>>4300746
Nah i scan .dng raws with my dedicated scanner then batch process them in photoshop or lightroom.

Its a bit slower than dslr scanning and a bit lower res but ir dust removal and set&forget is pretty worth.

Not at home right now but basically, after using a tripod for some years, I managed to find an old overhead projector that I now strap my camera to.

The holder is some Chinese AliExpress find and it's probably the one bit I'm thinking of replacing since I want the sprocket holes to show as well.

And since I almost exclusively shoot black and white these days I use an old LED panel that works fine enough, and with a Laowa macro (highly recommended) with my xpro3.

God! what a headache this thing is.

I've tried it. Built a rig and everything but honestly it was a chore compared to my Dimage 3200 with a spare frame.

>Apparently the main caveat of buying a dedicated scanner is slow as shit workflow
Not really a problem if you just shitpost while you wait.
The real problem with scanners is they're all ancient mechanical devices designed with absolute minimum effort and break in any number of ways for no good reason. Mostly they use ancient SCSI controllers with various adapter chips in front to make it work with computers made this millennium.
My old minolta scanner would get desynchronized and stop mid-scan on a regular basis. The firmware had no way to rehome itself so I'd have to open the case and manually turn the screw back to the beginning and try again. My new plustek also randomly locks up mid-scan but so far it only happens occasionally and power cycle has been able to fix it. It's probably only a matter of time. My epson usually works but you have to start it in a specific sequence - first plug in power, wait a minute for it to calibrate, then plug in USB, wait a little while and finally start the software.

In summary, scanners fucking suck. When these shit the bed I guess I'll go DSLR scanning but I don't expect to get as much resolution with 120.

>>4300746
>Apparently the main caveat of buying a dedicated scanner is slow as shit workflow, having to edit the picture with shitty software BEFORE moving to the next frame, and feed it manually yet.
I hate Silverfast with a passion.

>>4300816
Just scan in SilverFast as 48-bit HDR RAW, no edits.
Scan, move to next frame, scan. Repeat. Do until full.
Process images later?

Like, what the fuck why are you actually trying to have SilverFast post-process your film? This is textbook definition of doing it wrong.
User error for sure.
Also, SilverFast is too GUI based and lacks control by number inputs so fine tuning crops is always a pain in the dick. VueScan is easier to use if you know how to click through some settings.

>>4300746
I use Epson scan, don't have to edit before moving on
>Slow
Sure but I can do other stuff, like editing the previous ones actually. I find DSLR scanning very finicky and for 120 I'm fine with the quality I get

>>4300818
For 120 you really need a lot of megapixels and a good lens to get anything better than a flatbed, not surprised.
110 film kind of requires camera scans though. It's all about the magnification/physical size of the film, 120 is big, 35mm is mid, anything smaller is tiny.

>>4300819
it really also depends on the print size you need
at 3000x4500 (Dimage 3200 which costs similar to an average macro lens, 35mm film) you get approximately 38x25cm at 300dpi so very usable for the standard "30x40cm" lab print size
how often do you print above that?

>>4300821
>how often do you print above that?
always, if a picture is good enough to print why the fuck would you print it postage stamp size? 11x14 is the absolute minimum to be worth considering

>>4300823
I guess you are using inches? so 11 inches by 14 inches is 28cm by 35cm so enough for 3000x4500px scan. And this old dimage is by no means a high res scanner .
But let's say you print in order to get the photo framed and perhaps exhibited? A normal silver print in a darkroom would be the way to go then wouldn't it?
I just can't imagine the scenario I need an extremely high res scan for printing very large prints. Then again I don't shoot colour.

also "postage stamp" is usually below like 15cm by 23cm
everything above that until like 30cm by 45cm is considered a normal print size

>>4300817
>Like, what the fuck why are you actually trying to have SilverFast post-process your film?
Lol no, but it's still a clunky piece of shit even if just doing raw scans. It also does 0 compression on DNG files. If you do a lossless re-compress with Adobe DNG Converter you save 50%+ in file size.
>VueScan
The only thing I miss for VueScan is that it can't save the IR data as separate layer but there are better ways to deal with dust and scratches anyway.

