An adventure into the Panasonic GX85/80 begins - and a look at the Leica Nocticron for Micro Four Thirds

[sir_danish]

4 hours ago, John Matthews said:

I'll be sharing some tests tomorrow concerning specular highlights and optimal contrast settings.

That would be awesome!

[John Matthews]

So, I spent about 3 hours trying to figure out exactly what the best contrast setting would be for specular highlights. The goal was to avoid any high-contrast lines between blown-out portions which is often a tell-tale of poor video quality.

Methodology
I placed a shiny rounded object on a surface and had a light source in the distance. I adjusted camera settings so that focus was sharp, on a tripod, fixed aperture (F1.7), shutter at 1/50, ISO 200, fixed WB. With Zebras set at 105 IRE, I made sure part of the shiny object was overexposed then recorded 4k video in Natural profile (sharpness= -5, NR = -5, saturation = 0). Contrast was first set at +5, then went down from there to -5 with a total of 11 videos takes.

I made these observations:

1. Panasonic does a great job at keeping a subtle degradation from blown-out areas to areas with detail.
2. Contrast settings seem to simply shift midtones up or down, there was only a small shift in the shadow floor, if any.
3. I found it difficult to match 2 shots with different contrast settings; so, something else might be happening with the curve.
4. Increased midtones in post yielded superior results (noise & artifacts) when contrast was at “-5”.

Conclusions
There’s no real benefit for specular highlights in terms of them looking more “video-like” when contrast was set at any of the 11 settings. However, if you plan on increasing midtones in post, you’ll have superior results with contrast set at -5. Otherwise, you’ll need to live with no visible detail in the dark parts of the image.

 

[sir_danish]

3 hours ago, John Matthews said:

So, I spent about 3 hours trying to figure out exactly what the best contrast setting would be for specular highlights. The goal was to avoid any high-contrast lines between blown-out portions which is often a tell-tale of poor video quality.

Wow, thanks for the test! This will help me a lot...

[Vesku]

In my tests with GH4 and Natural profile I have not found any quality differences between contrast -5 and contrast 0. The banding, noise and cleannes is about the same. -5 has lighter or flatter image (more what eye sees) but otherwise the quality is the same. Choose the look you prefer.

P.S. Both has banding in skies and fine gradations. I use deband filter to remove that banding.

[John Matthews]

2 minutes ago, Vesku said:

In my tests with GH4 and Natural profile I have not found any quality differences between contrast -5 and contrast 0. The banding, noise and cleannes is about the same. -5 has lighter or flatter image (more what eye sees) but otherwise the quality is the same. Choose the look you prefer.

P.S. Both has banding in skies and fine gradations. I use deband filter to remove that banding.

Yeah, I didn't find a massive difference between -5 and 0 and I couldn't say that one was THAT much better than the other. However upon increasing the midtones, I saw a little more noise (in terms of frequency) in the shadows. Unfortunately, you cannot see it much on the youtube video mentioned about due noise suppression. I'm now coming to the belief that the best profile settings (for edit in post) on a 4:2:0 codec would be to match a jpeg to an unaltered raw photo (no post-processing), but that needs to be confirmed.

[DPC]

45 minutes ago, John Matthews said:

 I'm now coming to the belief that the best profile settings (for edit in post) on a 4:2:0 codec would be to match a jpeg to an unaltered raw photo (no post-processing), but that needs to be confirmed.

Not sure what a "unaltered raw photo" means. As a general rule, the less data your image contains, the nearer you want to be to your final result in camera. 

 

[John Matthews]

11 minutes ago, DPC said:

Not sure what a "unaltered raw photo" means. As a general rule, the less data your image contains, the nearer you want to be to your final result in camera. 

 

What happens when you don't know what you want the final image to look like? What settings should you use? That's the question I would like an answer to.

My reference to "unaltered raw photo" was in terms of contrast, sharpness, NR, and saturation matching the look of a rw2 file opened WITHOUT any extra processing. My theory is that Panasonic engineers might have optimized their processing to the output of a rw2 file- the resulting jpeg or mp4 matching that look would have the least amount of macro blocking and artifacts. I could be very wrong though... Do you have any information on that?

[Vesku]

1 hour ago, John Matthews said:

What happens when you don't know what you want the final image to look like? What settings should you use? That's the question I would like an answer to.

My reference to "unaltered raw photo" was in terms of contrast, sharpness, NR, and saturation matching the look of a rw2 file opened WITHOUT any extra processing. My theory is that Panasonic engineers might have optimized their processing to the output of a rw2 file- the resulting jpeg or mp4 matching that look would have the least amount of macro blocking and artifacts. I could be very wrong though... Do you have any information on that?

I think the basic general image for everything is JPG Standard profile 0,0,0,0 without any additional features. JPG and video are very similar when using levels 0-255 (I dont know if the GX85 has levels 0-255). I dont think the reasonable image settings affects much. NR, idynamic, sharpness, high shutter speed in motion makes more differences in codec result.

