to showing zebras, and the change is easily visible in the waveform.

Note: My previous post above shows results with ProRes HQ (which is 422 10-bit). The snapshots attached here show h265 "MOV" 422 10-bit, so the results are the same as far as I can tell.

Of course, this is kinda pixel peeping. What's more important is how good or bad does the final video actually look, regardless of what charts say.

 

[CharlesH]

Here is an instructional video from Panasonic that explains the exposure tools in Lumix cameras.

 

[CharlesH]

Here is an instructional video ...

And I learned here about setting zebras to the Base Level/Range. I missed that in several years of using Lumix video cameras. There is the old adage, "You learn something new every day." Thanks for the thread Peter.

 

[PJD]

Here is an instructional video from Panasonic that explains the exposure tools in Lumix cameras.

Thanks for the link! I'd missed that one. A very helpful video.

I soon plan to do a careful comparison of recording S5iiX video in auto iA mode vs. recording video of the same scene in full manual mode. The differences will be interesting. Cheers!

 

[PJD]

And I learned here about setting zebras to the Base Level/Range. I missed that in several years of using Lumix video cameras. There is the old adage, "You learn something new every day." Thanks for the thread Peter.

Base Level/Range for use with V-Log is new to me, and I've started using the Luminance Spot Meter. Very interesting.

I shot a few V-Log h265 MOV 422 10-bit test clips yesterday, and I found there's no discernible detail in very bright white areas of a frame where the Luminance Spot Meter indicated +6.2 stops (converted to Rec 709 in Davinci Resolve Studio). But maybe this requires HDR processing & monitoring? (I monitor in SDR.) I plan to shoot more tests at using ProRes (422 10-bit) at +6.2 stops, and at slightly lower spot readings, to see at what point details can be seen in bright white surfaces. TBD.

 

[CharlesH]

I shot a few V-Log h265 MOV 422 10-bit test clips yesterday, and I found there's no discernible detail in very bright white areas of a frame where the Luminance Spot Meter indicated +6.2 stops (converted to Rec 709 in Davinci Resolve Studio). But maybe this requires HDR processing & monitoring? (I monitor in SDR.) I plan to shoot more tests at using ProRes (422 10-bit) at +6.2 stops, and at slightly lower spot readings, to see at what point details can be seen in bright white surfaces.

My experience is the same, even in HDR and with an HDR monitor. I try to limit exposure to about +5 stops; this is about one bar down on the waveform monitor. This works out as about 800 nits, which is enough for me. Some very small areas pushing +6 might be okay, these won't show on the spot meter but on the waveform monitor. I'm interested to see what you find.