>>4300826
Filesize/compression to me is irrelevant but here's the thing.
You don't need DNG format, these aren't camera raws. They're RGB linear data. You can always bulk convert them to 16bpc+alpha PNGs if you want, for free, without any meme Adobe software dependency or proprietary shit.

Also VueScan save the IR data as a separate layer.. the alpha channel. Whenever you use 64-bit RGBI file output, this lets you use it however you want later, or toss it, but it's there if you want it.

>>4300830
That makes a lot of sense. Thanks for the info anon!

>Also VueScan save the IR data as a separate layer.. the alpha channel. Whenever you use 64-bit RGBI file output,
I feel like an idiot for not figuring that out sooner, it's literally in the manual
>64 bit RGBI 8 bytes per pixel 4 samples per pixel 16 bits per sample
>RGBI

>>4300824
Why would you make a darkroom print when modern printers are superior in every single way?

>>4300839
they can't replicate the texture, the feel, the sheer variety of shades of gray that a silver print can
especially on big prints
plus it puts you above everyone else
and it's not like it's a huge endeavor to enlarge a print if you have access to a decent darkroom

on the other hand I've recently gifted twelve A3+ B&W wedding photos to a friend of mine, all printed on a midrange Epson printer and upscaled from the Dimage 3200 scans and they loved it
so for normies darkroom prints are an overkill I agree

I ain't gonna lie when I scan it's mostly for sharing stuff online

File: scan.jpg (113 KB, 1500x1316)

113 KB JPG

>>4300746
Used a Canoscan for years, mostly when my cameras were 12mp and below. Tried to the copy stand approach, but found it just as cumbersome to set up. Now use this, super easy and consistent.

>>4300844
>they can't replicate the texture, the feel, the sheer variety of shades of gray that a silver print can
can we see your collection of silver prints?

>>4300847
no

>>4300844
Yes they can. Thats the whole point. Modern printers eclipsed all else.

And if your “art” needs a primer on printing tech to stand above anything its probably not any good

>>4300851
modern digital cameras also eclipsed everything film related yet here we are

>>4300853
They didnt. Noise isnt grain and bayer still makes up imaginary colors.

>shoot film
>still has to use a digital camera to see their snapshit
l m a o i n g

>>4300856
this
arguably, one could say that B&W (no color filter) surpassed film but in truth that's also not true because almost nobody makes a usable B&W digital camera that isn't tuned for bayer.. scraping CFAs off still applies the base voltage gain from expected loss of light from a bayer filter so noise is still there
plus some fancy B&W films like adox win in resolution still, but nobody has an effective way to truly scan it at a high enough resolution to surpass digital, but at a minimum for 35mm adox you'd need medium format digital

>>4300846
What is that? Where you get it?

>>4300864
You could use a metallurgy microscope with a digital camera, a thin lightbox, and an arduino hooked up to servos that precisely move the microscope stage, and automates taking 100 pictures that you then composite into one 10k+ dpi image. I doubt there would ever be a reason you would want that aside from microfiche-esque use of film, but it could a fun project.

>>4300831
>>4300830
How do you use the IR layer to manually remove dust afterwards? Do you get better results than with automated/integrated ICE stuff?

I respect you guys for still shooting from, but who are we kidding here? You've just spent a small fortune on film and Dev, and now you have to bust out the DSLR with some esoteric scanning contraption. There are better ways to waste your life bros.

>>4301075
The results are worth it. Plus now I don't feel bad that I never actually use my digital camera anymore because it still has a purpose

>>4301025
Vello / JJC 35mm Film Digitizing Adapter kit, some combo of those words will get you there. Some come with the light, some don't.

>>4301075
Bro, my dad left me like a thousand home photos, I am not going to scan all the shitty faded prints at this point, If I'm doing anything I will do it the best I can.

>>4301072
The automated stuff in Silverfast has never really impressed me.
They sell a Photoshop plugin for $200 to do it in post but I have no experience with that (https://www.picture-plugins.com/)

For VueScan files you will have to use Photoshop to turn the alpha channel into a mask and then use the mask in combination with healing or spot removal. Something like that would probably work.
There might be other workflows that I'm not aware of.

>>4301097
Cheers anon indeed simply using it as a mask could be good, I'll try it

>>4301072
>How do you use the IR layer to manually remove dust afterwards?
You can extract it and use it as a layer mask to apply manual touch-ups quite easily.

