Bayer filter or Math issue?

PJ,
This is complex.

Some of the cameras I've looked at in the past have strong overlaps in the blue to green and green to red dichroic filters, and the coating passes some yellow wavelengths into the green sensors.
Part of this is to be backward compatible with the older National and International definitions of color in video systems, and part of it is due to legacy modulation where various color domains had "weights" or fractional multipliers applied to Red, Green, and Blue subcarriers. One of the few things I like about ATSC(Digital) in the US, vs NTSC (Analog Legacy) in the US is that flesh tones are actually related to human flesh, and golf courses actually have green grass.

NTSC (World Designations N and M, etc. ) had a 3.58 Mhz subcarrier that was phase controlled and QAM modulated to carry color in a modified classical monochrome video carrier of 7 Mhz bandwidth, and that was sort of SSB, the vestigial sideband was filtered out at the transmitter to reduce bandwidth while conveying nearly all the modulation. This was done to make the color system compatible with older black and white video so millions of people did not have to replace their TV sets. This was done at the loss of some serious amounts of monochrome luminance data. When color was implemented, the deflection format was changed, digital counters were used for generating sync and color subcarrier, and video resolution / line count was degraded a bit vs the pure original monochrome standard. Even the monochrome standard used a mix of blue and yellow phosphors to achieve white, to make an acceptable picture tone. Its a "weird" white if you ever see one of the older monitors, has a weird blue / grey tint to it as the two phosphors have different decay times.

PAL (B,D,G,I. etc) mixed the three colors into two subcarrier modulations and one Luma carrier. PAL had superior color where NTSC systems came with chroma, tint, and fleshtone controls, which, until its last days, led to constant family and Bar fights over how to adjust the TV,,,,

This is why, until DTV made its mark and has been more perfected, RGBY Laser show systems made a color gamut that was unseen before by most viewers. PCAOMs only made this gamut ever larger.

I don't have a great answer for you, but everything Video is an engineering compromise. I suspect with BAYER cameras its a bit of both of what you express. OLED and Color Plasma sets are still far better appearing then most LCD based systems. OLED is stunning, nearly the gamut of a really good RGBY laser. You can spend a lot of time staring at CIE charts trying to figure this out, then factor in International Standard Illuminants etc.

I had a long debate with the engineers at my former medical employer with CRT vs LCD systems. The new stuff shipped with LED, resulting in a temporary deep market for third party medical Grey Scale, Monochrome, VGA or better CRT systems. Radiologists overwhelmingly preferred monochrome CRTs with deep luminance , as the image sensors in a good X-ray, or CT machine were 12 , 14, or more bits. If you want to go off into a tangent, look up "Houndsfeld Number", which applies a sliding scale to the Luminance data to compress it into the CRT or LCD's capability. Some MAC users and a few PC users could not wait to snap up those expensive "Paper White" high performance displays. I'd love to have one for Engineering PDFs.

This drove me nuts when I tried to store three monochrome images in a NTSC recorder for 3D video at work. The color weighing and summing caused serious crosstalk. I returned to storing two images in monochrome alternating fields, or in alternating lines.

Video is, and always will be, a compromise.

This is start, I know you are capable of digging deeper based on this:

https://www.techbriefs.com/component...ves-of-cameras

Steve