hmmmmmmm

getting really weird with the editing here
Q+03, raws nondemosaiced, channel mixed to hell, contrast blasted, 1-bit dithered at half size

just put a piece of paper on top of the scanner and boom it fixes the light angle problem. still gotta remove the lights from the scanning head though it washes out the image .. and maybe gotta look into writing custom scanner firmware to actually control exposure cuz once those clouds went away i was just getting pure white captures (oh wait i could just make an aperture for the lens. lollll). this shot turned out nice though

not the best finder and not lined up at all cuz i couldnt build much of a good contraption. but hey, "medium format" shot

continued experiments in editing

i could probably spend hours trying to get the texture to look right, but this is okay.

reverse tone work ?

playing around with this one

raw from the Q, not demosaiced, channel mixed, tone curved, and then b&w converted in rawtherapee. editing a non-demosaiced image is uhhh annoying. the preview is no help, even the 1:1 zoom is not exactly accurate either.
then, export to darktable, blur tool used on 4px radius, local contrast tool blasted because the image was too flat, trial and error exporting until result is acceptable.

is the end result any different from the demosaiced image? uhhh it's nearly impossible to compare, the second image is all the exact same edits, just with AMAZE demosaic in rawtherapee and the blur turned off in darktable, and it's a completely different result so no 1 to 1 comparison can be made. if anything at all it illustrates how distinctly different these 2 editing pipelines are despite only swapping one module for another.

for fun, same process just without the b&w conversion in rt and with white balance in dt in images 3 and 4

there are a lot of subtle things i tried out editing this one. might be on to something...

pulled the fresnel lens out of the overhead projector and placed it on top of the scan bed. uh. i dont really know what i'm doing here. but i think??? that the fresnel might "correct" the light heading to the sensor? to maybe yield a more consistent exposure over the whole frame? shot one is with the fresnel facing in, second is with it facing out. in and out here relative to where the light comes from in the overhead projector since im attempting to reverse its optics for capturing images? uhhh im fairly certain the fresnel needs to be a set distance away from the scanner to actually "work" but its certainly Doing Something here. oh and all the crap on the photo is from the nasty plastic screen on the scanner itself, i'll remove it once i'm more confident with what i'm doing here.

tried to make an aperture plate for my 03 lens, put it behind the back element, turns out the actual existing aperture is in between elements, and so: circle!

PINHOOOOOOOOOOOOOOOOOOOOLE!
i am really loving the process of lugging around the tripod and really setting up the shot and exposing for like a full second in broad daylight all to get a photo that looks like this. it's quite cool.

i am Enjoying having access to a tripod finally.
30 sec exposures, experimenting with turning on and off phone screens for face lighting. shout out to my friends for being way more creative with that concept than i was initially even thinking

shot at f22, 1/8000.

dithered to 1bit because it's not like any individual pixels had any meaningful data on their own. filesize as very lossily compressed jpg? 20mb+. filesize as a lovely lossless gif? 4mb. lol.

I wanna try this out with some abandoned architecture in good light like this. should look quite nice.

3 monochrome shots, one filtered red, one filtered yellow, one unfiltered
edited in rawtherapee then loaded into gnu imp. subtracted red shot from yellow shot to produce a psuedo green image, then subtracted the red and pseudo green from the unfiltered to produce a psuedo blue image, then composed the three as RGB. phone photo of the scene is attached. I moved my camera slightly and fucked up the focus when unscrewing a filter so it ain't perfect but it is surprising how "accurate" the colors came out.

it's uh. it's uh. it's.... something
SOMETHING!
that first shot is at f8! others are f1.9. all the 1.9 shots metered at faster than 1/1000 despite being nearly completely dark lol.
the images are more usable than idve thought. sort of. i say sort of because all my software struggles to handle these absurd images.
rawtherapee and darktable both fail to display the image accurately while editing so theres a LOT of guesswork involved. and also the filesizes are absurd cuz all the noise. but if i try scaling down the image. well, it looks very different due to the noise being so prevalent that when the pixels get merged they have a big effect.
digikam doesnt even display an ISO value past 65k when browsing my collection lol.

edit: last image shows a pic at 1:1 zoom for an accurate view while editing. reaaaaal small bit of image to go off of on just 1080p x . x

3 yellow filters stacked, doesn't quiiiite darken the sky more than foliage, but it does get close! this is also an evening sky so a little more turquoise than a deep blue. i guess my dream for this look would be like. an industrial machine-vision longpass filter that blocks nearly all blue light. still neat though

normal shot.

normal shot layered with back focused shot with depth of field increased to cover everything but the subject, and a minimum focus distance shot with minimum depth of field.

same as above but "soft light" layered.

same but composed as RGB.

the 43mm f1.9 has quite a lot of focus breathing so its not too good for this effect, still pretty neat though