This test is an ISO test, focusing mostly on noise performance in typical filming conditions (not extreme low light)
In this test, both cameras sat at the same aperture, but different lenses and shutter speeds in order to let each camera be properly exposed. Properly exposed in this test was defined as a point at which there is no clipping in the highlights. So both cameras were exposed to the right as far as they could go without any clipping.
Because noise performance is hard to see clearly on compressed web video - I have saved these off as high quality PNG files.
I know the BMCC has a native iso of 800, and its only really metadata - whereas the 5dmkii has a built in analogue gain.
Both clips were processed the exact same, except that their exposure level was neutralised to about the same point from original ETTR exposure points - because the BMCC has higher dynamic range, it had to be pulled back down slightly further. This might lead some to claim the test is unfair - but in a real world shooting condition, when filming with raw video, ETTR is a good idea, and the degree at which each camera is able to be filmed this way is a clear indication of the overall performance of a camera, as such it is a valid comparison. Because both cameras were prevented from clipping, both cameras were able to use their full range. Neither cameras were pushed in post production.
To reiterate:
- SAME ISO EACH TIME -
- BOTH CAMERAS EXPOSED TO THE RIGHT BUT PREVENTED FROM CLIPPING -
- APART FROM EXPOSURE ADJUSTMENT DUE TO ETTR, BOTH CAMERAS WERE PROCESSED THE EXACT SAME WAY -
- NO SHARPENING OR NOISE REDUCTION WAS USED -
- DIFFERENT LENSES BUT AT SAME APERTURE EACH TIME -
- FILMED AT THE EXACT SAME MOMENT TO PREVENT AMBIENT LIGHT SHIFT -
- VARIATIONS IN LENS GLASS MAY INTERFERE WITH COLOURS, BUT SHOULD NOT AFFECT ISO NOISE PERFORMANCE -
DOWNLOAD LINKS:
ISO 200: http://we.tl/tUlTZdthKc
ISO 800: http://we.tl/Xiu1s8lsaB
ISO 1600: http://we.tl/srZvEgKoKJ
Quick Jpeg preview versions (quality level 12) for web browser viewing:
https://dl.dropboxusercontent.com/u/544 ... iso200.jpg
https://dl.dropboxusercontent.com/u/544 ... iso800.jpg
https://dl.dropboxusercontent.com/u/544 ... so1600.jpg
Noise Performance considerations
The 5D Mkii does very well in low light, high iso situations. This is due to many reasons including its large photosite size, good internal noise reduction and analogue gain on the sensor. The BMCC does not have analogue gain or internal noise reduction, but it actually has larger photosites and better quantum efficiency than the sensor inside the 5d Mkii. So it would be reasonable to suggest that up to around iso 800, the bmcc could have better noise performance than the 5d Mkii. When the 5d Mkii is in iso 1600, it has analogue gain applied, which really helps to boost SNR, when the bmcc is at iso 1600 it is actually shooting at 800 with digital gain. While it is true to say that at very high iso's, the 5d Mkii is undoubtedly better in terms of noise than the BMCC - I have often found that the BMCC is actually cleaner at 800 and below than the 5d Mkii. Considering ive always tried to avoid filming above 800 on my 5D when it was my primary camera, this is quite interesting to me now that I have the BMCC. I am interested to see your own thoughts on this and if you agree with my own assessment, or perhaps wish to offer your own.
Remember, this situation is not a low light test. At ISO1600 in these shots, both cameras were at f10 - which if such a scene demanded a deep DOF, you would need to shoot at 1600 even in a well lit scene like this.
In this test, both cameras sat at the same aperture, but different lenses and shutter speeds in order to let each camera be properly exposed. Properly exposed in this test was defined as a point at which there is no clipping in the highlights. So both cameras were exposed to the right as far as they could go without any clipping.
Because noise performance is hard to see clearly on compressed web video - I have saved these off as high quality PNG files.
I know the BMCC has a native iso of 800, and its only really metadata - whereas the 5dmkii has a built in analogue gain.
Both clips were processed the exact same, except that their exposure level was neutralised to about the same point from original ETTR exposure points - because the BMCC has higher dynamic range, it had to be pulled back down slightly further. This might lead some to claim the test is unfair - but in a real world shooting condition, when filming with raw video, ETTR is a good idea, and the degree at which each camera is able to be filmed this way is a clear indication of the overall performance of a camera, as such it is a valid comparison. Because both cameras were prevented from clipping, both cameras were able to use their full range. Neither cameras were pushed in post production.
To reiterate:
- SAME ISO EACH TIME -
- BOTH CAMERAS EXPOSED TO THE RIGHT BUT PREVENTED FROM CLIPPING -
- APART FROM EXPOSURE ADJUSTMENT DUE TO ETTR, BOTH CAMERAS WERE PROCESSED THE EXACT SAME WAY -
- NO SHARPENING OR NOISE REDUCTION WAS USED -
- DIFFERENT LENSES BUT AT SAME APERTURE EACH TIME -
- FILMED AT THE EXACT SAME MOMENT TO PREVENT AMBIENT LIGHT SHIFT -
- VARIATIONS IN LENS GLASS MAY INTERFERE WITH COLOURS, BUT SHOULD NOT AFFECT ISO NOISE PERFORMANCE -
DOWNLOAD LINKS:
ISO 200: http://we.tl/tUlTZdthKc
ISO 800: http://we.tl/Xiu1s8lsaB
ISO 1600: http://we.tl/srZvEgKoKJ
Quick Jpeg preview versions (quality level 12) for web browser viewing:
https://dl.dropboxusercontent.com/u/544 ... iso200.jpg
https://dl.dropboxusercontent.com/u/544 ... iso800.jpg
https://dl.dropboxusercontent.com/u/544 ... so1600.jpg
Noise Performance considerations
The 5D Mkii does very well in low light, high iso situations. This is due to many reasons including its large photosite size, good internal noise reduction and analogue gain on the sensor. The BMCC does not have analogue gain or internal noise reduction, but it actually has larger photosites and better quantum efficiency than the sensor inside the 5d Mkii. So it would be reasonable to suggest that up to around iso 800, the bmcc could have better noise performance than the 5d Mkii. When the 5d Mkii is in iso 1600, it has analogue gain applied, which really helps to boost SNR, when the bmcc is at iso 1600 it is actually shooting at 800 with digital gain. While it is true to say that at very high iso's, the 5d Mkii is undoubtedly better in terms of noise than the BMCC - I have often found that the BMCC is actually cleaner at 800 and below than the 5d Mkii. Considering ive always tried to avoid filming above 800 on my 5D when it was my primary camera, this is quite interesting to me now that I have the BMCC. I am interested to see your own thoughts on this and if you agree with my own assessment, or perhaps wish to offer your own.
Remember, this situation is not a low light test. At ISO1600 in these shots, both cameras were at f10 - which if such a scene demanded a deep DOF, you would need to shoot at 1600 even in a well lit scene like this.
Tom Majerski
Colourist at Tracks and Layers
http://www.Tracksandlayers.com
Motion Graphics - Colour Grading - VFX
Colourist at Tracks and Layers
http://www.Tracksandlayers.com
Motion Graphics - Colour Grading - VFX