Hi there
My applications commonly need low light performance.
If you want to understand why, please check our website : http://kimchiandchips.com
We have a BlackMagic Production 4K Camera. We bought it for a number of purposes (including scanning/calibrating).
I've noticed that its low light performance doesn't match a DSLR (neither does a RED for that matter). But for our purposes it's really important to record with a global shutter and we'd like to use our BMPC in low light situations.
(low light + global shutter + 4k is basically impossible I know)
Generally our delivery format is 1080p, and i'm wondering if there's a way to squeeze better low light performance out of the camera by adding together the pixels of each 2x2 pixel block when down-sampling to HD.
In theory, down-sampling should therefore give us 2 stops of extra low light performance. I'm curious if selecting the HD mode in camera (downsample before recording) has the same result. And if anybody has any experience with this?
In theory, if we record in 4k, and then downsample to HD afterwards, then the downsampled video will have less 1/4 as much noise per pixel. Meaning that if we add 4x gain in post (e.g. DaVinci Resolve), the post-amplified image should be the same as just adding together the pixel values in the camera
Pixels A, B, C, D in a 2x2 block:
Downsample then amplify x4 : (A + B + C + D) / 4 * 4
Simply add values : A + B + C + D
(these two are equivalent)
Personally, I think it would be great just to have higher ASA options (1600, 3200) in HD mode. This could be added with a firmware update and would kick up a lot of positive interest for the camera. But it may turn out that the same amount of noise at HD is unacceptable, whilst at 4k, it's too small to notice.
(PS : I don't have the camera with me now, it may turn out that there's already high ASA options for HD recording already. Which would be great!)
Elliot
My applications commonly need low light performance.
If you want to understand why, please check our website : http://kimchiandchips.com
We have a BlackMagic Production 4K Camera. We bought it for a number of purposes (including scanning/calibrating).
I've noticed that its low light performance doesn't match a DSLR (neither does a RED for that matter). But for our purposes it's really important to record with a global shutter and we'd like to use our BMPC in low light situations.
(low light + global shutter + 4k is basically impossible I know)
Generally our delivery format is 1080p, and i'm wondering if there's a way to squeeze better low light performance out of the camera by adding together the pixels of each 2x2 pixel block when down-sampling to HD.
In theory, down-sampling should therefore give us 2 stops of extra low light performance. I'm curious if selecting the HD mode in camera (downsample before recording) has the same result. And if anybody has any experience with this?
In theory, if we record in 4k, and then downsample to HD afterwards, then the downsampled video will have less 1/4 as much noise per pixel. Meaning that if we add 4x gain in post (e.g. DaVinci Resolve), the post-amplified image should be the same as just adding together the pixel values in the camera
Pixels A, B, C, D in a 2x2 block:
Downsample then amplify x4 : (A + B + C + D) / 4 * 4
Simply add values : A + B + C + D
(these two are equivalent)
Personally, I think it would be great just to have higher ASA options (1600, 3200) in HD mode. This could be added with a firmware update and would kick up a lot of positive interest for the camera. But it may turn out that the same amount of noise at HD is unacceptable, whilst at 4k, it's too small to notice.
(PS : I don't have the camera with me now, it may turn out that there's already high ASA options for HD recording already. Which would be great!)
Elliot