Alan, I like your scans. I used to own a V550. It cost me new the equivalent of 200$. I thought it had an unbeatable price/performance ratio especially with medium format negatives, provided they were well exposed and developed correctly.
But then, if you think about it, actually even the line sensor in these cheap flatbeds is, in a way, a better scanning device than the interpolated sensors in mirrorless/DSLR cameras.
Just to clear up a few things. In essence, the raw output of Bayer-filter cameras consists of a so-called Bayer pattern image: an arrangement of colour filters on a square grid of photosensors. In the Bayer arrangement, this filter consists of a matrix of repeating 2x2 pixel patterns, one coding for red, one for blue, and two for green. Importantly, each pixel is filtered to record only one of three colors:
Bayer Filter
[source
https://en.wikipedia.org/wiki/Bayer_filter]
So the key thing here is that each pixel of the sensor is behind a colour filter and the output is an array of pixel values, each indicating a raw intensity for one of Red, Green or Blue. This arrangement needs an algorithm to estimate for each pixel the colour levels for all colour components, rather than a single component.
This is called `demosaicing'. There are different implementations of this - think of it as a type of signal interpolation. Now compared to the initial, raw intensity images, the reconstructed image is typically accurate in uniform-coloured areas, but will have a loss of resolution (detail and sharpness) and edge artefacts. This is a big topic and one that is valid for the output of X-trans sensors too (perhaps even more valid as the output of an X-trans array has historically been found to be more difficult to demosaicise).
But to go back to scanning, most dedicated film scanners do not rely on Bayer (or equivalent) pattern matrices and the raw output they produce does not require demosaicing. The so-called `line CCD sensors' in a scanner are, at a very raw level, better than any camera sensor because they do not interpolate and because they use only a single line using the best part of a sharp dedicated lens so there is little or no optical distortion or other lens flaws added.
One consequence of the lack of a Bayer array+demosaicing is that when a scanner like the Coolscan 8000/9000ED or the Pacific Image is scanning 90mp, those are 90mp of full color data. Digital camera colour data is only 1/4 of the stated resolution due to the above. So even, say, a Fujifilm GFX 100 (a 102mp sensor, 10K$ camera) is only getting 25mp of full color data (and another 25mp of extra green [luminosity] data) from its 100mp of photosites. There are workarounds to limit that: eg pixel shift, but you are still left with the limitations of digital camera colour, lens flaws, having go through the hassle of stitching when scanning 120, plus any other issues inherent with the specific home-made scanning setup used (vibrations of the repro stand? imprecise sensor/film alignment? poor quality/evenness of the retro-illumination; and much more). Orange mask removal is another story and so is the lack of IR (infra-red channel) for dust removal in home made DSLR scanning rigs, but you know that already.
In any case - apologies to OP for further contributing to derailing this thread, which IMU wasn't about DSLR scanning at all.
One of the reasons to macro and stitch, or sensor shift or do manual sensor jitter and stack-stitch is exactly to take most or all of the effect of demosaicing out.
Some scanners use line CCDs. CCDs transport the whole image, in this case a line, over the sensor via an analog bucket brigade. A technique which on the physically long lines and old processes, like these low production run silicon products are made on, is not exactly harmless to signal quality.
The stepper motor moving the line along the other axis of the image, is the sole arbiter of resolution potential along usually, the long dimension of the photo.
The two above “gotchas” means lots of added noise and and false grain from mechanical imprecision and analog sampling lines.
The RGB strip lines might seem an advantage at first, but is actually not in the way it is often implemented. Often the whole width of the line is treated as one pixel that is stepped over the image in whole increments.
That actually means worse colour precision than with Bayer arrays.
If you have an IR filtered flash there is noting stopping you from doing dust removal with camera scanning too.
It’s just often not necessary because the materials used in camera scanning doesn’t attract as much dust (less static).
Amazing info.
It was not my intention to turn this in to a scanner vs digital camera.
