Low Light Streaks - Is my camera faulty?

For what it's worth. I set the camera up a few feet from an off white wall and set the image and exposure to bring out the harshest FPN I could get. Lens cap off, LCD shut and left it overnight with out changing any light levels. Did this a couple of nights and noticed quite a change. FPN is still there but nothing near as a bad. Others have tried different tactics with varying results.

Agreed, ISO isn't the problem; it's exposure levels. I also think some of us GH1 owners have different levels of problems with this issue so mileage may vary. I briefly had 2 cams and one was, without a doubt, slightly worse than the other for FPN. I, sadly, didn't have it long enough to really analyze the difference. Sorry about that.
 
Okay, test is underway. I took a shot under controlled conditions, 3 stops under, and yep, there's banding.

So now it's going to "burn in" for around 24 hours, then I will take the exact same shot again and be able to compare as to whether there's any validity to the "burning in" theory.
 
The crazy thing is that for me, the FPN shows up on areas of the image NOT underexposed...

Let's say I have a shot where half the frame is dark, and the the other half of the frame is properly exposed but is a solid color wall. The FPN doesn't show up in the dark shadowed half of the frame, just on the properly exposed wall.
 
Hi again,

just to clarify:

I followed what I read here in this thread simply out of desperation. I did my first shoots in that same bar and I was really disappointed by those streaks. Before that I never even read this thread and just thought by the headline that some poor dude had a faulty camera :)

Lens on? yes (kit lens).
Procedure? Pointed at really dark area, ISO 1600, left running one night long.
Improvement? Yes! No streaks at ISO 1600.
Could I be better exposing now that I did in my first clips? Possible. I have better glass now than I did before, but I want to emphasize that I never went for underexposure. So if the image looks equally exposed to me and has the same ISOs and shutter speeds in both streaky and non-streaky cases something else has to have changed, doesn´t it?

I´m really looking forward to the results of Barry Green´s tests.

Just imagine if he proves the burn-in an illusion! Which would mean that I and others simply improved our gh1-skills. :) not the worst outcome...
 
I realise that the problem is underexposure but - you could say it's overexposure of the bits where banding is showing.

Think of it this way. If I record a shot in my studio here, where there's a 9 watt fluoro in the the desk lamp and no other illumination (well, apart from 2 22" displays!), and say I've got a white document on the desk - if I expose the scene so that the print on the white document is legible and it's not overexposed and thus all white, then the light cream coloured walls and white book-cases disappear into black - no banding. If on the other hand I open the aperture a bit, so that the paper on the desk is a blown highlight, then the background is also lightened, no longer crushed to black, and banded.

So if you've got at least something in the shot which you can correctly expose without using very high iso or whatever, then the rest should be ok. If however the whole scene is effectively too dark to get any reasonable result from (lowish contrast within the shot probably), and you crank up the gain, you'll see the bands because you are over-exposing stuff which should be dark.

Does that make sense? OK, ok....
 
Ozpeter said:
I realise that the problem is underexposure but - you could say it's overexposure of the bits where banding is showing.

...

So if you've got at least something in the shot which you can correctly expose without using very high iso or whatever, then the rest should be ok. If however the whole scene is effectively too dark to get any reasonable result from (lowish contrast within the shot probably), and you crank up the gain, you'll see the bands because you are over-exposing stuff which should be dark.

Does that make sense? OK, ok....
Click to expand...

Ok, besides that I obviously will look at my display and do whatever it takes to get rid of banding you basically suggest "threshold-based lighting"?
Like if I can only get it (it being the background or anything but the main object of the shot) lit with banding I should just go for black and bandingfree instead?
Pure black is definitely more appealing than banded noise.

Hmmm...
 
Okay, I set up a shot which involved a white sheet of paper, some ND gel on part of it, some minusgreen gel on other parts of it, and a cardboard box which was roughly flesh-toned, and shot the very first shot on the GH1. ISO 200, f/4.3, 1/60 shutter, everything about 3-stops underexposed. Easily visible banding on the white paper and some on the box.

I then let the system "burn in" for 24 hours, as prescribed in post 305.

I then shot the exact same setup, same exposure, same lighting, same everything, same tripod position (the camera hadn't moved, the shot was identical from before). And the banding is identical.

If I was "MythBusters" I'd say the myth about "burn-in" is "Busted".
 
editman said:
So it's busted and we all have to do proper exposure and lighting as it should be.
Click to expand...

true. Still, I feel a bit silly now for believing in voodoo instead of trying to improve my skills.

