GH2 ISOs to avoid...Test results!

roei z said:
thanks,
but your GH13 test method is not the same as what he did on the GH2.. try duplicating the same test as he explained
Click to expand...

Um i did exactly what he did according to his explanation. Lens cap on, SMOOTH -2 settings, Recorded each iso, FCP 3way CC with mids maxed to the right and saturation maxed. Is there something I am missing?

Also I even tried to see if there would be a difference in f4 and f22 with the lens cap on to see if there would be any noise difference, but it was the same. Maybe its just me or I'm baffled @ how much better the test results yielded for the GH13. Like I said unless I'm missing something.
 
cannot help you with that.. sent him a message and ask him, cause there's obviously something different.
it should look at least close to what he did and it's definitely not.
 
roei z said:
cannot help you with that.. sent him a message and ask him, cause there's obviously something different.
it should look at least close to what he did and it's definitely not.
Click to expand...

It might be the codec, I dont have any NR or IS on, but even then when recording pure black the mbps shouldn't produce high numbers. I'll shoot him a PM, but if we did everything thing the same... HURRAY to the GH13. LOL!
 
don't think so, because the banding line shows where it usually at on a full frame. or i might be mistaken.
and also, the GH13 test somehow shows NO noise. must be some kind of different settings.
 
@Jonathan Ruiz...There's something wrong with that test. All these cameras produce noise at different levels...it doesn't matter the brand. I just sold my GH-13...it...like every other camera has noise especially when you go up the ISO ladder (and I'm not even mentioning the FPN and banding).
 
I performed an ISO test, too and contrary to your noisefloor test in real world ISO 320 and 640 are worse than 200,400 and 800!
http://www.vimeo.com/18246626

320 and 640 have colored nasty noise while the other modes only have luminance noise.
Maybe ma camera is different. German PAL modell.
ISOs 2500 and 3200 look awful flickering in my camera.
 
Jonathan12uiz said:
Um i did exactly what he did according to his explanation. Lens cap on, SMOOTH -2 settings, Recorded each iso, FCP 3way CC with mids maxed to the right and saturation maxed. Is there something I am missing?
.
Click to expand...

Jonathan,

I think perhaps what you did was adjust the "Blacks" slider instead of the "Mids" slider all the way to the right. You may want to check that. Let us know. In the meantime, I may test my unhacked GH1 and see if it behaves differently than the GH2. Might provide some good comparison info., especially if you provide GH13.
 
FelixGER said:
I performed an ISO test, too and contrary to your noisefloor test in real world ISO 320 and 640 are worse than 200,400 and 800!
http://www.vimeo.com/18246626
320 and 640 have colored nasty noise while the other modes only have luminance noise.
Click to expand...
Wow, you're right. Those conclusions are somewhat different. I can still see the luminance noise increase on 250, 500 etc. But yes, 320 and 640 look quite terrible. Looks like another round of shooting is in order for me (or somebody). I suppose my test of shooting only black is of limited use. Perhaps there is noise that only presents itself when pixels are exposed to light. I remember on the GH1 that the ugliest stuff showed up in the mid to deep shadows. Not black, not midtones, but right in between. In your test as well, the bright wall in the middle looks OK at all ISOs but the darker areas really suffer from chroma noise at the ISOs you mention.
Somebody did a test early on of NR levels on high ISO clips. It might be interesting for you to do a similar test but set NR to zero. If that gets rid of the 320/640 ISO noise problem without affecting resolution and clarity, then maybe it's a non-issue.
 
dvbrother said:
Jonathan,

I think perhaps what you did was adjust the "Blacks" slider instead of the "Mids" slider all the way to the right. You may want to check that. Let us know. In the meantime, I may test my unhacked GH1 and see if it behaves differently than the GH2. Might provide some good comparison info., especially if you provide GH13.
Click to expand...

Nope, I've been editing with FCP since 2004, I've adjusted the Mids, you can tell because it would be black not gray. Also there is Noise at the far right apparent @ 1250 and 1600 iso, just not a lot so its hard to miss.

Maybe its the fact that my cap is actually a good seal, not leaking any light in, where as your cap was not fully on... letting enough light get in to show differences in the sensor.

Also I shot in FHD, did you shoot in SHD? if you give me the exact method then I will do it again, but as far as I know I did everything exactly as you did. I took a screenshot earlier, but I didn't upload it and already deleted it. It shows the CC in FCP maxing the mids and saturation.
 
That 2nd test couldn't have been done right cause it's a lavender color.. unless that was tinted in post to see something different.


BTW... The short sketch at the end looked really good... Very nice, very cinematic looking... What lens did you use?
 
Justyn said:
The short sketch at the end looked really good... Very nice, very cinematic looking... What lens did you use?
Click to expand...

I used the Panasonic 20mm f1.7 for the shots in the computer room. The first 2 shots I put a .7 Wide Angle Adapter on the 20mm. Some of the other shots were done with an old Nikon 50mm f1.4. Just got it and had to try it out. Thanks!
 
Justyn said:
That 2nd test couldn't have been done right cause it's a lavender color.. unless that was tinted in post to see something different.


BTW... The short sketch at the end looked really good... Very nice, very cinematic looking... What lens did you use?
Click to expand...

It was done replicating the same steps as the OP. no other modifications.
 
FelixGER said:
I performed an ISO test, too and contrary to your noisefloor test in real world ISO 320 and 640 are worse than 200,400 and 800!
http://www.vimeo.com/18246626

320 and 640 have colored nasty noise while the other modes only have luminance noise.
Click to expand...

