Dark Current Noise at DVinfo.net

Aaron Shaw · December 16th, 2004, 05:11 PM

Hi everyone,

I'm just curious if prosumer cameras actually remove the dark current noise from the CCD before recording to tape. I'd like to know so I can be prepared for it if it happens.

I'm not sure if this would matter much at the exposure lengths used with video though.

Anyone have any information about this?

Vladimir Koifman · December 19th, 2004, 01:39 PM

I dont think they do. Leakage strongly depends on temperature, one needs to measure it dynamically. It's too much a complication.
Also, leakage subtraction process increases other noise sources, so it does not makes sense to do it unless leakage is by far the dominating noise source.

Andre De Clercq · December 19th, 2004, 02:47 PM

Dark current is the most important noise source in CCD cams, and strongly determines the dynamic range. The average (temp dependant ) component (DC component) is always substracted (black level reset). The AC component cannot be canceled (random) and it is this component that makes pics noisy at high gain settings. The HAD (hole accumulation diode) architecture from Sony significantly reduces dark current (AC an DC components).

Aaron Shaw · December 19th, 2004, 03:55 PM

Wow, thanks for the feedback guys.

I was not aware that dark current was tied so strongly to dynamic range. Very interesting.

Vladimir Koifman · December 20th, 2004, 02:13 PM

<<<-- Originally posted by Andre De Clercq : Dark current is the most important noise source in CCD cams, and strongly determines the dynamic range. The average (temp dependant ) component (DC component) is always substracted (black level reset). The AC component cannot be canceled (random) and it is this component that makes pics noisy at high gain settings. The HAD (hole accumulation diode) architecture from Sony significantly reduces dark current (AC an DC components). -->>>

Actually, a portion of this AC component is not entirely random and can be subtracted, at least theoretically. For instance, some digital still cameras subtract this fixed pattern leakage noise component as measured from a dark frame exposure right after a shot taken.
However, it's more complex to implement this in video cameras.

Andre De Clercq · December 20th, 2004, 03:20 PM

Right Vladimir. Indeed in still camera's and industrial linear CCD scanners the fixed pattern noise and QE differences are being compensated in by automated lookup achitectures in order to get reliable imaging.

December 16th, 2004, 05:11 PM	#1
Aaron Shaw Major Player Join Date: Jul 2004 Posts: 587	Dark Current Noise Hi everyone, I'm just curious if prosumer cameras actually remove the dark current noise from the CCD before recording to tape. I'd like to know so I can be prepared for it if it happens. I'm not sure if this would matter much at the exposure lengths used with video though. Anyone have any information about this?

December 19th, 2004, 01:39 PM	#2
Vladimir Koifman Regular Crew Join Date: May 2003 Location: Israel Posts: 115	I dont think they do. Leakage strongly depends on temperature, one needs to measure it dynamically. It's too much a complication. Also, leakage subtraction process increases other noise sources, so it does not makes sense to do it unless leakage is by far the dominating noise source.

December 19th, 2004, 02:47 PM	#3
Andre De Clercq Major Player Join Date: Sep 2002 Location: Belgium Posts: 804	Dark current is the most important noise source in CCD cams, and strongly determines the dynamic range. The average (temp dependant ) component (DC component) is always substracted (black level reset). The AC component cannot be canceled (random) and it is this component that makes pics noisy at high gain settings. The HAD (hole accumulation diode) architecture from Sony significantly reduces dark current (AC an DC components).

December 19th, 2004, 03:55 PM	#4
Aaron Shaw Major Player Join Date: Jul 2004 Posts: 587	Wow, thanks for the feedback guys. I was not aware that dark current was tied so strongly to dynamic range. Very interesting.

December 20th, 2004, 02:13 PM	#5
Vladimir Koifman Regular Crew Join Date: May 2003 Location: Israel Posts: 115	<<<-- Originally posted by Andre De Clercq : Dark current is the most important noise source in CCD cams, and strongly determines the dynamic range. The average (temp dependant ) component (DC component) is always substracted (black level reset). The AC component cannot be canceled (random) and it is this component that makes pics noisy at high gain settings. The HAD (hole accumulation diode) architecture from Sony significantly reduces dark current (AC an DC components). -->>> Actually, a portion of this AC component is not entirely random and can be subtracted, at least theoretically. For instance, some digital still cameras subtract this fixed pattern leakage noise component as measured from a dark frame exposure right after a shot taken. However, it's more complex to implement this in video cameras.

December 20th, 2004, 03:20 PM	#6
Andre De Clercq Major Player Join Date: Sep 2002 Location: Belgium Posts: 804	Right Vladimir. Indeed in still camera's and industrial linear CCD scanners the fixed pattern noise and QE differences are being compensated in by automated lookup achitectures in order to get reliable imaging.