But thanks for making that test Barry Green, clarity once and for all.

Maybe your results should find a place in the faqs or get sticky or something. Once Panasonic starts delivering real amounts of GH1s this streaky question flow will rise again.
 
Ok, can't argue with Barry's test, and thank you, but I wasn't alone in seeing a lessening in the FPN with either the burn in or simply usage over a period of time. I know I saw a difference. Kholi, commanderspike, Martti Eckland and others experienced less FPN as time went by. Is there anything else we can think of that would explain the camera's FPN reduction as usage increases? My camera still has banding but it is not anything like it was when I first got it. For now I'm happy voodoo worked for me.
 
Barry_Green said:
ISO 200, f/4.3, 1/60 shutter, everything about 3-stops underexposed. Easily visible banding on the white paper and some on the box.
Click to expand...

Doubt it makes any difference but I made sure the ISO was set to 1600, shutter 1/60, no idea on the f stop, just to ensure I was generating as much noise as possible. By the time I began it looked like I was shooting through a brown cheese cloth. Simply as bad as I could make it.

Sorry guys, I'm not trying to beat a dead horse here but I think there's a connection between camera usage and FPN reduction. I just don't know what it is.
 
Last edited:
Nighthawk said:
Is there anything else we can think of that would explain the camera's FPN reduction as usage increases? My camera still has banding but it is not anything like it was when I first got it. For now I'm happy voodoo worked for me.
Click to expand...

Familiarity with it would probably make a difference. The more you shoot with it the better you get at not underexposing shots.
 
Isaac_Brody said:
Familiarity with it would probably make a difference. The more you shoot with it the better you get at not underexposing shots.
Click to expand...

You're definitely right about that when it came to actually shooting with the cam but it doesn't explain the change I saw when doing the burn in. I was also wondering about batch lots as far as sensor production was concerned as it seemed some of us had bigger issues than others about FPN. I guess I'll just have to let it go and be one of those in the 'voodoo' camp for now. After all I was covered in chicken blood at the time.
 
Nighthawk said:
Doubt it makes any difference but I made sure the ISO was set to 1600, shutter 1/60, no idea on the f stop, just to ensure I was generating as much noise as possible. By the time I began it looked like I was shooting through a brown cheese cloth. Simply as bad as I could make it.

Sorry guys, I'm not trying to beat a dead horse here but I think there's a connection between camera usage and FPN reduction. I just don't know what it is.
Click to expand...

I'm not sure but if you had ISO already set to 1600 it means you were trying to compensate for light which means that it might be case that you were underexposing due to not enough light.

Just to test the case of streaks due to less light, I tested the GH1 in normal room lights. Initially there were streaks when I captured the room surroundings(except pointing towards light source), but as soon as I pointed the camera towards lights and the exposure was balanced, the streaks disappeared.

I did this test with brand new GH1 without any burn in period as such. Maybe you can try it yourself in the average lit room - you will see the streaks when you will pan around the room without pointing at the light source, but as soon as you will start to point towards the light, the streaks will start to disappear making it evident that it's the underexposure issue...
 
Nighthawk said:
Doubt it makes any difference but I made sure the ISO was set to 1600, shutter 1/60, no idea on the f stop, just to ensure I was generating as much noise as possible.
Click to expand...
I did the same.

I took my shot at 200 because that generated the streaks. Then I changed it to 1600 and let it sit overnight. Then I came back 24 hours later, changed the ISO back to 200, and took the follow-up shot.

Sorry guys, I'm not trying to beat a dead horse here but I think there's a connection between camera usage and FPN reduction. I just don't know what it is.
Click to expand...
It's the exact same phenomenon we see with every other camera. Some horrible flaw pops up, and then, a few days later, you just don't notice it anymore, whether it's HVX "noise" or EX1 "jello" or JVC "split screen" or every CMOS camcorder's FPN... Did it fade away? No, it's that as you work with the camera more, you learn to use it properly, and you just don't get in the circumstances that cause the issue.

The very first shots I ever took with a GH1, back in April at NAB, had horrible streaking/banding throughout the subject's face. Now, we shot two entire projects without a streak to be seen. What's the difference? We know how to use it now.

I can produce the banding at will, any time I want. And I can make it go away, by shooting at proper exposure.

I think it's just experience with the camera that is lessening the appearance of the streaking.

But hey, if anyone wants to burn in for 24 hours, you're not going to hurt anything. Just don't let that become a substitute for getting experience with the camera, learning its exposure characteristics, learning the histogram, and getting better with using it.
 
You must log in or register to reply here.
Back
Top