Retro

Colours
I started collecting vintage digital cameras in 2020 and soon combined it with my other pandemic hobby, photographing squirrels. I started to notice a pattern that the older cameras produced much nicer colours, especially in the forest during summer (lit by green bounce-light). The first four old cameras that I tested all had CCD sensors, so I had assumed that was the cause. However, after posing my findings on Reddit, several people told me that my initial assumption was incorrect. After taking all of their points on board, I am now working on a second part to this project. Until that is ready, you can read on to see my initial findings below...
New Cameras
To help illustrate what sparked off this project, here are some sample images from my modern mirrorless cameras - the: Sony A7, Sony A9, Sony A1, Sony ZV-E10 (which are all using CMOS sensors). To address the elephant in the room here, Sony cameras do have a reputation for having a bit of a green colour-cast, but that is largely fixable in post (as I understand it) and the newer models have addressed this. This "Summer Squirrel Saturation Situation" (as I'm going to call it), seems unfixable by simply moving some sliders around on the RAW file (at least from what I can tell). I will put the date that each camera was released on each image, below...


Old Cameras
Editing this next image was the main spark for this project. The colour separation of the red squirrel's fur against the green foliage jumped out at me. Shot in the same place as the last image (above) you can see that the red fur is much more achievable here. This was shot on a 2.7 megapixel DSLR (the first purpose built one, actually) from 1999, that's more than a quarter of a century old! These next images are all from CCD sensors. In order, these cameras are: Nikon D1, Fujifilm Finepix S2 Pro & Nikon D200...
Sony A390 [2010]
This next CCD camera is the last one that Sony made. Apart from a 14 megapixel APS-C sensor, it also featured full autofocus in "live-view" mode, with a tilting LCD screen and in-body image stabilization. On top of all that, it also supported some interesting fast Zeiss lenses, like this 85mm f/1.4 Planar. For this next phase of the project, I did a comparison between this camera and the Sony A1.
Sony A390 + Zeiss Planar 85mm f/1.4
I did my best to match each image from the two cameras. The same squirrel, at the same time, in the same lighting, position and pose. Unfortunately I needed to use different lenses, but both were high quality Sony 85mm prime lenses, so hopefully that helps to keep some level of consistency. Both images were taken in RAW format and manually edited in Adobe Camera RAW to provide the best combination of White Balance Temperature and Tint.
Red squirrels were looking far more red on these older cameras. This was just what I was getting on the Nikon D1. It has been a conscious thought that something was lacking on the modern camera's colours. Previously, they felt washed out, like looking through a brown haze. The older cameras are providing more of that separation, a more saturated red, without looking contaminated by green. If I try simply boosting saturation of a CMOS image, their fur goes a rather toxic yellowy orange. Images from the older camera's might be over-saturated at times, but here they just make everything look better.
OK, One More
Both of these images were taken from the same position. The A1 images will be cropped down to match the APS-C area. I managed to get the poses even closer this time, and each image was only a few seconds apart. Both images used exactly the same settings, so this will also be an interesting comparison for exposure. The white balance for these will be calibrated to the same area.
Assumptions & Next Steps
I know that, technically, whether CCD or CMOS, digital camera's pixels don't capture colour information (only greyscale). The colour of the filter array (CFA) in front of each pixel determines the wavelengths of light getting to the photodiode (although how well the photons are turned to electrons for each wavelength could still be different). I had not considered that the existence of the CFA disqualifies sensor types from being responsible for the balance of colour. However, many people do, and I can understand their assumption. It's true that early CMOS cameras also have pretty colours for general photography (as subjective as that is). Whether this is also the case for this my "Summer Squirrel Saturation Situation" however, we will see...
Thus, the next phase in my testing will be to add samples from a Canon D30 (the oldest Canon CMOS from 2000), as well as early full frame CMOS cameras from Canon, Nikon, and Sony (2005-2008). Then I will compare those to CCD and newer CMOS cameras. I will also use the same lens (where possible), colour calibrate all new samples, and experiment more with colour profiles and processing.

It's true that fundamental changes in digital sensor colour treatment happened at a similar time that the industry switched from CCD to CMOS? Whether that was the camera's internal processing (colour science) or the CFA will probably remain a mystery for me, but after the next batch of testing I hope to find out whether CCD is responsible for these nicer squirrel colours, or not.
I will update this article with phase two of the testing, once all the samples, experimentation and processing is done (after the summer).