GH2 Tests (with a gh13 perspective)

Thanks Hunter. Great tests. I don't understand why people would think that after all the commotion about the mud in the GH1, the poor codec, etc. Panasonic would come out with the next generation of the GH line and it would be inferior to the GH1, even hached. Remember the HVX200 and then the HVX200A? It was a no brainer that the GH2 was going to be much more efficient. And remember that 24mbits is the proper implementation of the AVCHD codec. Can it be shot higher? Yes it can. Will Panasonic risk any further complications by doing this? I don't think so.
 
Great test! Would really consider a gh2 now. No 25p is a bummer though. Lets hope panny releases a firmware ugrade for pal users soon.
 
24pN looks fantastic. I can live with 1080p24 for now. Hopefully Panny updates firmware to upgrade 720p60 to 24Mbps.
 
can you confirm this??
no more or very minimal aliasing(line skipping) moirè , mud and rolling shutter??
thanks
 
Visible here is improved:
DR
Color fidelity
Resolving power
Codec efficiency
& noise
-over the gh13
Indeed. It appears that image processing capability has been greatly improved and updated (B-frame capable) codec is really holding up all of which leads to 24p footage that is clearly a notch above (overall) that of GH13 and two notches above the GH1.

Thanks for putting this test together, Hunter. Would be great to see the comparison in 720/60p modes next. My gut feeling is that GH2 will not come out as strong in this test, due to lack of B-frames and no benefit of increased bitrates. However we will be able to see just how much image processing mojo the new 3-processor Venus HD is packing. Thanks again!
 
Of course there will be some moire, aliasing, mud, and rolling shutter. For pete's sake, it's a camera.
 
Even then, this isn't as good as it gets. The AVCHD in the Gh2 is much better than the GH1's version, but it's still not up to the standards that the AF100 can do. Bigger hardware, bigger memory buffers, more efficiency. Even at the same bitrate.

I really do hope this dispells once and for all the notion that "more megabits = better footage" because, very frequently, that just isn't correct. It's the type of megabits, and how they're used, that determine the efficiency and how good the final pictures look. The only way you can say "more = better" is if you're talking about the exact same thing (i.e., AF100 at 21mbs vs. AF100 at 17mbps... in that case, yes, more = better).

No doubt. Gh2 doesn't have Cabac, and you can see its confused by the noise structure & pattern on the dark parts of the far house. (certain hack is called for, vk ;))
Redcode is a cool example of intelligently used bits; 8:1 storage (36), visually lossless
 
can you confirm this??
no more or very minimal aliasing(line skipping) moirè , mud and rolling shutter??
thanks

A lot less aliasing than gh1, you would really have to pixel peep to find something objectionable- and no noticeable color moiré- this is as good as it gets for a dslr.

Rolling shutter seems about the same as the gh1, maybe a bit better. Not enough to be a huge difference though.

GH2 (attempting to get some color moiré/aliasing issues)- iso1250, NR -2:
huntformiore.jpg
 
Even then, this isn't as good as it gets. The AVCHD in the Gh2 is much better than the GH1's version, but it's still not up to the standards that the AF100 can do. Bigger hardware, bigger memory buffers, more efficiency. Even at the same bitrate.

I really do hope this dispells once and for all the notion that "more megabits = better footage" because, very frequently, that just isn't correct. It's the type of megabits, and how they're used, that determine the efficiency and how good the final pictures look. The only way you can say "more = better" is if you're talking about the exact same thing (i.e., AF100 at 21mbs vs. AF100 at 17mbps... in that case, yes, more = better).
Based on the raw footage you posted from Osaka the AVCHD implementation in AF100 also features CABAC, which GH2 lacks. Hopefully someone who gets ahold of both (hint-hint) can post a side-by-side. AF100 holds a lot of theoretical advantages over the smaller cousin, but it would be great to see exactly how much advantage the video-centric OLPF, professional level hardware and encoder have when dealing with difficult scenes.
 
A lot less aliasing than gh1, you would really have to pixel peep to find something objectionable- and no noticeable color moiré- this is as good as it gets for a dslr.

Yes, a big win.
-What keeps me up at night is color science before compression... everything else I love.
 
Yes it looks great!

Hunter, I've been shooting my hacked GH1 with "C" settings, usually outside, and I'm not seeing the codec breakup as you showed in your first set of pictures. I know it's a hassle but can you try it with "C" settings and no native 24p?
 
Yes it looks great!

Hunter, I've been shooting my hacked GH1 with "C" settings, usually outside, and I'm not seeing the codec breakup as you showed in your first set of pictures. I know it's a hassle but can you try it with "C" settings and no native 24p?

That's what I was saying. I shoot amlost the same settings and have never seen anything like this. Even when waving the cam around!
 
Gh13 users, if you want to try and break the codec like I did on the gh13- put a wide-ish lens on stopped down to get a long dof and put a super high shutter on, shoot a high detail scene (like I did with trees and grass) while quickly panning around and youll probably see what I did if you do frame by frame play back. Again, It doesnt look that bad in motion- this is just an extreme stress test.

Of course, It may or not be better in "c" settings with native 24p not checked, but I dont shoot like that so I didnt test those settings. YMMV.
 
Gh13 users, if you want to try and break the codec like I did on the gh13- put a wide-ish lens on stopped down to get a long dof and put a super high shutter on, shoot a high detail scene (like I did with trees and grass) while quickly panning around and youll probably see what I did if you do frame by frame play back. Again, It doesnt look that bad in motion- this is just an extreme stress test.

Of course, It may or not be better in "c" settings with native 24p not checked, but I dont shoot like that so I didnt test those settings. YMMV.

Ok I will try again tonight, but I did a bunch of tests when I first applied the hack and never got anything like this. I don't have 24p checked so maybe that is why. I guess that will always be the problem. Every GH13 is different based on the hack settings. I was just really taken back on how bad those grabs looked. Thanks again...
 
How about HDMI out?
it's been a while
still a lot speculation out there that says that the port is not available.
it's funny that we still haven't had a solid answer for it.

Thanks for the tests, Hunter
 
I'll try it again too. I was recently shooting across a field into a group of trees through a 16mm lens at F16.. I didn't see anything like that, but maybe I missed it. I'll check the footage again.
 
Ok I will try again tonight, but I did a bunch of tests when I first applied the hack and never got anything like this. I don't have 24p checked so maybe that is why. I guess that will always be the problem. Every GH13 is different based on the hack settings. I was just really taken back on how bad those grabs looked. Thanks again...

You'll need a lot of light (to stop down and get very high shutter), so try to do it in the daylight.
 
How about HDMI out?
it's been a while
still a lot speculation out there that says that the port is not available.
it's funny that we still haven't had a solid answer for it.

Thanks for the tests, Hunter

If you read the whole post youll get a somewhat solid answer (at the very bottom).
 
Back
Top