Thread: Log to rec709

Page 2 of 3 FirstFirst 123 LastLast
Results 11 to 20 of 21
  1. #11  
    Senior Member rick.lang's Avatar
    Join Date
    Apr 2012
    Location
    Victoria BC Canada
    Posts
    3,956
    Pretty good match. When I tried ACES, colour was a mess. But I didn't know what I was doing so I went back to the real world and deliverables are fine.
    Reply With Quote  
     

  2. #12  
    Senior Member
    Join Date
    Apr 2016
    Posts
    101
    i find ACES / Resolve easier if i use contrast / pivot / exposure / sat only in the first node, much like using Baselight's default "filmstyle' gradeing


    and do not touch lift / gamma / gain or the curves menu untill the image has been normilised using the filmstyle tools
    Reply With Quote  
     

  3. #13  
    Senior Member shijan's Avatar
    Join Date
    Apr 2012
    Location
    Odesa, UA
    Posts
    520
    Quote Originally Posted by david evans View Post
    Great tips! I tried the LUTcalc with these parameters:

    Rec Gamma - BMD Film 4.6k
    Rec. Gamut - Blackmagic Passtrough
    Out Gamma - Amira 709
    Out Gamut - LC709

    Before the LUT, I created the color transform node with the following:

    Both inputs : as timeline
    Output Gamma: Amira709
    Output Gamut: L709

    I'm getting a slightly more pinkish image than I usually do, but it's definitly softer on the highlights then the BMD LUT. Seems to retain more information. Do you know if it's better to use different parameters?

    Thanks
    Need to test it, but not sure if "Rec. Gamut - Passtrough" is equal to BMD film gamut and will work correctly. The LUT needs correct input color space (gamut) to able to convert (compress) colors in smaller color space. Usually "Passtrough" assumes that you don't expect any color space correction in future. Most likely you got a LUT which just assigns (but not converts) smaller LC709 color space to large BMDfilm color space and so you got dimmed desaturated colors. Its probably same as assign sRGB to ProPhotoRGB image in Photoshop.
    It was discussed before that BMD film gamut is kind of exposure dependent and can not be reproduced with common methods. There is also no any open specs about it. Thats the reason why it is not used in LUTcalc.

    With earlier versions i used input - Data, Output - Legal. But i see that they add 2 new presets (LUT Type) for Resolve which are unknown for me and need to be tested and understanded.
    BTW here is last beta version of LUTcalc http://cameramanben.github.io/LUTCalc/testing/ seems they bring back lovely Amira709 and Alexa-X-2 color spaces. And also seems Amira709 and Alexa-X-2 gammas now have slightly different contrast levels.

    In common both have very smooth highlights rolloff:
    Amira709 - filmic colors and more contrasty, Alexa-X-2 - filmic colors and less contrast in shadows.
    LC709 - more saturated colors and more contrast, LC709A - more dimmed colors and too low contrast as for me.

    As to input color spaces for LUT, you can use any wide gamut supported by Color Space Transform node in Resolve. Sony, Panasonic, Rec2020, Arri wide gamut. They just works slightly different when you color correct image. Test them.

    Read LUTcalc manual and tooltips and decide for yourself which is better for which situation.
    As for me ACES is limited in Resolve, it don't allow to use LUTs, Color space transforms and limited in export formats. I don't see any reason to us it untill you in ACES VFX worcflow or some kind of it.
    With Resolve Color Managed LOG workflow you can get 100%same tools (gain as expose, contrast without clipping) and color corerction methods as with ACEScc.
    Last edited by shijan; 01-06-2017 at 07:12 PM.
    Reply With Quote  
     

  4. #14  
    Senior Member shijan's Avatar
    Join Date
    Apr 2012
    Location
    Odesa, UA
    Posts
    520

    Reply With Quote  
     

  5. #15  
    Senior Member shijan's Avatar
    Join Date
    Apr 2012
    Location
    Odesa, UA
    Posts
    520
    Quote Originally Posted by david evans View Post
    Great tips! I tried the LUTcalc with these parameters:

    Rec Gamma - BMD Film 4.6k
    Rec. Gamut - Blackmagic Passtrough
    Out Gamma - Amira 709
    Out Gamut - LC709

    Before the LUT, I created the color transform node with the following:

    Both inputs : as timeline
    Output Gamma: Amira709
    Output Gamut: L709

    I'm getting a slightly more pinkish image than I usually do, but it's definitly softer on the highlights then the BMD LUT. Seems to retain more information. Do you know if it's better to use different parameters?

