Colour space for RED log files

Home Page forums Applications Nuke Colour space for RED log files

Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
    Posts
  • #202891
    Mel
    Participant

    I have a DPX sequence that was transfered from the r3d’s as log then graded in baselight & output Log.
    I want to know what colourspace i read these files in as. REDLog looks on the money but REC709 asnd sRGB look good too. I am looking at this dpx sequence in Flame which is right next to the Nuke to get the match right.
    (most comps done in the flame so i know that whatever the flame is outputting is what is approved)
    P.s Must keep the metadata in tact.

    Thanks
    Mel

    #217909
    Keefa Chan
    Participant

    The dpx sequence has come from baselight. We are not using the R3d;s, as there has been a slight grade applied, hence it coming from Baselight. The dpxs where output from baselight as log. I am reading them into Nuke. I need to output exactly what i am given. It is a potential film out.

    Cheers

    #217907
    guillem ramisa
    Participant

    @Mel 32507 wrote:

    The dpx sequence has come from baselight. We are not using the R3d;s, as there has been a slight grade applied, hence it coming from Baselight. The dpxs where output from baselight as log. I am reading them into Nuke. I need to output exactly what i am given. It is a potential film out.

    Cheers

    If you are unsure what colorspace the DPX has you can turn off any color-conversion in Nuke: choose linear or raw data in the read node, disable the sRGB curve in the viewer (choose none) and make sure to choose linear or raw data in the write node. This is of course not ideal if you are mixing your source clip with other stuff (like CG) so it depends a bit what you want to do. For simple retouching it might be fine.
    A filmscanned log DPX would normaly be interpreted as Cineon (in Nuke). But you said you had matching(?) results with redlog.
    To get it perfectly right I guess you need to know a bit more how the conversion from Red to DPX was done pre grading and exactly how the output from the baselight works. (example: red raw -> to dpx with redlog? -> subtle grade in baselight -> dpx with xxx colorspace conversion). To me it sounds like the files could be converted to redlog dpx files and that they got a slight grade and then just epxported (still redlog)? But as I said you probably need to dig a bit deeper to know exactly what type of gamma your files have.
    But just as long as you write out your DPX files the same way you bring them in (same colorspace), you are safe.
    It’s just when you start mixing stuff with material with other colorspaces (linear files or srgb files etc) that the internal conversion of your dpx (from whatever colorspace you have) to linear is getting important.

    Cheers,
    K

    #217911
    Keefa Chan
    Participant

    THanks, yeah the r3ds were debayered PD log 985, RED space with the exposure @ -1. It was tHen graded slightly in baselight & output Log as a dpx sequence.
    When i view the dpx sequence in Nuke as RAW with no viewer lut, it looks the same as in the flame, ( with the flames colourspace: video. which i never understood but thats another topic)
    Also, i output a frame in the colourspace REDLog & one in Cineon and imported into the Flame. Cineon was a deadmatch, but the REDlog space was the one that matched the flame when viewing on the monitor in Nuke.
    Cool, so i get that reading in & out RAW would keep the colourspace. So why would REDLog look correct on the monitor (with sRGB viewer) but be the wrong colourspace when checking in FLame?
    My aim is to output Log since thats what baselight gave me, and to be able to view this file correctly in nuke.

    Thanks heaps for your help , i look forward to your reply

    #217910
    Keefa Chan
    Participant

    @Mel 32511 wrote:

    THanks, yeah the r3ds were debayered PD log 985, RED space with the exposure @ -1. It was tHen graded slightly in baselight & output Log as a dpx sequence.
    When i view the dpx sequence in Nuke as RAW with no viewer lut, it looks the same as in the flame, ( with the flames colourspace: video. which i never understood but thats another topic)
    Also, i output a frame in the colourspace REDLog & one in Cineon and imported into the Flame. Cineon was a deadmatch, but the REDlog space was the one that matched the flame when viewing on the monitor in Nuke.
    Cool, so i get that reading in & out RAW would keep the colourspace. So why would REDLog look correct on the monitor (with sRGB viewer) but be the wrong colourspace when checking in FLame?
    My aim is to output Log since thats what baselight gave me, and to be able to view this file correctly in nuke.

    Thanks heaps for your help , i look forward to your reply

    Also, is it just that im not viewing the log files correctly in NUke? DO i need to look through a log to Lin viewer or something?

    regards,
    Melissa

    #217908
    guillem ramisa
    Participant

    Maybe you already know this but it is kind of important to understand how Nuke works with color spaces (so my apologies if this is too “granular” for you). What Nuke does is to try and convert all footage to linear space upon import. In the viewer it applies a lut (default is srgb) to make things look “normal” and then when you write your files it applies the appropriate conversion again.
    You can disable this behavior if you like (described in my earlier post) but it will not be correct if you mix material with different colorspaces.
    If you read in raw and disable the viewing lut and write out raw, your dpx will not change and it will have the same colorspace as it had. This will also mean that when you are working internally in Nuke your file and comps will be in log space.
    Another question, what is your monitor setup for Nuke? Do you have a proper calibrated grade monitor hooked up? Without that it would be a bit difficult know or say if your settings would match anyways.
    Also actually I’m not sure what type of log curve your DPX’s have after all it’s gone through in the baselight system. First converted to pdlog you say and the graded and then output to dpx? Did you grade with a filmlut or was the monitor rec709 for TVC or how was your baselight set up? But that might not be so important anyhow, depends a bit what you are going to do in Nuke. The more color critical your needs are the more important I guess.
    I didn’t really follow how you set up your read/viewing lut/write nodes in Nuke when comparing different log formats (like cineon vs REDLog)? What is important though is that you use the same settings in the read and write nodes to make sure that you don’t change the DPX’s original colorspace.
    So if you get a good match on your monitor using REDLog in the read node then use that. Just make sure to use REDLog when writing out to apply the same color conversion as you did when importing. If using REDLog is a 100% correct conversion to linear for working inside of Nuke I cannot say for sure. Probably not after all that DPX has gone through? But maybe that is not as important. Depends a bit on what you are going to do as mentioned before.
    Sorry if I’m boring you to death. Have to stop babbling now 🙂

    K

    #286109
    Charlie Flint
    Participant

    How incomprehensible it can be when people who know a lot about a subject are talking about it. I, for example, have never really understood the subject, but I have always been interested to understand what you are talking about. Can you use anonymous file sharing and send me a link to a video tutorial on how to properly communicate on such an intelligent topic on the Internet?

Viewing 7 posts - 1 through 7 (of 7 total)
  • You must be logged in to reply to this topic.
Copy link
Powered by Social Snap