|
|
Thanks a lot Evangelos for sharing this :)
Yes indeed!
I don't think it's perceptually linear either though... It certainly differs from the L part of the Lab equation. In the design of the TV system, the monitor decode curve isn't the same as the encode curve anyway, imparting contrast to the image which is a non-linear effect. That the monitor decode curve (originally the CRT response curve) is close to being perceptually linear is a useful bit of good luck that helped TV work and was not part of the initial design. Poynton has some good stuff about this in his book.
If we can keep the word linear for things that are a straight line, that helps a lot.
well ,
i don't totally agree ...
here seems that :
-10bit DPX is the only available "container"
-alos seems that when talking of 709 gamma space , 709 matrix is the only option to associated with it (meaning that the image can be processed as 709 gamma (like redgamma 2) and use Camera RGB as color matrix )
so for example the image can be graded as regamma 2 (gamma space) and camera rgb (color matrix) as DPX 16bit (or obviously Redcode) for DCP and than use the 709 conversion lut (this apply for digital only delivery )
reading the article seems that without a log + print lut workflow the image will be inferior ...
as colorist i actually think that use a film print Lut when not needed , is really not ideal , as the color palette of film is quite different than the color palette of a Digital cinema camera and sometimes create quite weird color shifts ....
my 2 cent
Hi Graeme :)
This confusion comes from that there is no define standard on rec709 on camera and TV like deanan said. I understand what you mean that the encode curve is different from the decode curve, which is why it makes it not perceptually linear, well ideally it isnt. But i know people like Michael Chenery from Cinespace refers rec709 as perceptually linear because the grayscale gradient effects seem like linear or close, eventho it isnt accurate 100% technically.
But i really dont know what else that i could refer rec709 to if this is the case. It's not log and not really perceptually linear nor true linear. I tend to lean over that it is perceptually linear category.
But for argument sake, we could call rec709 as Video Gamma Space, hence removes confusion for those who still try to differentiate between the two.
Rivai
Yes, we generally refer to REC709, sRGB etc. are gamma encoded or video gamma. Terminology-wise we leave linear for linear light encodings and log for encodings based on a log function.
Yes indeed REC709 is very confusing - and it only ever specified the display, not how the image to be displayed should be derived.
Graeme
Because I want to keep it simple for regular people to understand, if I remove the word "Linear" and insted say REC709 would be sufficient?
(I also know that Mike C. calls it perceptually Linear... because we see it as linear, when combining camera curve and display curve)
Gabriel there are very few people that have the ability to store 16bit images in HD or 2K with video storage... you know better the top Video is HDcam-SR 444... which is 10bits... and very few are 12bits... and even much fewer are 16bits...
When people says "I will color grade your movie in REC709" very very rarely mean 16bits DPX... also very few software allow that...
But it's not only the container format its also what you see with what you render is the SAME... in Log they are not... you see through a LUT and you render without the LUT... so what goes to file is much bigger than what you see...
Think about that... its paramount difference...
Yup, just REC709 is perfectly correct and adequate.
Sure it's linear when you see it - but that's because you've got a non linear file on a non linear display (or a linear display with a non-linear curve to compensate!) - although the end result is linear, each of the stages is not :-)
Graeme
BTW Evangelos....quick question. What method/tools you use to handle camera that doesnt give you any RAW access like RED/Arri.
Let's say you shoot with Sony F3 which has a larger color space than rec709. It is a fact that the picture is gamma encoded. How do you convert those image to standard density cineon log files if they dont give you the specification ? Let's skip its S-Log discussion for now.
Is it possible that you could transform the image to true linear and apply lin2log function for us to be able to get into log space for DI ?
PS. I used Sony F3 as an example, but we could use any other type of camera that has color space bigger than rec709
Thanks
Rivai
| « Previous Thread | Next Thread » |
| Tags for this Thread |