Philip Bloom´s Christmas Shootout Part 2 - skin tones, codecs and low light

Well, yeah you COULD have slowed the shutter speed on the GH2, but I think he wanted to use the same shutter speed on all the cameras....

My point was -- drawing the conclusion that the GH2 is "bad" in low light based on Bloom's test -- is a bad conclusion.



I think the GH2 is the HV20 of today. The little train that could. Low price for surprisingly good quality, lots of workarounds needed to make it pro, but possible. A real lovechild of the forums.

Two weeks ago (before I started playing with this "hacked" GH2) I'd have probably argued that the HV20, in it's time, was a more substantial camera than the GH2. After two weeks of playing with it and manipulating the (147 Mbps, mpeg-4, intraframe) files in post -- I'd say that characterization is off-base in a very, very big way.


Only problem is that it isn't clean. So I have replaced moire and aliasing on my 5D's with a [hacked] GH2's noise in the shadows (even on low ISO's) and even worse, banding in the sky or on gradients. The banding is a real and serious problem

You need to be aware of the "ISO bug" and how the get around it. The GH2 is cleaner than the 7D/60D/T3i/T2i.

On the hacked GH2 I've had ZERO banding issues. ZERO. Including at ISO 160. This indicates that the banding problem with the AF100 and un-hacked GH2 is an encoding problem, not a sensor or "8 bit" or filtering problem.


Regarding some of the other responses in here -- I am absolutely not a member of the GH2 "cult," I own an FS-100 and various Canon HDSLRs, and it's entirely possible that if the 1Dx doesn't deliver I'll own a C300 this year. Two weeks ago, before playing with the high bitrate (defined as over 100 Mbps, intraframe) hacks I'd have been more than happy to be condescending to writers saying nice things about the GH2. But I'm telling you -- as a guy who truly despises micro-four-thirds -- these new hacks (starting in about November) have created something totally different and truly mind-boggling. Pulling two to four stops of usable detail out of seriously dark shadows is real. The image detail is in EX1r land, with the low light of the 7D, reliably recorded to a codec you can't even find in an external recorder.

The thing that people don't seem to understand is that these "hackers" didn't just increase the bitrate, in many cases they've actually rewritten encoding algorithms. That "Driftwood" hack featured in the first Bloom video had much more going on with it than just increasing the bitrate of a stock GH2.
 
My point was -- drawing the conclusion that the GH2 is "bad" in low light based on Bloom's test -- is a bad conclusion.

Nowadays pretty much all new cameras are really
good. I would not say the GH2 was bad in low light.
I would say it was not as good as the FS100, the F3,
or the C300
 
I guess that's the problem I have with Panasonic and the AF100 at the moment. If these guys can hack the GH2 and produce a superior image (in this case, a FAR superior image), then why can't Panasonic either update the AF100 or release a camera that can compete with Sony's Super35 cameras and Canon's C300? I'm not asking them to go away from 4/3 sensors. I recognize that they have significant capital invested in that format. I just want to get the best image out of it as possible and the Driftwood hack of the GH2 proves that it is possible. Give me a high-bit rate AF100 that record to P2 or, even better, to SD or CF cards and you've got one heck of a camera.
 
Back
Top