GH3 So what's the general consensus of the GH3 moire issue so far...

Gary Huff said:
It's funny that they would add a codec option that's almost useless to shoot with.
Click to expand...
Who cares. If you don't like it don't use it. It has All-I and IBP in .mov and AVCHD. You have three to choose from.
THe more I use it the better it gets. Still need to get out the door and give it a fair test though.

Let's keep in mind that when the Gh2 launched, everyone said they were sticking with their Gh1. Then the Gh2 got hacked and even Coppola wanted one. I have the feeling same will happen for the Gh3. A firmware will come along and make it better and possibly lead to a hack.

All in all, it's a better camera the Gh3.
 
TheDingo said:
...Thanks. I didn't notice that. ( most people upload a full rez file )

The author should have kept things at the original image size.
Click to expand...

Yeah. I actually typically do 720p. Sometimes I work in a 720p timeline so that I can zoom in a little if needed. But even if I do 1080p, I just tend to transcode to 720p. Maybe I should stop. lol.

But I did leave a couple of comments noting that to him. Hopefully it makes since to him if his English isn't fluent. I think that the downrez-ing could likely be creating just a little bit of the moire that is seen, but not the majority-- just a little bit.
 
dustylense said:
Then the Gh2 got hacked and even Coppola wanted one.
Click to expand...

I think that's overstating the actual event.

I have the feeling same will happen for the Gh3. A firmware will come along and make it better and possibly lead to a hack.
Click to expand...

Great! I didn't buy my GH2 until this time last year, so being a late adopter didn't affect me at all.
 
Barry_Green said:
Yes, that is the question. But see, the thing about a properly-shot chart is, it shows you ALL potential levels of detail. So if moire doesn't show up on a chart, then it will never show up in the real world either. And if it does show up on the chart, then that means that it will also show up in the real world either. The question of how frequently is largely answered by -- how much of the chart is covered with moire? If it's only on the littlest tiniest patches, then you can expect that it will be a pretty minimal problem in the real world too. But if the chart is showing huge colored moire swaths on anything past a few hundred lines (like a Canon DSLR does), then you can expect that you'll encounter moire in many situations in the "real world".

Cameras don't operate in two different modes, one for "charts" and one for "real world" -- they operate exactly the same. So if it shows up on the charts, it will show up in your real world footage.


Well, that's a fundamental misunderstanding of the concept of moire then. If there's a certain pattern of detail that causes moire on the chart, then that same pattern of frequency of repeated detail will cause moire in your image. And if it's only a little patch of the chart that shows it, then only that frequency of detail will show moire in your images. But, ANYTHING at that pattern or frequency will moire in your images. So you could get a corduroy jacket exhibiting moire hell on a camera that shows only a small patch of bands of frequency that cause moire, IF that jacket happens to be at just the perfect size in the frame that its corduroy pattern is being sampled at that moire-causing frequency.
Click to expand...

Thank you for your explanation .
 
Just pulled this grab from some quick footage I shot with my Gh3 and it makes me want to cry. Don't have my GH2 with me right now so can't shoot comparison...but come on. Lets ignore the noise and stuff because the lighting is crap and just focus on the aliasing. Window blinds, the upper window frames, etc. Wow just...wow. :( I generally like the camera (and I'm not sure why someone said 72 Mbps is no good for I-frame? Seems pretty decent to me..) but just...wow. What a terrible image on so many levels. Mud, noise, aliasing, crap. Decent highlight falloff though on that lamp in the foreground, out the window not so much.

This is at sharpness -5.

7-14mm f/4.0 (at 7mm, def noticing great crop factor, wide open), 1/30th, ISO 800. ALL-I 72 MBPS 24p.

gr91v.jpg
 
Last edited:
Ben...did you shoot this in ALL I? DustyLense said that he saw bad artifacts in ALL i...then said better results are to be found with the other options. Just curious..
 
I shot ALL-I, just ran through all the other options, no improvement. Moire went away when I defocused slightly, which is I suppose how you would have to deal with this situation. Present in all MOV and AVCHD modes. Also present on these same blinds with other lenses from same angle, but to a lesser extent. Still there even on the 20mm f/1.7, even with it being a longer lens and getting more actual detail.
 
Last edited:
Wow...we've had some far different results and opinions on this issue. Some footage I've seen has it all over the place, then other clips are seemingly devoid of it.
 
I had some great stuff with very high detail areas with no moiré, and yet had some stuff like this with just bizarre amounts of it. I did have an issue where areas of very sharp and high detail critical focus seemed to have noisy aliasing in them, that would move around and be very distracting. However, turning sharpness down in camera seems to have fixed it.

Overall I am somewhat disappointed with this camera, but I like that it has much more flexibility than the hacked GH1/GH2, but I will definitely be keeping mine around. Probably time to retire my GH1, but they're likely not selling for much anyway. I was hoping that the GH3 would do everything better, but it doesn't seem to.

Also, as an aside, anyone else notice that fixed pattern banding is back?! BLARG.

Blame all issues on Sony.
 
...I am happy if the amount of moire is similar to what the GH2 produces. ( this is one reason why I won't sell off my GH2 until I am certain I won't need it again )

I am also curious to see how an HD ENG camera behaves in the same environment.
 
dustylense said:
Beb_B. Try shooting some avchd and IBP .mov on that scene and see if you get a different result.
Click to expand...

I did, no change, at least not discernible to me. Pixel peep might show a difference but honestly they're all across the threshold of acceptability.
 
It's a quick screenshot from full-screen playback on an iMac. Just for illustrative purposes, not technical ones. I assure you it looks pretty much the same.
 
TheDingo said:
It mimics the Canon 5D MK3 format specs, and maybe this can be fixed with a higher bitrate patch. ( 70 Mbps is far too low for ALL I-Frame )
Click to expand...

The problem is Canon and Panny went with VBR instead of CBR (variable not constant bit-rate). The MKIII maxes out at 92Mbps but it can drop to the low 30's in low detail scenes creating all sorts of yuckiness (excuse the highly technical terminology). Nothing a hack can't fix. I am liking what I've seen so far and I wasn't a GH2 fan. Dare I say it looks a bit more Canon like (I'm talking colour and dynamic range not softness).

A 60p codec hack is a must given that you can't get a 60p 1080p recorder without taking out a mortgage. You'd think after 3 years of hacking these companies might get a clue. :nads:
 
You must log in or register to reply here.
Back
Top