GH2 Tests (with a gh13 perspective)

Svart said:
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.
Click to expand...
Its the high shutter/ very fast pan the broke it on a complex scene- not something you would do intentionally if actually shooting something.
 
mojo43 said:
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...
Click to expand...
Objectively speaking, GH13 with 24pN patch works in sometimes unpredictable manner. One of the known issues resulting from such unpredictability is the so called "bitrate pulsing problem," wherein a specific frame (usually the last one in GoP) will have very little bitrate allocated to it, resulting in the "mudded" frame, which is likely what we are seeing in the second "pan" screenshot. In my shots I have rarely seen the problem, but plenty of evidence has been posted that the problem is real. Unfortunately, to date, no solution has been found. So far, the 24p mode looks largely problem free on GH2 (at least as far as problems that plagued GH1/GH13 are concerned). Looking forward to seeing additional side-by-side comparisons. For me 60p is important for "sports" footage.
 
I shoot high shutters but I've only used around 1/800 when I was really stuck and didn't have my NDs on me.
 
svecher said:
Objectively speaking, GH13 with 24pN patch works in sometimes unpredictable manner. One of the known issues resulting from such unpredictability is the so called "bitrate pulsing problem," wherein a specific frame (usually the last one in GoP) will have very little bitrate allocated to it, resulting in the "mudded" frame, which is likely what we are seeing in the second "pan" screenshot. In my shots I have rarely seen the problem, but plenty of evidence has been posted that the problem is real. Unfortunately, to date, no solution has been found. So far, the 24p mode looks largely problem free on GH2 (at least as far as problems that plagued GH1/GH13 are concerned). Looking forward to seeing additional side-by-side comparisons. For me 60p is important for "sports" footage.
Click to expand...

I just checked 24 frames around the grab I posted form the gh1 and its not a codec pulse, all the frames in the pan show break up as in the grab (some more some a bit less). But like I said, when you watch it played in motion, you dont really notice it.

If someone wants to host a prores 422 hq of the panning from both camera's let me know and Ill post them.
 
Last edited:
stonebat said:
Of course there will be some moire, aliasing, mud, and rolling shutter. For pete's sake, it's a camera.
Click to expand...

In extra teleconversion mode that might not be true. From the samples I have seen the ETC mode with the 20mm pancake will produce some astonishing results.
 
thanks for this great test Hunter..
you made me happier with my Japanese purchase.
i'll just have to be careful on those underexposed sensor horizontal atrocities i've got.. and hope those will be firmware fixed.
 
hunter richards said:
IS that a shocker? It shocked me. I was thinking the opposite was going to happen. (before you think I made a huge mistake like I did at first, Yes the gh1 really does have the 44mbs hack enabled Ive triple checked! I used the max 44mbs suggested values in Ptool)
Click to expand...
I appreciate your performing these comparisons, however, I see some serious issues that undermine your conclusions.

In the first pair of images, lens flare from the direct sunlight in the upper left corner severely reduced the contrast and clarity of both images. The composition of the shots was not identical and the GH13 version is overexposed compared to the GH2. Sorry, but I think these problems make the comparison inaccurate.

The second pair of photos makes for a better comparison, but the composition is still not identical. There is enough variation in the framing to account for the subtle difference in exposure and contrast in the two images. For these tests to be valid, the cameras really should be mounted on the same tripod and the documented discrepancies between GH1 and GH2 ISO sensitivity accounted for:

http://www.dxomark.com/index.php/en/Camera-Sensor/Compare-sensors/%28appareil1%29/677|0/%28appareil2%29/630|0/%28onglet%29/0/%28brand%29/Panasonic/%28brand2%29/Panasonic

The third pair of images also shows the built-in discrepancy between GH1 and GH2 exposure settings. When I disregard this factor and examine resolution details the two images look virtually identical to my eyes.

The rest of the images are of the GH2 only and demonstrate that it does indeed produce high quality images. However, I don't think the three comparisons between GH2 and GH13 were precise enough to produce conclusive results. In addition, the 44Mbps GH13 patch that was tested dates back to one of the earliest versions of PTool and does not make use of more recent refinements that Vitaliy has built into PTool 3.5. In my own tests, I have found that these improvements enable higher AVCHD bitrates and can make a noticeable difference in image quality.
 
not to hijack the thread but Lpowell can you point me to those refinements? I have not kept up with the incremental changes at all and as usual everything seems buried in the threads.
 
Guys, regardless of any details with comparision the real point is that the GH2 can do that without breaking. I don't really care if maybe the GH13 can or can't...after all the talk about the Gh2 codec not holding up, we've clearly seen that it can and does. So don't go complaining about how you'd never shoot in this situation, etc. We know! The point is if the GH2 can do this it can handle whatever you throw at it.
 
lpowell, Thanks, I promise im not trying to make the GH13 look any worse (im still keeping my gh13 for use btw) I agree my tests were nominal at best and unscientific, I was just trying to get an out of the box idea on performance.

lpowell said:
I appreciate your performing these comparisons, however, I see some serious issues that undermine your conclusions.

In the first pair of images, lens flare from the direct sunlight in the upper left corner severely reduced the contrast and clarity of both images. The composition of the shots was not identical and the GH13 version is overexposed compared to the GH2. Sorry, but I think these problems make the comparison inaccurate..
Click to expand...

I dont know what to say here, they cameras are literally on top of one another. Let me better explain how I went about this test, I knew that the gh1's iso's were more overrated than the listed ones for the gh2. So I used scene evaluative exposure settings on both to get "the same" exposure (looking at the +/- EV scale in each viewfinder) So the gh1 outside was actually a slightly lower iso to match the GH2.

Here is what I found by shooting a few more scenes this way, even though the histogram and exposure reading matches in camera between the two, the gh1 image always looks milky and the gh2 more punchy. Ill post some better examples of this later today.

Regarding resolution, the gh2 really has an edge, you can see when you get your hands on one- and in my defense, the muddy gh1 frame would still look very similar even if the contrast was higher. (and again, you wont see this issue when shooting normally with it).
 
Great test - I would like to see each camera with the same lens on as a video test.
Gh2 with 20MM anfd then the Gh1 with 20MM.
The one clip right after another and then some slow wipes between them
and then split frames.

And some with (what is the lowest ISO) - 160 or 200??
And then the same thing inside at 800 ISO....

I am getting a GH2 with 14-140 and plan to do some testing
between that and a T2i and my HVX200.
I also am getting adapters for Nikon, Canon and C mount to go the the m4/3
I have plenty of glass in all flavors.
 
Svart said:
not to hijack the thread but Lpowell can you point me to those refinements?
Click to expand...
See the link below in my signature.

Let me emphasize that I understand this was a casual comparison of the two cameras rather than a rigorous test. However, I think the most anyone can honestly claim at this point is that both cameras are capable of producing remarkably high quality results. As with any tool, what's really helpful is to take careful note of its limitations and full advantage of its strengths.
 
Last edited:
lpowell said:
See the link below in my signature.

Let me emphasize that I understand this was a casual comparison of the two cameras rather than a rigorous test. However, I think the most anyone can honestly claim at this point is that both cameras are capable of producing remarkably high quality results. As with any tool, what's really helpful is to take careful note of its limitations and full advantage of its strengths.
Click to expand...

I followed the link in your signature originally but I saw no explanation of "refinements". I see the same settings that I used with 3.2. So let me clarify, what exactly are the refinements/differences that seem to improve the quality? What *is* the increase in quality?
 
Svart said:
...what exactly are the refinements/differences that seem to improve the quality? What *is* the increase in quality?
Click to expand...
My 50Mbps AVCHD Fast Action 3-Frame GOP Patch is able to record FHD and SH videos at up to a 55Mbps maximum bitrate. With well-illuminated, sharply-focused, highly-detailed subject matter, the increased bitrate can make a visible difference in image quality. The 720p30 SH mode produces videos with a 3-frame GOP. When filming rapid motion, recording a new key frame every tenth of a second can make a visible difference in details the AVCHD encoder's motion vectors are not able to capture.
 
lpowell said:
Ok, I was being polite. In my view, the problems I pointed out invalidate the original poster's conclusions.
Click to expand...

This is a rhetorical question and doesn't require an answer. But I will pose it anyway.

If the results had been reversed, and the GH1 with hack had shown better results, would you have still invalidated the OP's response to his tests?

I take issue with the constant naysayers. When you have both cameras and have done your own tests, then I will pay attention. Otherwise it's all pure speculation.

Here at least there is some substance.
 
Hunter;
how did u record the image from HDMI port?
during the live feed mode, do u think a recorder still can record that full quality 1080i image?

Thanks
 
Can you post the natives of the files from all these tests? I would love to see them as I had opposing results as you with motion breakup with lots of foliage in central park. While still shots on a tripod, the GH2 did look better, in massive movement and crowds and trees in wind, I saw a lot of issues, as I posted in my other thread that nearly got my head cut off. ;)

So this leads me to believe that either a> I did in act have a pre production model and something has been improved, or b> you didn't capture as much motion as I did.

I hope it's A, but I really want to see the native files of both rather than stills.

- Blackout
 
You must log in or register to reply here.
Back
Top