C300: C 300 MK II - White Paper (PDF)

AndreeOnline said:
I know it's tempting to just allocate fewer values per stop and think you get more stops, but you don't. Those would just be fractions of a stop in relation to real light.
Click to expand...

Thank you for that list of values. It helps to see it laid out that way. I think I now know why I've never been able to understand this discussion about bit rate and stops. I've never really thought that the purpose of a video camera was to capture all of the data available in real light. I always thought in terms of wanting an image to have blacks that appear black and whites to appear whites. In low contrast situations, this requires either a picture profile that reduces captured number of stops or color correction that increases the image contrast. For low bit rate codecs, its best to change the picture profile or lighting rather than wait until post. In very high contrast situations, it's nice to have a camera that can actually capture a high dynamic range and compress it so that the bright objects don't blow out and look unnatural. In the end, the actual dynamic range of the shot is going to be compressed way down for the viewing screen anyway. Very few viewing screens actually exhibit a wide dynamic range and the audience is used to seeing a highly compressed image.
 
it seems like shooting at 2K RGB 444 12-bit produces the best image out of the sensor or am I missing something...
 
cpreston said:
Thank you for that list of values. It helps to see it laid out that way.
Click to expand...

Yes, I think it does too. And at the same time it becomes easier to understand why the last few additional stops are hard to come by. You need to significantly improve the signal-to-noise ratio just to get one more stop (in addition to upgrade the internal components to a higher precision like 16bit).

And just like with displays that—in order to achieve higher contrast numbers, it's mostly a question of lowering the black values—in cameras you want to lower the sensor/readout/photon noise as low as you can as that will help DR a lot. That has long been Sony/Nikon/Pentax advantage in still cameras in low ISOs.
 
jtten said:
it seems like shooting at 2K RGB 444 12-bit produces the best image out of the sensor or am I missing something...
Click to expand...

It's the highest spec format, if you leave resolution out of it.

But since the C300 mkII does some interesting internal upscaling of the individual color components before reconstructing the 4k image, it will be interesting to compare the two options for 2k delivery.

Once the 4k image is scaled to 2k in post, I don't think there will be any advantages in terms of resolved detail compared to the 2k 12bit image, since that too is derived from internal 4k upsampling of the color components and THEN downsampling.

The 12 bit 2k image will be the fatter file, while 4k has the advantage of re-framing/stabilizing.
 
On the C500, the difference between 12bit and 10bit was just not existent in the images. They looked the exact same colours even after scientific bit depth analysis and skin colours.

The 4K ProRes 10bit downscaled produced the best image. If the C300II is the same I wouldn't leave the extra resolution for 12bit, unless of course no post-production for downscaling is scheduled.
 
The DP and director called out 12-bit 2K capture on the C500 as having the best look. Just as countless DP's and directors prefer to shoot on Alexa vs. 4K cameras. In an HDR display world, tonality will ultimately trump resolution, IMO.
 
You must log in or register to reply here.
Back
Top