GH4 Best foot forward

M

Mestizo Devon

Well-known member
I'm looking for a profile that will put the GH4's best foot forward, something that will make it shine so to speak.
Where have we settled with profiles??? James miller, cine D/V log like.... 235/255, Sharpness, saturation all down?

Any thoughts
 
My contention is that a baked in image in cam is far superior with the GH4 to anything shot flat for adjustment in post.

My personal flavor is to shoot CineV at 16-235 with all zero's and expand the range slightly with an s-shaped color curves adjustment in post. I still firmly believe that Panasonic is using full raw sensor data before the codec in cam to get to 16-235 then compressing. This is far superior to this all happening after compression in post.

Cheers,
Pete
 
I agree that baked in is not a bad way to go... especially as I often hand over the footage right away or upload it on set for client remote review. Depending on the situation I tend to shoot Natural with sharpness and saturation turned down a between -1 to -3 and the rest all zeros. Shadow / Highlight either off or -2/+2. Master Ped often zero and no higher than +5. iEverything OFF. White Balance is so easy to do - don't trust auto.

For the edit - yes I have a 1080 timeline and just drop the 4K footage into it. I shoot Cinema 4K so (in Premiere CC) I set frame to size then increase it to 50% to get rid of black bars... and can scale up to 100% as needed - but usually not more than 90%.
 
CustomerGH4 said:
c3hammer, why not shoot in 255? It can always be brought down to 235 in post very easily.
Click to expand...
from my testing I've seen a couple of advantages of shooting 16-235. Firstly the GH4 seems to protect the highlights better. Secondly it is using raw sensor data to push and pull into the smaller range then compressing. When doing it in post your computer only has 4.2.0 8-bit data to push and pull.




I do lots of full sun landscapes and blown clouds and highlights don't clip as bad in 16-235. The in cam processing with full sensor data seems to do a very good job with this.



cheers,
Pete
 
Mestizo Devon said:
Are you guys having to drop the 4k footage in a 1080 timeline to see better IQ?
Click to expand...

The GH4 performs optimally in 4K mode, so to minimize aliasing and moire, and to maximize image sharpness, you want to shoot in 4K as much as possible.

However, if your delivery target is YouTube then it may not make much difference.
 
Mestizo Devon said:
Are you guys having to drop the 4k footage in a 1080 timeline to see better IQ?
Click to expand...

Not "having to", but so far I've been dropping 4K footage onto a 4K timeline. After editing in 4K I render to 1080p. Tends to look a bit "richer" to my eye.

Of course your computer and editor have to be able to handle 4K. I'm using an almost 2yr old HP with 2nd gen Intel Core i7 2600 (definitely not the latest), 8GB RAM, and GeForce GT 530 graphics card with 2GB RAM. Seems to be enough to do the job.
 
Consider this as your base setting:

4K
24P
Natural
Contrast -5
Sharpness -5
Noise Reduction 0
Tint +2
Shadow +2, Highlight -2
Range 0-255

I came to these settings after shooting a lot of test charts (Datacolor SpyderCheckr) and various skin tones. I was originally predisposed to use the Cine-D profile, but I found that while it does yield a bit more usable latitude, it has color problems that are nearly impossible to fix in post.
 
I would highly suggest people don't use Shadow / Highlight adjustments.

They don't increase dynamic range, and introduce strange colour issues in shadow and highlights areas, and increase noise.

I have done my own tests to verify this, raising the shadows in post produces much cleaner results, and yields the same dynamic range, but without all the colour issues.
 
Also true.

I also like Scenery, -2, -5, -3, -3, +2. But not always. Sometimes CineV -2,-3,-2,+1,+1. Sometimes other settings. Sometimes contrast and saturation are wanted, sometimes not, etc etc. I think this is yet another one of those cases where there is no single 'best,' and it all comes down to personal preference.
 
ukfilmer said:
I would highly suggest people don't use Shadow / Highlight adjustments.

They don't increase dynamic range, and introduce strange colour issues in shadow and highlights areas, and increase noise.

I have done my own tests to verify this, raising the shadows in post produces much cleaner results, and yields the same dynamic range, but without all the colour issues.
Click to expand...

My experience differs. I found that small (1 or 2 notches) bumps open up the shadows and provide a more natural roll-off on the highlights *pre-compression* yielding better results than trying to do the same in post.

With certainty, I found that small tweaks to the highlight and shadow curve utilizing the Natural profile yielded more accurate color than using the Cine-D profile and leaving the curves at zero.

Whether or not you gain any *real* dynamic range increase is questionable, but it's a really common complaint that the output from the GH series (discounting the hacked GH2) is too video-like. Knowing reasonable steps to take to tame (flatten) the output can be the difference between a happy client and a client that feels like the image looks too "cheap."
 
joe12south said:
My experience differs. I found that small (1 or 2 notches) bumps open up the shadows and provide a more natural roll-off on the highlights *pre-compression* yielding better results than trying to do the same in post.

With certainty, I found that small tweaks to the highlight and shadow curve utilizing the Natural profile yielded more accurate color than using the Cine-D profile and leaving the curves at zero.

Whether or not you gain any *real* dynamic range increase is questionable, but it's a really common complaint that the output from the GH series (discounting the hacked GH2) is too video-like. Knowing reasonable steps to take to tame (flatten) the output can be the difference between a happy client and a client that feels like the image looks too "cheap."
Click to expand...

I experimented with small increments as well as large ones, my experience was still the same.

I realise coming on to a forum such as this and posting minimal details of my testing methods and preaching about the results is a bit silly... but like many others I've not got time to write pages and pages about how I tested this.

I think you're right in regards to the image certainly LOOKING flatter, but it's flattening range that is already there, and doing so in my experience (not yours obviously) introduced more chroma artifacts than it was worth when applying a similar curve in post. It turned my shadows which were largely monochrome, into a green casted nightmare and would have been horrible to grade.

Kudos for testing yourself, I'll bare what you said in mind if I decide to do more tests in the future.
 
You must log in or register to reply here.
Back
Top