Page 1 of 1

444 OR 422 - color space can shoot BMCC ?

PostPosted: Fri May 31, 2013 6:05 pm
by Valdo Rochas
Hi everyone
can anyone tell me about BMCC color space i mean it has 444 or 422 ?
how much color space is when i shoot in RAW dng ?
or maybe RAW has 422 and Prores has 444 ? I AM REALLY CONFUSED
how does SDI output related to 444 ?
hellp ...

Re: 444 OR 422 - color space can shoot BMCC ?

PostPosted: Tue Jun 04, 2013 6:45 pm
by Christine Peterson
Moved to Cinematography...

It's 4:2:2. You can find these details on the specs page, here: http://www.blackmagicdesign.com/product ... techspecs/

Re: 444 OR 422 - color space can shoot BMCC ?

PostPosted: Wed Jun 05, 2013 6:07 pm
by Jonas Bengtson
RAW is RAW. No subsampling. Just pixels from the sensors.

SDI out is 4:2:2

Re: 444 OR 422 - color space can shoot BMCC ?

PostPosted: Wed Jun 05, 2013 6:10 pm
by Tom
Jonas Bengtson wrote:RAW is RAW. No subsampling. Just pixels from the sensors.

SDI out is 4:2:2



This.

Also, as far as I know, chroma subsampling is totally different from colour space.

Re: 444 OR 422 - color space can shoot BMCC ?

PostPosted: Thu Jun 06, 2013 3:21 pm
by Nick Smith
RAW is 12 bits, which would be the equivalent of 4444 though right? Or does that number have nothing to do with the bit depth?

Re: 444 OR 422 - color space can shoot BMCC ?

PostPosted: Thu Jun 06, 2013 3:42 pm
by Jonas Bengtson
Nick Smith wrote:RAW is 12 bits, which would be the equivalent of 4444 though right? Or does that number have nothing to do with the bit depth?


Bit Depth has nothing to do with chroma subsampling.
They can be related in encoding terms but not in the terms we are discussing right now.

Re: 444 OR 422 - color space can shoot BMCC ?

PostPosted: Thu Jun 06, 2013 3:47 pm
by Tom
Nick Smith wrote:RAW is 12 bits, which would be the equivalent of 4444 though right? Or does that number have nothing to do with the bit depth?



It has nothing to do with it.

4:4:4, 4:2:2, 4:2:0 etc are all descriptions of a form of lossy compression which reduces the resolution of the chroma channels of video. They do not describe colour space nor are they in any way related to bayer patterns or bit depth. It is worth pointing out, that because bayer pattern data needs to be debayered before it can be viewed - the actual quality of the footage is very much dependant on how it is debayered.

The confusion often arises from the fact that people talk about chroma resolution in terms of having a high enough sensor resolution in order to have 1 coloured photosite per channel. This is because a bayer pattern sensor uses photosites which are either red, green or blue. 1 pixel on the raw image originally came from a photosite which was either red green or blue, not one with full RGB data. The RGB colour for each pixel is derived by interpolating the data from the surrounding pixels - how well this is done can make a big difference to the overall image quality. The argument is that because each 1 pixel does not have proper non-interpolated RGB data, the sensor should have a higher resolution than the final video resolution, in order that the final resolution ends up having x1 photosite of each colour per pixel. The BMCC sensor does not have a high enough resolution so that when converting to 1920x1080 HD video, each pixel has x1 red, x1 green and x1 blue photosite behind it. This would suggest that in a similar way to how 4:2:2 has a lower chroma resolution, the bmcc sensor also has a lower chroma resolution. The problem with this is that when compressing to 4:2:2, the effect is uniform and specific. When debayering the raw data, the effect is very variable and dependant on the debayering method. Hence one should not confuse chroma-subsampling with bayer pattern sensors.

The BMCC shoot raw uncompressed video as bayer pattern data. This means that it is true to say you can take that footage and turn it into 4:4:4 compressed footage. The Raw footage is also 16bit data stored in a 12bit log file.

The BMCC also shoots ProRes and Avid DNxHD - which are compressed formats with variations of their compression methods. The Methods used in the BMCC are both 10bit 4:2:2 formats.

Re: 444 OR 422 - color space can shoot BMCC ?

PostPosted: Fri Jun 07, 2013 1:53 am
by sean mclennan
^^^ read...understand...learn.

Great explanation and breakdown Tom!

Re: 444 OR 422 - color space can shoot BMCC ?

PostPosted: Wed Jun 12, 2013 5:15 am
by niulinfeng
Tom wrote:
Nick Smith wrote:RAW is 12 bits, which would be the equivalent of 4444 though right? Or does that number have nothing to do with the bit depth?



The BMCC shoot raw uncompressed video as bayer pattern data. This means that it is true to say you can take that footage and turn it into 4:4:4 compressed footage. The Raw footage is also 16bit data stored in a 12bit log file.



sir, could u please further explain that

Re: 444 OR 422 - color space can shoot BMCC ?

PostPosted: Wed Jun 12, 2013 10:35 am
by Tom
niulinfeng wrote:
Tom wrote:
Nick Smith wrote:RAW is 12 bits, which would be the equivalent of 4444 though right? Or does that number have nothing to do with the bit depth?



The BMCC shoot raw uncompressed video as bayer pattern data. This means that it is true to say you can take that footage and turn it into 4:4:4 compressed footage. The Raw footage is also 16bit data stored in a 12bit log file.



sir, could u please further explain that



Which bit?