[John Matthews]

39 minutes ago, Vesku said:

I think the basic general image for everything is JPG Standard profile 0,0,0,0 without any additional features. JPG and video are very similar when using levels 0-255 (I dont know if the GX85 has levels 0-255). I dont think the reasonable image settings affects much. NR, idynamic, sharpness, high shutter speed in motion makes more differences in codec result.

The GX85 is 4:2:0. The result is an image with values of 0-255 in RGB. The problem with Standard (0,0,0,0) in 4k is that the image looks like bad 1080p, the major culprit is NR, followed by too much sharpening. IMO if you're after a more "filmic" image, they should be dialed down. In all, there are 7 picture profiles and 18 different filter modes I think. All of them treat the image differently with a number of settings for each. One of these modes/settings is going to be a little or significantly better than other ones. This is what many of us are after.

[Cary Knoop]

The GX85 delivers video levels (16-235) but does record "illegal" values as well. 

That's why it is essential to ensure values are brought within legal range in post otherwise you may actually clip or crush data.

 

[John Matthews]

21 minutes ago, Cary Knoop said:

The GX85 delivers video levels (16-235) but does record "illegal" values as well. 

That's why it is essential to ensure values are brought within legal range in post otherwise you may actually clip or crush data.

 

Yes @Cary, my mistake. I think we settled this 50 or so pages back in this thread. I completely forgot. However, at least when you bring into FCPX, it spreads it out and you'd never know.

[Cary Knoop]

1 minute ago, John Matthews said:

Yes @Cary, my mistake. I think we settled this 50 or so pages back in this thread. I completely forgot. However, at least when you bring into FCPX, it spreads it out and you'd never know.

So how does that work in FCPX, does it reduce contrast linearly to map those values to video levels?  (I am not familiar with FCPX). 

I work with Premiere Pro and in Premiere you do have to bring values into legal range (if applicable of course) otherwise you will clip values.

 

[John Matthews]

4 minutes ago, Cary Knoop said:

So how does that work in FCPX, does it reduce contrast linearly to map those values to video levels?  (I am not familiar with FCPX). 

I work with Premiere Pro and in Premiere you do have to bring values into legal range (if applicable of course) otherwise you will clip values.

 

In FCPX, there's nothing to do. It handles it automatically if I'm not mistaken... that's probably why I forgot. It just knows how to handle it by default.

[Chris Oh]

slightly on topic. It's currently on sale at BH!

$697.99

https://www.bhphotovideo.com/c/product/1302079-REG/panasonic_dmc_gx85kk_lk_dmc_gx85_digital_camera_black.html

GX85 with Kit lens
Panny 45-150mm
$100 gift card

 

[KnowsNothing]

That's a fantastic deal. I'm biding my time wondering if the G85 is going to go on sale, it's seeming unlikely for this month.

[sir_danish]

Yey! I got my GX80 today. I've never had a camera with IBIS, so I wonder if it's normal that there's always some sort of wobble going on inside the camera body. It feels like a rather big component has come loose... :D

[John Matthews]

1 minute ago, sir_danish said:

Yey! I got my GX80 today. I've never had a camera with IBIS, so I wonder if it's normal that there's always some sort of wobble going on inside the camera body. It feels like a rather big component has come loose... :D

Yes. It's normal. Actually, I wonder if the G80/G85 does the same since they have slightly different mechanisms.

[sir_danish]

43 minutes ago, John Matthews said:

Yes. It's normal. Actually, I wonder if the G80/G85 does the same since they have slightly different mechanisms.

Ok... But that "clack" when I tilt the camera forwards doesn't inspire my confidence.  It's hard to believe that this internal movement helps stabilize the image, rather than make it jittery. Not to mention the clack sounds being recorded by the internal microphone. Well, let's see... I need to charge the battery and start capturing some footage.

[John Matthews]

1 minute ago, sir_danish said:

Ok... But that "clack" when I tilt the camera forwards doesn't inspire my confidence.  It's hard to believe that this internal movement helps stabilize the image, rather than make it jittery. Not to mention the clack sounds being recorded by the internal microphone. Well, let's see... I need to charge the battery and start capturing some footage.

Good luck! I hope you enjoy it. Get a MFT Sigma lens and you'll be amazed how many "clacks" a camera can make- borderline maraca! This is not my worry though with the mechanism. I more worried about the wiring behind the sensor... wouldn't it break after all the bending around? Time will tell. Let's hope that Panasonic did their due diligence in their design.

[mercer]

7 hours ago, John Matthews said:

Yes. It's normal. Actually, I wonder if the G80/G85 does the same since they have slightly different mechanisms.

I don't think mine does that. Now I need to check because I have not noticed any loud quack.