GH5 GH5 Moire

Cary, 4k resolution on vimeo. The original file is 4k too. I see the problem perfectly on my 27' Imac. On my FullHD TV set is even worse. Maybe it's a downscaling issue but I need to downscale my 4k footage to 1080p. I never had any problem with the 4k downscaled GH4 footage.

No, it's not a downscaling issue. I also looked at your 4K file on a 5K display and there is something strange going on. For a few seconds the image is fine, and then it sort of begins to pulsate. Having read Lpowell's post, I did see strange little filaments at the top of the lampshade. Whether or not those were his "gritty form of modulation distortions" I don't know, but there is an issue there. Have you tried a manual focus lens to see if it's a combination of IBIS and electronic focusing interacting in a strange way?

Personally, that wouldn't stop me from getting a GH5. It is an aberation that I would try to reframe or relight if it was a critical image, which in this case it does not appear to be....
 
Thank you everybody. Stoneinapond, I have only Panasonic lenses so I can't try to repeat the test with a manual focus lens right now. The issue is very bad when you see it on a normal full hd tv. Maybe I'm doing something wrong, I use an Imac connected to a hd tv via blackmagic ultrastudio express, FCP X 10.3 (no problems at all with GH4 4k footage). Anyway, like I said, I can't see a real alternative to this camera for my needs therefore I'll keep it. It's a great camera in every other aspect.
 
Cary, my IMac monitor is 2560x1440; my TV set is FullHD. On both I see a huge amount of moire in this picture. I don't have any real 4k monitor. Maybe my computer and my TV aren't able to render the picture properly, I don't know. If you don't see what I see then I guess I'm doing something wrong. I really would like to understand what.
 
Cary, my IMac monitor is 2560x1440; my TV set is FullHD. On both I see a huge amount of moire in this picture. I don't have any real 4k monitor. Maybe my computer and my TV aren't able to render the picture properly, I don't know. If you don't see what I see then I guess I'm doing something wrong. I really would like to understand what.
What I would recommend you try is to crop (not resize!) the video to the native resolution of your monitor and view it.
 
It's a common misconception that eliminating the OLPF increases the "natural" detail of a demosaicked image. What it actually does is spray ultra-fine edge details across the image, but corrupted in a spatially inverted manner. The resulting edge scintillation appears to enhance sharpness, but it's not natural at all, it's a particularly gritty form of modulation distortion produced by violating the sampling theorem with cavalier disregard for the sensor's Nyquist frequency. In other words, it's an insidious contamination that cannot be squeegeed clean in post, at least not without permanently degrading genuine details as well.

The GH5 has a 20Mpix image sensor and lacks an OLPF, compared to the GH4's 16Mpix sensor with OLPF. In this regard, Panasonic is following Nikon's example of omitting the OLPF on recent cameras that use sensors higher than 16Mpix. The theory is presumably that past a certain Mpix threshold, ultra-fine aliasing artifacts are no longer discernable (in most cases...). However, Nikon seems uncertain whether 20Mpix is enough, as they omitted the OLPF in the D500 but retained it in the D750. And it's not because the D750 is full-frame - aliasing is strictly a matter of sampling granularity rather than image sensor size. Here's detailed info from Nikon on OLPF issues:

http://www.nikonusa.com/en/learn-and-explore/a/products-and-innovation/moir%c3%a9-false-color.html

Back to the GH5, while its lack of an OLPF makes its still shots even more excrutiatingly "sharp" than the GH4, that's not likely the culprit producing moire in 1080p and 4K video modes. What's critical there is exactly how Panasonic demosaicks the 20Mpix sensor down to video resolution. We don't know the sticky details of this algorthm, as neither Panasonic nor any other camera manufacturer are willing to disclose their technical trade secrets. But with moire, the bottom line is what you can see with your own eyes, particularly while panning. For those who want to see what the GH5's moire actually looks like, use a sharp lens to pan across a sinusoidal Zone 720 resolution chart like the one below, and view it at 100% magnification on a one-to-one pixel monitor:

http://www.bealecorner.org/red/test-patterns/Zone720-sine-g22-b.png

Oh, this is giving me a headache.....I am going to go lie down and dream of simpler times.
 
Cary, my IMac monitor is 2560x1440; my TV set is FullHD. On both I see a huge amount of moire in this picture. I don't have any real 4k monitor. Maybe my computer and my TV aren't able to render the picture properly, I don't know. If you don't see what I see then I guess I'm doing something wrong. I really would like to understand what.
I've an iMac 27" too (2560x1440) and I'm seeing the moire pattern.
 
Go into your display settings and set your resolution to 5K, then play it again full screen and let us know if you still see the moire patterns.
5K obviously can handle UHD but if you scale UHD to 5K you may get some artifacts.
Now the ratio of 5K to UHD is large enough but for instance scaling 4K to UHD which is a very small ratio is more problematic.
It has something to do with Nyquist's sampling theorem.
 
Cary, my IMac monitor is 2560x1440; my TV set is FullHD. On both I see a huge amount of moire in this picture. I don't have any real 4k monitor. Maybe my computer and my TV aren't able to render the picture properly, I don't know. If you don't see what I see then I guess I'm doing something wrong. I really would like to understand what.

This really sounds like your displays are not able to properly downscale 4K UHD footage to their native resolution. I'm going to try doing a proper downscale to 1080p and will post the footage to view later tonight.
 
5K obviously can handle UHD but if you scale UHD to 5K you may get some artifacts.
In order to evaluate moire properly, you must view the original footage at 100% magnification on a one-to-one pixel monitor. If your monitor displays in higher resolution than the original footage, you should NOT view that footage full-screen, you should display it in a window EXACTLY the same pixel dimensions as the original footage. Otherwise, any moire you see is just as likely to be a display artifact as a recording artifact.
 
In order to evaluate moire properly, you must view the original footage at 100% magnification on a one-to-one pixel monitor. If your monitor displays in higher resolution than the original footage, you should NOT view that footage full-screen, you should display it in a window EXACTLY the same pixel dimensions as the original footage. Otherwise, any moire you see is just as likely to be a display artifact as a recording artifact.
That is exactly what I was saying.
 
Hmmm... Using the original UHD footage from Vimeo which shows no moire when I view it at 1:1 pixel size, it seems to always show moire in the center lamp when I downscale it. I tried a few different apps to downscale, including using the Cineform CODEC which has never failed me before, and none of them were able to give me a clean 1080p version of this footage.

I see NO MOIRE in the original UHD footage when viewed at 1:1 pixel size.
 
Just watched the first clipagain on a 5K monitor but at the native size. The lampshade is full of moire. I do think the native lens has something to do with it.
 
You might want to try putting the camera on a tripod, switching off in-lens and i- body stabilization, and perhaps put the lens in manual focus. What does that look like?
 
Back
Top