Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

is there an option to record 16bit redcode instead of 12 bit, in scarlet?

T

Togan Gokbakar

Active member
Joined
Dec 7, 2011
Messages
42
Reaction score
0
Points
0
Hi I am using my scarlet happy ever after. But i wonder if there is an option to record 16bit redcode instead of 12 bit, in the camera menus?

Because with simple math i figured out that scarlet shoots 12 bit instead of 16bit as announced.

1 frame of 6:1 red code footage is 2.2 Megabytes.

2.2 MB is 18,454,937.6 bits

Compression ratio is 6:1 so uncompressed data is 18,454,937.6 times 6 , which is 110729626 bits.
And then when we divide this data to pixel number we will find the bits per pixel which is the color data.
4K red scarlet frame is 4096 x 2160 which is 8847360 pixels.
So lets divide the data bits to pixels: 110729626 / 8847360 which is 12,5155556

So 12,5 bits per pixel. 0.5 is probably the metadata.
So as we can see my footage is originally captured as 12 bit. Which makes scarlet a 12bit camera not a 16bit camera.

Is this a default thing or is there an option to change it that i don't know?

Thanks.
 
From what i've read here, you can't assume that every frame has the same size. It depends on the content you shoot. Compression isn't just a hard 6:1-dropping of data, there is way more math behind it.
 
Data rate is aprox. the same, one frame could be 2.2 mb one 2.1 one 2.3 mb, but not 5 or 6. So it is same most of the time. And when it comes to compression , if it is 6:1 it is 6 to 1, no more math actually.
 
If you go on with the math, 16bit means 16 times more data than 12bit, which seems way too much to record.
I'm happy with the quality regardless of the bits, anyway.
 
looks like red is not really being clear on this , but the most i understood in this matter is that internal prices is 16 bit but it record 12 bit ...

many people asked same question and RED never (or at least i did not catches that ) properly answered stating that it record 16 bit ...

g
 
While you figure out your need-of-16-bit dilemma, the film industry will continue their scan-through-print work with only 10 bits.
 
Yes but it is 10 bit per channel , so it is 3x10 bit 30 bit ( in dpx standard of kodak), and it is 30 bit logarithmic sampling not linear. Red has only 12 bit linear per pixel not channel. if you want to get the full color data with a bayer pattern you need 4k x 3 resolution which our cameras dont have. So i love my scarlet , great camera for the price , but it is nor true 4k nor 16bit
 
Its a master plot!

Obviously the folks at RED are holding your other 4 bit hostage until Xmas.

So, you gonna be naughty or nice?
 
4bit means 16 times more data which is not less. As i said i love my scarlet. Discussing camera specs are not a bad thing. Dont be so fanatic to support red. Evaluate and critisize everything. At the end red is a company for us, like all other companies.
 
You know you can't get 65 thousand levels ( 16 bits ) out of a tiny 5.4 micron pixel, don't you ?
Just not possible.

Educate yourself: look up 'full well capacity'

beyond that, Red smashes out the darkest part ( lower bits ) to make it more suitable for compression. Look at the darkest of the darks, see the lack of stochastic noise in there.


All said, just use the camera an shoot. It looks good for now.
 
Last edited:
i agree with you we dont need 16 bit for video, but they market the camera like that, thats the problem. by the way leica and hasselblad have 16 bit linear recording.
 
There is nothing wrong kemalettin, you can do your own calculation.
 
did my own math on epic and found 14bit..weird isnt it? (on 4:1 ratio 5K,and another test with 12:1 ratio result came up 10) there must be different test to find out raw bit.not like this imho
btw R3D. and Redcode is not real uncompressed i guess.
 
REDCODE is based on wavelet compression, so yes, it's lossy. This makes the math a bit trickier than just 6:1, and then there is also the bayer-pattern-thing to keep in mind. I also don't understand the RC-options we have as a way to primary set filesizes, there is more priority in controling the quality. Size can differ from frame to frame, it depends on content. That alone gives a good hint that there is more than just a hard 6:1-reduction.

So if your calculations say something then its at best a average amount of bits per pixel. But i guess we are far away from being accurate. ;)
 
You must log in or register to reply here.
Back
Top