    Thanks
    You done things wrong

    Try this:
    LUTcalc with these parameters:

    Rec Gamma - BMD Film 4.6k
    Rec. Gamut - REC 2020
    Out Gamma - Amira 709
    Out Gamut - LC709

    In Resolve Before the LUT, create the color transform node with the following:

    Both inputs : BMD Film 4.6k
    Output Gamma: BMD Film 4.6k
    Output Gamut: REC 2020



    Adjust contrast and do color corerctions B E F O R E LUT generated in LUTcalc
    Add Film emulation or any other filmic color LUT A F T E R LUT generated in LUTcalc
    Last edited by shijan; 01-06-2017 at 08:42 PM.
    Reply With Quote  
     

  6. #16  
    Senior Member shijan's Avatar
    Join Date
    Apr 2012
    Location
    Odesa, UA
    Posts
    520
    it seems returned Amira709 color gamut produces strange ripple in highlights in LUTcalc v2.99Beta1. So be aware of it yet.
    Alexa-X-2 seems ok, but with slightly different contrast arrangement now.
    LC709/709A seems the same as in LUTcalc v2.91

    Screen-Shot-2017-01-07-at-3.58.50-AM.jpg
    Reply With Quote  
     

  7. #17  
    Senior Member
    Join Date
    Apr 2016
    Posts
    101
    As for me ACES is limited in Resolve, it don't allow to use LUTs, Color space transforms and limited in export formats. I don't see any reason to us it untill you in ACES VFX worcflow or some kind of it.
    i do agree with Dimitri most of the time, i do have a diffrent view on some of his thoughts...
    - LUT's can be used after the image is normilised, with all the dammage most LUT create, negateing the value of ACES mathmatical transforms, but if suboptimal images are the goal, ACES is the probbaly wrong way to go about it
    - color space transform are also on offer, in both Baselight and Resolve in ACES, and much cleaner than LUT's, i use them in both systems commonly
    - Export formats are just about anything and everything.... that one i do not understand atall...


    With Resolve Color Managed LOG workflow you can get 100%same tools (gain as expose, contrast without clipping) and color corerction methods as with ACEScc.
    That's because RCM (and truelight color spaces as well) is a subset of ACES, so yea, it makes sense tha the same tools are on offer.. as they are the same tools... Baselight is soooooo much more elegant on this file tho, diffrent world really


    more thoughts;

    The advantages of ACES are pretty significant... today i worked on a second pass thougth two and half reels of a feature, DP, director and producers in attendance, will do the other two and half reels tomorrow, roll across with mix and final tweaks on Monday.. i did not once touch hue offset balls, 95% of the grades are nothing but exposure/contrast/pivot/saturation then a second node with L*a*b, then a third node with seperate RGB exposure / contrast / pivot / and log grade tools for highlights and shadow trims

    in maybe 5% of the entire feature i used lift/gamma/gain -or- curves -or- keyers -or- windows/tracks.... 2400 shots, and in maybe 120 shots (but probbaly less) i used the tools that are really designed for display refered workflow... but that all i needed to use, the images looked very good, clean and natural when you don't beat them to death, it takes a light touch

    ACES works well if you use the tools that work well with it... IMHO lift /gamma /gain, hue offsets in regions, and curves are not really optimal choices given the regions they are designed to work in, and they large canvas that ACES represents...

    Speed, clean maths, and flexiable outputs are the advantages i see, with clean maths being the front runner, with raw camera sources skintones are not comprimised by mapping the linear 16 stops of the Alexa into 8 or less, it's noticable when you have linear ArriRAW and log encoded Prorez double recorded as i do...

