GH4 - Best Video Settings

yiannis_zach

@simo777 , yes, me too I enjoyed that trick when I found it, gave me less ISO stress and better -low noise- images :)

simo777

Look I know some may already have the feature enabled but I couldn't find any info on it online. Maybe it was already mentioned in this thread but I couldn't see it. Either way, theres no reason why this option is only available in the Photo mode menu. Having the ability to choose from 1000 and even 1250 ISO in video mode should be standard in these cameras, given their relatively poor low light capabilities. Anyway, I've changed it from "hack" to "trick" to clear up any confusion

Vesku

@simo777

That iso "trick" has been there many years from GH2 and GH3. I think it is an old bug and lazy programming from Panasonic. It is one odd feature in GH-cameras like no showing exposure values during video recording.

lmackreath

Are people turning down there sharpness religiously on this camera?.. Even if they are mostly downscaling their 4K footage to 1080p?

I am big fan of Natural on all zeros tbh!!

Mckinise

I'm a Sharpness Zero guy. I do play with other settings from time to time, but sharpness is usually left alone.

caveport

@Mckinise +1

mrbill

I get aliasing when I downscale 4K footage until I bring sharpness down to -3. At this value, all my aliasing disappears - double checked on grade 1 HD broadcast monitors at work.

balazer

If you get aliasing when downscaling 4k to HD, the problem is your downscaling, not the 4k.

Of course filtering out all of the high frequency detail before downscaling is going to fix that. That's what the downscaling is supposed to be doing itself, either as part of the scaling algorithm or as a pre-processing step. You shouldn't have to do it yourself. Find a better downscaler, and your problem will go away.

lmackreath

I notice that if I keep my shaprness at 0 the downscaled 4 footage looks good, if however I leave my sharpness at 0 and shoot footage at 1080p out of camera with no downscaling I do notice some aliasing. This change of resolution is something I have to do while on a shoot for slow motion but in these cases it may be worth turning the sharpness down to -3 for the 1080p footage that is not going to be downscaled?

joethepro

What is the real benefit of cinelike D? It doesnt change the white clipping point, and just seems like a standard gamma curve that you'll have to account for in post. Shooting flat in REC709 space never made much sense to me unless you were in very contrasty locations and truly needed a low contrast profile to get the proper exposure. Then it definitely can help, but whats special about cinelike D vs -3 or -5 contrast in any other profile?

balazer

"Shooting flat in REC709 space" is an oxymoron. The Rec.709 standard renders images in a specific way, with normal contrast. If the contrast is altered, it's not Rec.709.

The GH4 doesn't have any Rec.709 output option. Rec.709 is something you see on professional camcorders. Consumer cameras pretty much all render images in some non-standard way that the manufacturer came up with. That's part of the reason for each different manufacturer's cameras having a different look.

The GH4's Cinelike D is decidedly not Rec.709. It has reduced contrast applied in a funky way that will be difficult to correct without a LUT. So by itself, I'd say there's no advantage to Cinelike D. But good LUTs exist, to transform to a rendered image or to scene referred. So in that sense Cinelike D can be an advantage.

Getting a bit technical on Cinelike D, that was a gamma curve option on some older professional Panasonic camcorders. It has lower contrast than Rec.709, and all you need to do to transform it to a rendered image is to add some contrast with a curve. But Panasonic seems to have screwed up the implementation on the GH4. No curve can transform it to a properly rendered image. You need a series of curves and matrix transformations.

joethepro

@balazar Thanks for the information. I guess I was mistaken as I thought in general rec709 was a standard that has a defined range of values and how they are distributed from shadows to highlights for standard broadcast reproduction. Since consumer cameras do not record in a log format for the most part, I figured they just used the rec709 space. Anyway, it seems to me like the advantage of cineD is mainly to keep the low mid and shadows from getting muddy or blocky by raising them for recording. but the GH4 has such a good codec that it doesnt really even need it as you can lift the shadows quite a bit and still retain image quality. Now vlog-L is an obvious benefit as it gives you more highlight headroom to play with.

caveport

@balazer The GH4 shoots in sRGB or Adobe RGB colour space no matter what the contrast is. Colour spaces merely define the limits, not what is recorded. BTW gamma is not defined in rec709 but is commonly accepted as 2.35 for the EBU standard. I can't recall what is accepted for SMPTE standard.

balazer

A color space defines the relationships between colors and numbers. It's more than just the primary chromaticities, and it may or may not have limits. The GH4 doesn't "shoot in sRGB". Rather, all video shot in any photo style except for V-Log L, Cinelike V, and Cinelike D, is meant to be displayed in sRGB. (Or something close to it; Panasonic doesn't actually say. But sRGB is the de facto display standard). None of the GH4's video modes use Adobe RGB.

