GH4 Lots of tests with the new GH4

Hi Guys,

So is it better to set the camera to 16-235 or leave it at the 16-255 default?

If your post pipeline can handle 0-255, that's the best. Then you can have 256 levels (so theoretically the least banding).

If your post pipeline can't handle blacks below 16, but can handle superwhites, do 16-255. You get 240 levels.

If your post pipeline can't handle blacks below 16, and can't handle superwhites, do 16-235. You get 220 levels. EG the most color banding but also the most safe from clipping.

Bruce Allen
www.boacinema.com
 
Thanks Bruce,

I'm on Adobe CC so no problem with the pipe line. If I get really lucky the pipeline will start gushing raw crude oil and I will be able to take that much needed vacation.

Wayne Mann
www.HeliCamHDmedia.com

Sounds good. Be careful even with Adobe stuff. It can still be dodgy.

To me the only real way to test is to just send full-range video all the way through from recording to final delivery and see if it clips :)

Bruce Allen
www.boacinema.com
 
HVXguy, here's a raw still for you. And a JPEG. And 4K Quicktimes cycling through the ISOs :)

-5 -5 -5 -3 0 refers to -5 Contrast, -5 Sharpness, -5 Noise Reduction, -3 Saturation, 0 Color shift.

They are unsullied, except that I deleted the audio track using Quicktime 7 Pro.

https://drive.google.com/folderview?id=0Bx36eGbmhPWuMHhfUlpIODJwT1U&usp=sharing

This was to investigate some weird stuff I saw in the highlights on skin tones and looking for banding, etc. So please don't bother criticizing my choice of still life, lack of dynamic range, etc :)

Lighting was a lone Kino Diva (with 3200K tubes I think), plus some late afternoon light going through the apartment.

It is a good example of some of the issues with Cine-D shooting warm subjects with daylight WB (eg when the red channel clips... but also before the channel clips there are issues). I also tried 3200K WB and it was unpleasant as well. 1/2 CTB was perhaps slightly better. Will upload those later - as well as other color profiles, etc.

Anyway, look at the highlight on the onion: it goes from neutral to orange to yellow. And right before it clips, it goes to a desaturated green! This is also noticeable on the banana from 400 onwards. And the defocused color chips go pretty funky as the red channel clips.

I have shot I do wish this camera had a raw mode or just some kind of "show native response" mode so that we could bypass the stupid color matrixes Panasonic put in it.

Bruce Allen
www.boacinema.com
 
Last edited:
I've found Panasonic iso's to be very odd and inconsistent, here's what I mean, On the gh2, I found iso 200, 400, very clean and low noise, Also when picking iso's it has less noise when scrolling left to right when going through the different iso settings. On the g6, Iso 250 and 640 are by far the cleanest, next is 800. 800 is cleaner than 160 or 320. 250 I would have thought that to be an electronic non native iso, but for whatever reason there's very low noise in the setting. And the g6 also has he scrolling bug, right to left has less noise than scrolling the opposite. What I'm getting at is there's no consistency from camera to camera. It could be the gh4 needs testing in this regard, I would put on the lens cap and go through each iso setting to see what is clean instead of assuming the norm iso's being the cleanest.
 
HVXguy, here's a raw still for you. And a JPEG. And 4K Quicktimes cycling through the ISOs :)

-5 -5 -5 -3 0 refers to -5 Contrast, -5 Sharpness, -5 Noise Reduction, -3 Saturation, 0 Color shift.

They are unsullied, except that I deleted the audio track using Quicktime 7 Pro.

https://drive.google.com/folderview?id=0Bx36eGbmhPWuMHhfUlpIODJwT1U&usp=sharing

This was to investigate some weird stuff I saw in the highlights on skin tones and looking for banding, etc. So please don't bother criticizing my choice of still life, lack of dynamic range, etc :)

Lighting was a lone Kino Diva (with 3200K tubes I think), plus some late afternoon light going through the apartment.

It is a good example of some of the issues with Cine-D shooting warm subjects with daylight WB (eg when the red channel clips... but also before the channel clips there are issues). I also tried 3200K WB and it was unpleasant as well. 1/2 CTB was perhaps slightly better. Will upload those later - as well as other color profiles, etc.

Anyway, look at the highlight on the onion: it goes from neutral to orange to yellow. And right before it clips, it goes to a desaturated green! This is also noticeable on the banana from 400 onwards. And the defocused color chips go pretty funky as the red channel clips.

I have shot I do wish this camera had a raw mode or just some kind of "show native response" mode so that we could bypass the stupid color matrixes Panasonic put in it.

Bruce Allen
www.boacinema.com

Arghhh, Adobe can't open GH4 RAW yet…..
What about that Panasonic RAW converter?
I thought I had that installed, I can't find cause I never used it….
I really need RAW support before I can buy this camera.
 
Struggling to upload at full res...



Stills from video. All shot at 200iso, F7.1. 14mm on the 4k shot, 18mm on the 1080 shots, newer 14-140 lens.

1 - 4k rendered to 1080
2 - 1080 200mbps
3 - 1080 96 fps (from 100mbps stream)
 

Attachments

  • 3-GH4-1080_96fps.jpg
    3-GH4-1080_96fps.jpg
    56.3 KB · Views: 0
  • 1-GH4-4k.jpg
    1-GH4-4k.jpg
    59.6 KB · Views: 0
  • 2-GH4-1080_200mbps_2997.jpg
    2-GH4-1080_200mbps_2997.jpg
    56 KB · Views: 0
In a support call to a major hardware provider, I accidently got an expert on the phone, one who had tested (or his tech folks had) the GH4. He warned me that the DR of the GH4 was not what they expected, and suggested I might want to be aware of this before owning one. Of course, I have to see for myself, and my copy is scheduled to arrive here tomorrow. But I am ready to return it, if this is true.

This source has always been right on the money in the past, and in a position to know… and for some time. I am not at liberty to reveal the source, but I throw it into the pot here, since it may be helpful.
 
Not to belittle your post, but "Dynamic Range" seems to be the hot buzz word as of late.

I don't ever remember anybody complaining about the dynamic range of cameras much before 1 1/2 years ago.
 
Not to belittle your post, but "Dynamic Range" seems to be the hot buzz word as of late.

I don't ever remember anybody complaining about the dynamic range of cameras much before 1 1/2 years ago.

Not to mention 1 1/2 years ago people would have gladly paid 4K plus for what the GH4 is offering here. I wish people would keep things in perspective here. On another forum someone was comparing the GH4 to the Alexa and complaining about the GH4 image. If all this complaining isn't a sign we are spoiled for choice I don't know what is.
 
...On another forum someone was comparing the GH4 to the Alexa and complaining about the GH4 image. If all this complaining isn't a sign we are spoiled for choice I don't know what is.

I have to agree, but it's the same thing that's been going on for a few years now. Everyone expects a sub $5k (or sub $2k) camera to compete and have the features as a $20k (or more) camera, and are disappointed when they don't (with the refrain usually being "But Black Magic...). While I want to see improved DR, the reality is this is a $1,700 DSLR. I simply want to know if it's 1) a better image than the GH3 and 2) a better image than my AF100. If the answer to both is yes, I'm buying one (and all 8+ stops of DR) next week if not this Friday.
 
http://www.techradar.com/reviews/ca...l-slrs-hybrids/panasonic-gh4-1218311/review/3
Dynamic range charts for RAW and Jpeg. Raw is almost 13 stops and JPEG is 11, depending on ISO (800 being better than 400!). EOSHD has said it has 11 stops in video mode, which would be nice!

This is quite amazing. I'm not aware of another camera that shows so little drop off in dynamic range in raw mode as you move up the ISO ladder. Really something, and starting with a baseline of more latitude than the 5d3 (no, I didn't say better image...). They've pulled out some new magic with this sensor.

Based on what I'm seeing there, assuming the video mode is more directly related to JPEG's, ISO 200 is the sweet spot. I'm seeing quite a lot of noise creeping in by ISO 800 even though it's better with DR then iso400.
 
I did some quick DR tests yesterday with a friend's GH4 and our C300, Nikon D600, HPX2700, HPX250. The C300 was best for highlight handling in C-Log, the GH4 was last. I wasn't doing full DR tests, just highlight handling, placing all cameras at 30 and 40 IRE with an 18% gray card reference, then shot an 11-step chip chart.

I used Cine D gamma with the GH4. Unfortunately, we only tried ISO 400. I thought the GH4 looked quite good overall, comparing well to the C300 other than DR. This was just recording at 1080/24p in AVCHD.
 
Not to belittle your post, but "Dynamic Range" seems to be the hot buzz word as of late.

I don't ever remember anybody complaining about the dynamic range of cameras much before 1 1/2 years ago.

I think it's just human nature. Dynamic range has become the trait that obviously lacks in cheaper cameras compared to higher end ones, plus I think people are more educated on how important it is. The DOF and resolution frontiers have been conquered, on to bit depth and dynamic range.
 
We certainly were with film. You always want more.

Well, until the advent of the digital camera at popular prices, 'dynamic range' was the usual phrase used. "Latitude" was used both in the sense of what sort of illumination range, 'gamma' curve, and the more important question of,
"If I over expose by N stops, can I 'save' my shot with developing or printing adjustments'.

So a still film like 'Panatomic X', ASA 64, was consider to be a film with little Latitude, that is any over exposure was serious. Plus X was still pretty 'small' on latitude. Tri-X was consider to have more latitude, but the problem was 'grain'(aka noise) could become an issue.

And of course Slide Film was notorious for having very limited latitude... blown whites being the most obvious item.

Movie Film stocks had similar attributes, although the only movie stocks I ever shot stills on were certain color stocks, just to 'try it out'... I mainly stuck to B&W Tri-X or HP4.

It was some what easier to 'adjust' than current DSLR seem to be. The 'styles' seem to really be somewhat limited in really adjusting for shadow detail, while maintaining control on the highlights.

In the olden days one would expose for the 'shadows' and develop for the highlights, with in the limitations of the 'latitude/grain' criteria.

It is my belief that most of the 'goals' of DSLRs is 'direct viewing', and as such, the same goal was for slide film. So, for the most part the 'dynamic range' will be 6-8 stops at best. With some cameras other 'curves' are being used to get more 'stops' of dynamic range, but unfortunately there is pretty much chaos on how that is achieved.

There is the issue of the number of bits that are used to represent the light intensity, and how those bits are mangled to allow for a wider range to be represented that what the number of bits may suggest for a 'linear' relationship.
 
Back
Top