    The next advantage to me is speed, i get through a first pass miles faster in ACES than in display refered with log encoded sources, but i use the tools that work well in ACES already

    Last advantage is the flexiable and wide rangeing outputs without destructive LUTs and needless trim passes, i switch between 709 and DCiP3 commonly, no issues atall... export XYZ, sRGB, 709 and DCiP3, linearAP0 as needed with relaibity

    Display refered is not a bad choice, RCM is also not a bad choice, needless LUTs are a bad choice, mutiple LUTs stacked on top of each other are a really bad choice..

    Use diplay referd or RCM without LUTs, reverse engineer them into a grade node, (on a dare it took me under 14 min to reverse engineer a Koji LUT to a point that no one could tel what was my grade and what as the Koji LUT when put over a normal "Marcie" test image) or use ACES with tools that work well in ACES...
    Last edited by dermot shane; 01-07-2017 at 01:03 AM.
    Reply With Quote  
     

  8. #18  
    Senior Member shijan's Avatar
    Join Date
    Apr 2012
    Location
    Odesa, UA
    Posts
    520
    a little test here
    added node with Color space transform OFX before LUT with appropriate output color space.
    added node with saturation boosted to 75%
    added node and applied 4 different LUTs generated in LUTcalc with parameters BMDfilm to LC709 gamma, REC2020/Alexa Wide Gamut/Sony s-gamut/Sony s-gamut3cine to LC709 gamut:

    Just want to show how differently can be arranged colors when color correcting same image in different color spaces. (its better to download images and scroll one by one to see difference)

    REC2020 to LC709


    Alexa Wide Gamut to LC709


    Sony s-gamut to LC709


    Sony s-gamut3cine to LC709
    Last edited by shijan; 01-07-2017 at 03:18 AM.
    Reply With Quote  
     

  9. #19  
    Senior Member shijan's Avatar
    Join Date
    Apr 2012
    Location
    Odesa, UA
    Posts
    520
    Quote Originally Posted by dermot shane View Post
    i do agree with Dimitri most of the time, i do have a diffrent view on some of his thoughts...
    - LUT's can be used after the image is normilised, with all the dammage most LUT create, negateing the value of ACES mathmatical transforms, but if suboptimal images are the goal, ACES is the probbaly wrong way to go about it
    - color space transform are also on offer, in both Baselight and Resolve in ACES, and much cleaner than LUT's, i use them in both systems commonly
    - Export formats are just about anything and everything.... that one i do not understand atall...



    That's because RCM (and truelight color spaces as well) is a subset of ACES, so yea, it makes sense tha the same tools are on offer.. as they are the same tools... Baselight is soooooo much more elegant on this file tho, diffrent world really


    more thoughts;

    The advantages of ACES are pretty significant... today i worked on a second pass thougth two and half reels of a feature, DP, director and producers in attendance, will do the other two and half reels tomorrow, roll across with mix and final tweaks on Monday.. i did not once touch hue offset balls, 95% of the grades are nothing but exposure/contrast/pivot/saturation then a second node with L*a*b, then a third node with seperate RGB exposure / contrast / pivot / and log grade tools for highlights and shadow trims

    in maybe 5% of the entire feature i used lift/gamma/gain -or- curves -or- keyers -or- windows/tracks.... 2400 shots, and in maybe 120 shots (but probbaly less) i used the tools that are really designed for display refered workflow... but that all i needed to use, the images looked very good, clean and natural when you don't beat them to death, it takes a light touch

    ACES works well if you use the tools that work well with it... IMHO lift /gamma /gain, hue offsets in regions, and curves are not really optimal choices given the regions they are designed to work in, and they large canvas that ACES represents...

    Speed, clean maths, and flexiable outputs are the advantages i see, with clean maths being the front runner, with raw camera sources skintones are not comprimised by mapping the linear 16 stops of the Alexa into 8 or less, it's noticable when you have linear ArriRAW and log encoded Prorez double recorded as i do...