Rec.709 absolutely defines a gamma curve:

output = input * 4.5 if input < 0.18

else

output = 1.099 * input^0.45 - 0.099

I think what you mean is that Rec.709 doesn't define a display gamma. Rec.709 doesn't define anything about a display. It's not a display standard. It only defines how a camera should render and encode its output. sRGB is a display standard.

maddog15

Decided to get away from the CINE D profile and go with Portrait for this test. (Everything still dialed down to -4.) Many have commented that the Portrait settings renders caucasian skin better than CINE D. Anyway...aside from the god-aweful banding in the beginning moon rise....any and all input is welcome.

caveport

Rec. 709 is written as if it specifies the capture and transfer characteristics of HDTV encoding - that is, as if it were scene-referred. However, in practice it is output (display) referred with the convention of a 2.4-power function display [2.35 power function in EBU recommendations has also been changed to power function 2.4 since October 2014, according to EBU Tech 3320]. (Rec. 709 and sRGB share the same primary chromaticities and white point chromaticity; however, sRGB is explicitly output (display) referred with an average gamma of 2.2.

Poynton, Charles (2012). Digital Video and HD Algorithms and Interfaces. Burlington, Mass.: Elsevire/Morgan Kaufmann. p. 321. ISBN 978-0-12-391926-7

balazer

A display-referred color space, by definition, specifies the relationships between color values and the colors generated by a display. Rec.709 does not do this, and thus is not display-referred, implicitly or otherwise. Poynton was wrong to say it is. His confusion comes from the fact that images in a display-referred color space are supposed to be rendered for display. Rec.709 images are rendered for display. But that doesn't make them display-referred. He has conflated these concepts. Rec.709 is a bit of an oddball color space for being both rendered for display and scene-referred but not display-referred.

If you really care, I suggest reading https://www.w3.org/Graphics/Color/sRGB.html . Some of the numbers are wrong, but the explanation is correct, though hard to understand. I'd suggest also reading the RIMM/ROMM paper, http://www.photo-lovers.org/pdf/color/romm.pdf , which explains the difference between scene-referred and display-referred and also talks about display rendering. Note that neither RIMM nor ROMM are the analog of Rec.709: RIMM is scene-referred and unrendered; ROMM is display-referred and rendered. And if you care about color, make sure you understand the CIE 1931 XYZ color space really well. All of these other color spaces are defined in terms of XYZ.

caveport

While the rec709 opto-electronic transfer characteristics at source are clearly defined, I think it's wise to note the footnote in the rec709 spec:

In typical production practice the encoding function of image sources is adjusted so that the final picture has the desired look, as viewed on a reference monitor having the reference decoding function of Recommendation ITU-R BT.1886, in the reference viewing environment defined in Recommendation ITU-R BT.2035.

balazer

Sure, but then you're not using the Rec.709 encoding function. You're using an a modified or adjusted version of the Rec.709 encoding function.

In typical production practice, the Rec.709 encoding function is completely ignored. The camera manufacturer comes up with whatever display rendering it wants and uses some aspects of the Rec.709 digital representation, for viewing according to Rec.1886 & Rec.2035, or according to sRGB. The rendering is different enough from Rec.709 that Rec.709 is not even mentioned.

There are cameras that actually start with the Rec.709 transfer function and then apply their own adjustments like contrast, knee, etc. When you set the adjustments to zero, you get Rec.709. These cameras for the most part list Rec.709 in their menus.

konjow

Ive fallowed treads about GH4 settings when was relisted but decided to stay with GH3 for a while. I got GH4 finally today and do not remember now what should be used if editing in Premiere for online upload 0-255 or 16-235?

balazer

If you shoot in MOV format, use either 16-235 or 0-255. If you shoot MP4, use 16-235. Premiere doesn't read MP4 files in 0-255 correctly.

konjow

Thanks @balazer so if use mov 0-255 is fine and do not need to wary about straight output even to mp4?

"output"- export I mean

balazer

Yes.

konjow

thanks again @balazer

DZA

I'm on Portrait. -5-3-5-3 0, 16-235, Shadows 0 Lights +1. It's the best compromise after 1 year of tests.

Howdy, Stranger!

Categories

Tags in Topic

Top Posters