My intention was to praise the small companies that still provide us with products and to make the case that the quality of the product is very good and we should support them.
If that is really “amazing info” to you, a 101 in how sensors work, you might want to question the grounds and knowledge base for the rest of your steadfast reasoning.
Good luck achieving even illumination and excellent flat field performance at 1:1 magnification with newly made, low cost equipment and materials.
It is considerably easier to do that at the magnifications used in enlargers and projectors.
Don't get me wrong, I'm impressed with the efforts that people have devoted to using digital cameras to digitize film. I just know that when I've attempted to do it, the equipment I already have is ill suited for the procedure, and the investment necessary to end up with something that would be both good and convenient is in the same range as a new scanner, and would take a lot more space to use.
If you can come up with a decent quality "box" that handles the illumination and film handling and optical requirements and only requires me to mount my wife's micro 4/3 camera to it, then we can talk.
Even illumination is not that hard. Especially if you have plenty of light to waste in diffusion, as with a flash.
You are essentially doing 1:1 or larger macro, so not anything a projector or enlarger is not also doing.
Contact “printing” directly on a sensor with collimated/condensed light is an interesting unexplored option.
Of course a monochrome sensor would be preferable. And so would a liquid optical interface layer.
Not sure how thin a protective layer on sensors is possible.
When I referred to a "box", I wasn't being figurative. I meant a reasonably small box - the need for a copy stand (and the space for it) is part of my problem with re-purposing a camera and lens designed for general photography for this purpose.
And as much as I think my wife's micro 4/3 digital camera is great for what it is intended for, it isn't well suited to the equipment I already have - a 35mm slide copier originally designed for 35mm film - if I want to re-purpose it for film digitization.
I am perfectly capable of taking the steps necessary to convert products designed for other purposes to the process of film digitization. I have sufficient knowledge and experience to obtain excellent results from such a setup, if that were practical - but it really is not. The equipment would need too much space, would cost me too much money, much of it spent for functionality I don't use (I do relatively little close-up work) and sit unused for significant periods amount of time. And I would have to dis-assemble it after each use, store it, and then re-assemble it before each use.
The same concerns apply to my darkroom use. The difference being that I already own high quality darkroom equipment, including lenses suited to that use that are much higher quality than I could afford in a camera scanning setup. And yes I have investigated ways of using enlarger lenses in that application.
For other people, the functional compromises involved in re-purposing equipment designed for other uses would not be as significant. I encourage them to do so, if that is their wish.
And you mean something like the Filmtoaster?
https://www.filmtoaster.photography/
Problem with most people designing stuff like this, is that they implicitly also want to get rich quick, and have no idea about industrial production design or market research.
The film toaster is a prime example of that. Designing it with all these extras and options and even a toilet roll holder on the side is just at the same time glaringly arrogant and naive.
We need a simple box with a universal body mount, an enlarger lens screw-mount inside and with a backlight and a piece of liftable AN glass on top of the light.
And then a way to slide either the mounted camera or the mounted film around in a rigid plane.
Shouldn’t cost much to make or be too hard to get funded, if you look at the much more complex but much narrowly target grouped stuff that is crowdfunded several times over on a regular basis.
That camera and that lens together would cost me somewhere around $2800 USD to import if I bought them new.
Outside of film digitization, I would get little benefit from them compared to borrowing my wife's camera.
I would like to see what could be done by a savvy manufacturer incorporating a decent camera sensor in a digitizing box, or a digitizing box that would permit attaching my wife's camera.
While we wait for someone to do that, it doesn’t need to be that expensive (or hard) to do your own.
In fact a smaller higher pitch sensor is preferable. Exposure time and sensitivity is not the limiting factor it is with live stills.
A used M43, a macro or reversed enlarger lens, an old enlarger bed and column with the appropriate tripod mount and a backlight is really all you need for more than good enough results.
Far better than anything scanned on current consumer scanners.