    The next advantage to me is speed, i get through a first pass miles faster in ACES than in display refered with log encoded sources, but i use the tools that work well in ACES already

    Last advantage is the flexiable and wide rangeing outputs without destructive LUTs and needless trim passes, i switch between 709 and DCiP3 commonly, no issues atall... export XYZ, sRGB, 709 and DCiP3, linearAP0 as needed with relaibity

    Display refered is not a bad choice, RCM is also not a bad choice, needless LUTs are a bad choice, mutiple LUTs stacked on top of each other are a really bad choice..

    Use diplay referd or RCM without LUTs, reverse engineer them into a grade node, (on a dare it took me under 14 min to reverse engineer a Koji LUT to a point that no one could tel what was my grade and what as the Koji LUT when put over a normal "Marcie" test image) or use ACES with tools that work well in ACES...
    http://acescentral.com/t/acescc-vs-acescct/485/9
    just noticed new ACEScct in Resolve. so now we have: ACES linear, ACEScc, ACESproxy now ACES cct. And all of that in 2 color spaces AP0 and AP1. lol, this thing became out of control. its worse than new gamut and log curve with every new camera every year. add here hdr monitors with same mess...
    Reply With Quote  
     

  10. #20  
    Senior Member
    Join Date
    Apr 2016
    Posts
    101
    Quote Originally Posted by shijan View Post
    http://acescentral.com/t/acescc-vs-acescct/485/9
    just noticed new ACEScct in Resolve. so now we have: ACES linear, ACEScc, ACESproxy now ACES cct. And all of that in 2 color spaces AP0 and AP1. lol, this thing became out of control. its worse than new gamut and log curve with every new camera every year. add here hdr monitors with same mess...
    not really, unless one is easly confused -or- wants to be confused.... for eample the color spaces are defined by the ACES subset used

    The simple version..... two forks;
    dalies with a CDL?
    dalies in ACESproxy
    grade in ACEScc
    VFX in ACESlinear

    no dailes with a CDL?
    grade in ACEScct
    VFX in ACESlinear

    now is that that simple enough? And no LUT's needed or wanted, clean maths everywhere with zero LUT dammage, zero pre-compression of dynamic range top to tail = clean images out the far end of the pipe



    Some details for folks who want to look beyond the headlines;

    ACES liner is used by VFX, it's AP0, a very wide space, hard to grade in with traditional controls....i have the been there done that t-shirt on this one... hence AP1...

    AP1 is a bit wider thatn 2020, the controls in gradaeing systems are now more workable, colorists like me are happier....

    ACEScc is AP1, cc stands for "Color Correction" ACESproxy is also AP1, but shadws and highlights are mapped for rec709, it's only use is for on-set gradeing and creating CDL's that translate accuratly into ACEScc in a full gradeing system, so look at ACESproxy as a subset of ACEScc, only for onset use

    ACEScct (the cct stands for Color Correction Toe) is identical to ACScc aside from a soft shadow roll off.. this is for folks who were haveing a hard time with the shadows in ACEScc.. if i'm gradeing a show with no valid CDL to lean on, i'd chose ACEScct everytime as it's faster

    In Resolve you have to chose one option for the entire project, but you can move any shot from one space to another using DCTL's
    In Nucoda and Baselight you can chose by the shot or group, makeing the color space transform and DCTL's in Resolve redundant
    Last edited by dermot shane; 01-08-2017 at 01:45 PM.
    Reply With Quote  
     

Similar Threads

  1. UM46K BMD Film to REC709 LUT?
    By AaronChicago in forum General Discussion
    Replies: 19
    Last Post: 06-29-2016, 06:29 AM
  2. Alexa Rec709 LUT for BMPC4K
    By AlexB in forum Footage / Frames
    Replies: 2
    Last Post: 11-15-2014, 06:11 AM
  3. Replies: 4
    Last Post: 06-19-2014, 06:40 AM
  4. applying rec709 or custom LUT to monitor out?
    By jfilmall in forum General Discussion
    Replies: 5
    Last Post: 11-13-2013, 06:39 PM
Bookmarks
Bookmarks
Posting Permissions
  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •