Sunday, July 20, 2014

Panasonic GH4 "Range Wars" - Which Dynamic Range Test Do You Believe?

The $1699 Panasonic GH4 seems to elicit some pretty strong reactions. Opinions run the gamut from reviews like "Best Camera Ever?" to a "scientific" test called "Dynamic Range – Sony A7S vs. the others" where the results seem to confirm the testers' preconceptions, e.g,  "the rather videoish look and contrasty colors we can subjectively observe" from the GH4.

Here's the chart showing the results of this test (using the $3540 Xyla21 – 21 step Grayscale Combi-DX1):

GH4 CineLike Dynamic Range Test - Assessed as 10.9 stops by Cinema5D

The post describing these results gives us a few details about the methodology (e.g., the lens and ISO they used), but doesn't show waveforms or tell us which CineLike profile the testers used (this is important because CineLike V has less DR than CineLike D).

Fortunately, we can compare these results to other data-driven tests. The best I've seen is "Panasonic GH4: DR and Gamma, Timelapse, and More" from Adam Wilt writing for

Adam also used the Xyla 21 to test the GH4's Dynamic Range - but he tested it across various color and CineLike profiles. Here is his result for CineLike D:

GH4 Dynamic Range Test - Assessed as 11+ stops by Adam Wilt at dvinfo
Why the clear difference between the results of the two tests? Perhaps the Cinema5D testers used CineLike V instead of CineLike D. Or perhaps it was confirmation bias related to the subjective observations about the camera's "videoish" look.  Hard to know.

Personally, I don't see a "videoish" look from my GH4 unless I want it.  Neither do I see it from Phil Bloom's 4K Postcard from Phang Nha or Guillaume Le Berre's Kaydara.

That said, I really respect the effort the folks at Cinema5D put into this test, but in light of the significant difference with the dvinfo results, they might want to take another look at their conclusions.


  1. It is confusing for the readers. I appreciate GH4 tools for getting more details from the highlights and some from shadows but I must say I don't like sacrificing the skin light profile. I found out maximizing dynamic range through highlight/shadows reduces the data for the mid tone subtle light profiles. So the test need to take that into the account for comparison. People take Cine D footage and then they find it doesn't have good mid tone light so they pass it through a LUT in the post and again they loose some of that hightlight / shadows. A true high dynamic range sensor would reveal that detail in the highlight and shadows while keeping a nice light profile intact for the mid tones.

    1. I do wish the sensor had more DR - but it is very very good for the price if you know what you are doing. You might want to try Matt Mosher's "M-Log", which is the closest you can get to a Log profile from the GH4 until V-Log comes out:

  2. Totally agree with you!

    I've been investigating the GH4 for several months, and finally purchased it and searched even more to properly use it, and I'm still making tests months later to get the best out of it.

    The GH4 has proven to be one of the best cameras around since april, 2014... But it needs the user to have knowledge. I mean, no camera in the world, nor the best, can replace knowledge. The GH4, as any other camera (or maybe slightly more due to its great customization level), requires knowledge AND testing to get the best results in every situation. Even the GoPros... Even the Reds (especially cameras like those)...

    All those who say GH4 look videoish don't know a sh*t about photography, cinematography, multimedia, cameras... They believe a not-so-expensive and little camera can't be as good as a massive one. But we are in 2015 already! With a GH4, some Blackmagic or even an a7S it isn't mandatory to use so expensive and complex cameras if the production has a tight budget and can't afford the costs of all the gear (with all the needed accessories!) and all the staff and the time to process such amount of data, and you CAN get very good results if you use the tools you have properly!

    In addition, I saw the comments in the Cinema5D post, and the authors showed their ignorance and inability with their comments. They tested the GH4 dynamic range with the wrong settings! They used counterintuitive and irrational settings for DR... They even used “Master pedestal” at +15!!... Awful. Shameful.

    Ignorant people disguised as experts or as professionals say erroneous things, and other ignorants believe them without questioning anything. At first glance it's OK because, fortunately, there will always be righteous people who know how things are, but sadly all those who are wrong are the ones who stain the truth and the work of the truly qualified professionals and difficult progress...

    Thank you very much for this post. Best regards.

    1. Thank you for the thoughtful comment, Sergi. Good luck with your GH4!