GAM and NAM missing in AMF implementation (DaVinci Resolve 19.0.1 BUILD 6)

Dear everyone involved regarding the AMF development,

In DaVinci Resolve 19.0.1 BUILD 6 one may export a sequence as “Source Master”, “Dailies Request” or “VFX Request”.

Please, the post production company I work for needs two more AMF export options: “Graded Archival Master (GAM)” and “Non-Graded Archival Master (NAM)”.

Regarding both the GAM and the NAM, the human readability and understanding of the AMFs are important. E.g., was the Reference Gamut Compression “always on”, and which RRT, ODT or RRTODT was viewed, while grading the important Hero Grade?

Without knowing much about the AMF’s structure, and being confused about the true meaning of “true” or “false”, I attach two examples, including an informative description for human eyes.

  • I am sure you and/or Blackmagic Design can make corrections and improvements?

The GAM, draft 1:

    <aces:amfInfo>

            <aces:description>The project, Timeline 1, Graded Archival Master (GAM), Exported by A Colour Grading Tool</aces:description>

            <aces:dateTime>
                    <aces:creationDateTime>2024-10-03T10:52:29Z</aces:creationDateTime>
                    <aces:modificationDateTime>2024-10-03T10:52:29Z</aces:modificationDateTime>
            </aces:dateTime>

            <aces:uuid>urn:uuid:xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx</aces:uuid>

    </aces:amfInfo>

    <aces:pipeline>

            <aces:pipelineInfo>

                    <Description>The images in the accompanying SMPTE ST 2065-4 based OpenEXR files are encoded according to the Academy Color Encoding Specification, SMPTE ST 2065-1. Within this Graded Archival Master's images, the ACES Reference Gamut Compression and the ACES-based scene-referred colour grade are already baked-in. The "outputTransform" must be added to accurately display the intended Hero Grade.</Description>

                    <aces:dateTime>
                            <aces:creationDateTime>2024-10-03T10:52:29Z</aces:creationDateTime>
                            <aces:modificationDateTime>2024-10-03T10:52:29Z</aces:modificationDateTime>
                    </aces:dateTime>

                    <aces:systemVersion>
                            <aces:majorVersion>1</aces:majorVersion>
                            <aces:minorVersion>3</aces:minorVersion>
                            <aces:patchVersion>0</aces:patchVersion>
                    </aces:systemVersion>

                    <aces:author>
                            <aces:name>A Colour Grading Tool v99.9.9</aces:name>
                    </aces:author>

            </aces:pipelineInfo>

            <aces:lookTransform applied="false">
                    <aces:transformId>urn:ampas:aces:transformId:v1.5:LMT.Academy.ReferenceGamutCompress.a1.v1.0</aces:transformId>
            </aces:lookTransform>

            <aces:workingLocation/>

            <aces:outputTransform applied="true">
                    <aces:transformId>urn:ampas:aces:transformId:v1.5:RRTODT.Academy.P3D65_1000nits_15nits_ST2084.a1.1.0</aces:transformId>
            </aces:outputTransform>

    </aces:pipeline>

The NAM, draft 1:

    <aces:amfInfo>

            <aces:description>The project, Timeline 1, Non-Graded Archival Master (NAM), Exported by A Colour Grading Tool</aces:description>

            <aces:dateTime>
                    <aces:creationDateTime>2024-10-03T10:52:29Z</aces:creationDateTime>
                    <aces:modificationDateTime>2024-10-03T10:52:29Z</aces:modificationDateTime>
            </aces:dateTime>

            <aces:uuid>urn:uuid:xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx</aces:uuid>

    </aces:amfInfo>

    <aces:pipeline>

            <aces:pipelineInfo>

                    <Description>The images in the accompanying SMPTE ST 2065-4 based OpenEXR files are encoded according to the Academy Color Encoding Specification, SMPTE ST 2065-1. Within this Non-Graded Archival Master's images, nor the ACES Reference Gamut Compression (RGC), nor the ACES-based scene-referred colour grade, are baked-in. The RGC "lookTransform" (informative) was used to bring AP0-based colours inside the AP1 gamut, and the "outputTransform" (informative) was used for display, when creating the Hero Grade.</Description>

                    <aces:dateTime>
                            <aces:creationDateTime>2024-10-03T10:52:29Z</aces:creationDateTime>
                            <aces:modificationDateTime>2024-10-03T10:52:29Z</aces:modificationDateTime>
                    </aces:dateTime>

                    <aces:systemVersion>
                            <aces:majorVersion>1</aces:majorVersion>
                            <aces:minorVersion>3</aces:minorVersion>
                            <aces:patchVersion>0</aces:patchVersion>
                    </aces:systemVersion>

                    <aces:author>
                            <aces:name>A Colour Grading Tool v99.9.9</aces:name>
                    </aces:author>

            </aces:pipelineInfo>

            <aces:lookTransform applied="false">
                    <aces:transformId>urn:ampas:aces:transformId:v1.5:LMT.Academy.ReferenceGamutCompress.a1.v1.0</aces:transformId>
            </aces:lookTransform>

            <aces:workingLocation/>

            <aces:outputTransform applied="false">
                    <aces:transformId>urn:ampas:aces:transformId:v1.5:RRTODT.Academy.P3D65_1000nits_15nits_ST2084.a1.1.0</aces:transformId>
            </aces:outputTransform>

    </aces:pipeline>

Additional comments/questions regarding the current AMF-implementations in DaVinci Resolve 19.0.1 BUILD 6:

  • In the “Dailies Request” and “VFX Request” AMF-exports, the “inputTransform” is used. When there is no IDT defined by AMPAS, as an example the “CinemaDNG” Raw Profile, could DaVinci Resolve in the future please include an Resolve-unique informative description only understood by humans, but not machines? (In order for a human to read and manually choose the appropriate Raw setting, correct Raw-SDK-version etc).

  • In the “Dailies Request” and “VFX Request” AMF-exports, a separate CLF is created. The CLF contains a CDL or a 3D-LUT to transform the ACEScct colour volume. Why is such a 3D-LUT only affecting the limited 0-1 float range (with 65 points precision)? I.e., why not the complete ACEScct-range (with 129, 257 or 513 points precision and mandatory tetrahedral, not trilinear, interpolation) - in order for all users to be able to, always and fully, trust the CLF as ACES-safe without any risk regarding quality-issues?

Please, if you readers happen to know the appropriate colleague at Blackmagic Design, do not hesitate to let him/her know about the need for GAM and NAM AMF-exports!

In advance, thank you all for your help.

Kindly, Lars
Lund, Sweden. October 8, 2024.