Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
Ok EVERYONE, check out this link
http://vimeo.com/18580410
Some guy did a test and said (as witnessed by the video)
avoid ISO 250, 500, 1000, 2000
Noticed 640 looks cleaner than 400 for some reason, so maybe avoid 400 and use 320/640 instead.
I've done film mode testing on my GH13 and found that even at the GH1's noisy 1600 ISO, Vibrant film mode with saturation and noise reduction at +2, and contrast and sharpness at 0, dramatically reduces noise. I would assume the film modes would interact with noise similarly in the GH2, even though my impression is that the GH2 is far less noisy than the GH1. On my GH1 with the film mode I described, the well-exposed areas of the image have no noticeable noise, and the underexposed areas suffer more from the AVCHD codec falling apart than from sensor noise. The only drawback to that film mode is that it has higher contrast by default, so the overall image brightness is not quite as high as it would be with Standard at the same ISO. However, having little-to-no noise is my preferred option. You can recover the stop you lost in the film mode when you're grading.
I'm the guy who did the test, but I'm skeptical about my own conclusions.
ISO 800 is generally accepted to be the safe limit of the GH1. However, that general recommendation doesn't account for playing with film modes like I did, so film mode tests would be good.
I'm the guy who did the test, but I'm skeptical about my own conclusions. Somebody in Germany did a test that shows ISO 320 and 640 to have a ton of chroma noise in the shadows. I did my own follow-up and sure enough, in real world shooting (as opposed to shooting the lens cap like I did in my first test), ISO 640 looked noisy. Honestly, in real world shooting, I've found every ISO from 160 to 400 to look gorgeous. Above that, 800 and 1600 are my favorites. I avoid 640 because of chroma noise, and I avoid 1000 and 2000 because of how bad they looked in my original test. 2500 and 3200 are for absolute emergencies only.
Here's the other guy's test that shows 320 and 640 to look bad:
However, like I said above, in my experience 320 is fine.
Thanks for the test! May I ask how you composited the noise to look like that? Just curious.