Re: DXL Millennium and 8K Monstro
What OLPF was inside both cameras Pawel?
From: cml-raw-log-hdr@... <cml-raw-log-hdr@...> on behalf of Pawel Achtel ACS via Cml.News <pawel.achtel=24x7.com.au@...>
Sent: Friday, October 4, 2019 5:09 PM
To: cml-raw-log-hdr@... <cml-raw-log-hdr@...>
Subject: Re: [cml-raw-log-hdr] DXL Millennium and 8K Monstro
Jonathon Sendall wrote: “…is there any jiggery pokery before RAW gets dropped into any NLE or colour grading scenario…”
The test was performed using IPP2 pipeline, REDWideGamutRGB using RCX and outputting REC2020 to monitor.
Both sensors were exposed to all visible monochromatic spectrum projected directly on the sensors using laboratory-grade monochromator, which is much more accurate than using colour charts.
Both cameras reproduced greens and cyans significantly de-saturated compared to REC2020 target. Monstro appears to have slightly deeper, more saturated greens.
Monstro covered REC709 green saturation, whereas DXL2 didn’t. None of the cameras could achieve P3-DCI saturation of greens and fell way short of target REC-2020 green saturation.
Saturation as well as accuracy of blue colours was very good on both cameras.
Colour shifts in DXL2 were significantly larger than Monstro’s, turning red colours into magentas (line of magentas in CIE should be clear in an ideal case).
Below are CIE 1931 chromacity diagrams of all visible monochromatic colours reproduced by both cameras: Monstro and DXL2 respectively.
Figure 1RED Monstro
Figure 2Panavision DXL2
Below, vector scope outputs. In ideal scenario, the scope should resemble a pizza with magenta slice entirely missing (because magenta is not a monochromatic colour). Clearly visible is a shift of red towards magenta as well as desaturated greens observed on CIE chromacity diagrams.
Figure 3RED Monstro
Figure 4Panavision DXL2
Another problem we encountered was with RCX in that it significantly affected the colour vector as well as saturation of monochromatic colours with colour temperature adjustment.
In an ideal world, monochromatic colours should be invariant to colour temperature change, but this doesn’t appear to hold true in RCX (presumably a bug).
Hope it helps.
Pawel Achtel ACS B.Eng.(Hons) M.Sc.
“Sharp to the Edge”
ACHTEL PTY LIMITED, ABN 52 134 895 417
Mobile: 040 747 2747 (overseas: +61 4 0747 2747)
Mail: PO BOX 557, Rockdale, NSW 2216, Australia