I just did a more real-world test, and sure enough 640 ISO exhibits more chroma noise than 800, even when NR is turned up to 0, you can still see the difference. 320 didn't look bad in my test. My experiences with 800 ISO on the GH2 have been that it's acceptable, and in some ways it looks like film grain. Chroma noise on the other hand is just plain ugly. Looks like I'll be avoiding 640 ISO.
 
Last edited:
I experienced that NR only improves the picture above ISO 1250. Below it desroys the film-like grain and makes mud of it.
I´m using 160, 200, 400, NR-2 without tweking with Neat Video.
ISO 800 is as clean as ISO 200 if you gently denoise it with Neat Video without losing visible resolution (10 % luma / 30 % chroma setting).
 
Last edited:
dvbrother said:
Native ISOs in a camera use analog gain, while the others use digital gain or reduction. 200, 400, 800, and 1600 are the native ISOs of the GH2. The ISOs below these native ones (160, 320, 640, 1250) use 1/3 stop of digital gain reduction (kind of like some professional video cameras let you choose -3 gain for cleaner blacks). ISOs 250, 500, 1000, and 2000, which are the ones just above the native ISOs, use 1/3 stop digital gain to boost the sensitivity, and that's why they exhibit more noise than the others.
Click to expand...
Yet more evidence that the archaic ISO "film speed" model is obsolete and degrades the performance of digital cameras. I'll have to point out, however, that the concept of "native ISO" for a digital camera is bunk. The light intensity levels produced by the photosensitive elements in an image sensor are inherently analog. When video gain is properly applied to these signals before they are digitized, noise will be minimized and rise only gradually as the gain is increased. Altering the video gain after it has been digitized to simulate 1/3rd-stops inevitably cheapens the image quality of the camera. I'd find this kind of engineering compromise acceptable in an inexpensive cell phone camera, but for a top-of-the-line camera like the GH2, with a list price of $900? I certainly hope Panasonic has not shortchanged the AF100 in this manner as well.
 
lpowell said:
Yet more evidence that the archaic ISO "film speed" model is obsolete and degrades the performance of digital cameras. I'll have to point out, however, that the concept of "native ISO" for a digital camera is bunk. The light intensity levels produced by the photosensitive elements in an image sensor are inherently analog. When video gain is properly applied to these signals before they are digitized, noise will be minimized and rise only gradually as the gain is increased. Altering the video gain after it has been digitized to simulate 1/3rd-stops inevitably cheapens the image quality of the camera. I'd find this kind of engineering compromise acceptable in an inexpensive cell phone camera, but for a top-of-the-line camera like the GH2, with a list price of $900? I certainly hope Panasonic has not shortchanged the AF100 in this manner as well.
Click to expand...

I think it is common for CMOS cameras.
As I understand modern sensors have amplifiers near photosites. So, they can be source of limitations in analog amplification.
Plus ADC number is also high, so they also must be simple (and more programmed gain steps means complicated design, plus it is hard to make identical ADC at so many steps).
Look at this old paper - http://www.stw.tu-ilmenau.de/~ff/beruf_cc/cmos/cmos_noise.pdf
At least GH1 seems to be column based sensor (looking at FPN).
Gh2 seems to be row based (hence horizontal bands).
 
Jonathan12uiz said:
Um i did exactly what he did according to his explanation. Lens cap on, SMOOTH -2 settings, Recorded each iso, FCP 3way CC with mids maxed to the right and saturation maxed. Is there something I am missing?
Click to expand...

OK, are you ready for this? I tested my GH1 the same way I tested my GH2...and got the exact same results you did! No noise shows up until 1250 ISO!

So we ARE both applying the CC in the same way. The GH1 and the GH2 are different when it comes to noise in the blacks.

Just to make sure, I re-did my GH2 test with aperture at f16 and shutter speed at 1500 just to make sure it wasn't a light leakage issue. The GH2 performed like it did the first time! I also tested the GH2's 17mbps 1080i setting. Same noise as the first test. So there is a massive difference between the two cameras.
Certainly, the GH2 is not as noisy at 800 through 1600 ISO as the GH1 when shooting normal footage. So something interesting is going on. The recent post by Vitaly may be a clue. If the GH1 was a column based sensor (hence the FPN problem), and the GH2 is row-based (hence the horizontal band issue), maybe more noise creeps up into the deep shadows from the blacks in the GH2...enough to read when I crank the midtones in FCP. Bizarre.
 
Last edited:
dvbrother said:
OK, are you ready for this? I tested my GH1 the same way I tested my GH2...and got the exact same results you did! No noise shows up until 1250 ISO!

So we ARE both applying the CC in the same way. The GH1 and the GH2 are different when it comes to noise in the blacks.

Just to make sure, I re-did my GH2 test with aperture at f16 and shutter speed at 1500 just to make sure it wasn't a light leakage issue. The GH2 performed like it did the first time! I also tested the GH2's 17mbps 1080i setting. Same noise as the first test. So there is a massive difference between the two cameras.
Certainly, the GH2 is not as noisy at 800 through 1600 ISO as the GH1 when shooting normal footage. So something interesting is going on. The recent post by Vitaly may be a clue. If the GH1 was a column based sensor (hence the FPN problem), and the GH2 is row-based (hence the horizontal band issue), maybe more noise creeps up into the deep shadows from the blacks in the GH2...enough to read when I crank the midtones in FCP. Bizarre.
Click to expand...

Interesting find!
 
You must log in or register to reply here.
Back
Top