>Do you get better results than with automated/integrated ICE stuff?
Depends on the software.
For example EPSON's ICE usually bugs out and creates ugly artifacts, SilverFast is a hit or miss, and VueScan's IR is usually quite inoffensive compared to EPSON ICE.

>>4301089
A thousand ain't much, and would be quite easy to scan on a scanner.
Faded photos can usually be restored well. A hundred a weekend would have you done in no time, with little effort. You could do it while doing other things since a fair bit of the process is waiting.

>>4300746
If you buy a copy stand, it shouldn't be hard to keep it perpendicular. Kaiser consistently makes the best ones that I've used. Kind of pricey but sure beats the ol inverted tripod

>>4300746

I have old android phone with photoscan app, manfrotto phone clamp, tripod and move rig with hand as software tells

File: IMG_6433.jpg (1.29 MB, 3024x4032)

1.29 MB JPG

>>4300746
Haven't scanned anything yet as I just got it built, but I'm already looking for free alternatives to negative lab pro.

I don't even shoot enough film to warrant my own scanning rig. Just needed a project to keep myself busy.

[EXIF data available. Click here to show/hide.]

Camera-Specific Properties:
Image-Specific Properties:
Image Orientation	Top, Left-Hand
Image Width	3024
Image Height	4032

>>4300839
Don't care, still making my 8x10 contact prints.

>>4301564
Incredibly based.

>>4301420
>have a level on the rod
>dont bother to make the rod level
you should have two levels on it, 90 degrees apart

>>4300839
Artsy fartsy bullshit about quality aside, there is just something fun about the process anon, does that count for nothing to you? It’s like asking why someone would carve things out of wood when they have them at the dollar store injection molded and cheap. Sometimes the product isn’t the point. But I can only speak for myself. I just love seeing the image “magically” appear on paper in the tray.

>>4301569
Have you ever tried any precious metal toning?

File: IMG_3077.jpg (3.77 MB, 4032x3024)

3.77 MB JPG

>>4300746
My setup, nothing special but everything is replaceable. Pretty happy with Digitaliza+ by Lomography (price and usability are great) but only for 35 mm, I hate so fucking much scan 120 with this.

[EXIF data available. Click here to show/hide.]

Camera-Specific Properties:
Equipment Make	Apple
Camera Model	iPhone 12 mini
Camera Software	17.3.1
Sensing Method	One-Chip Color Area
Focal Length (35mm Equiv)	26 mm
Image-Specific Properties:
Image Orientation	Top, Left-Hand
Horizontal Resolution	72 dpi
Vertical Resolution	72 dpi
Image Created	2024:04:09 16:35:19
Exposure Time	1/33 sec
F-Number	f/1.6
Exposure Program	Normal Program
ISO Speed Rating	500
Lens Aperture	f/1.6
Brightness	-0.3 EV
Exposure Bias	0 EV
Metering Mode	Pattern
Flash	No Flash, Compulsory
Focal Length	4.20 mm
Color Space Information	Uncalibrated
Image Width	4032
Image Height	3024
Exposure Mode	Auto
White Balance	Auto
Scene Capture Type	Standard

>>4301566
Yea I'm still trying to find a way to shim it level. I think strips of aluminum cans should work fine. And you don't need two levels ya dingus.

>>4302673
Is it to do with film flatness?

File: AD196D9F-FC8A-480C-9CB6-D(...).jpg (74 KB, 640x951)

74 KB JPG

>>4300746
I use a 45mp Z7, with the Z 50 micro lens, es2 slide mount, and a george kovacs feather halogen lamp to position tye full spectrum bulb behind this tapered diffuser i made that gets thicker at the center to counter the center-edge falloff. BUT, Then I added three extension tubes between the lens & body, to scan 110 slides filling the frame at 45mp too. That worked surprisingly well. BUT THEN, i left the extension tubes on and stuck a 35mm slide back in instead, ritated the slide holder 90deg, and shot the 35mm film in three overlapping shots, then merged them as a pano in Affinity for a 12k x 8k pixel image with the most minimal slide frame border. BUT, also did each shot bracketed an exposure up & down too, and HDR merged them all together. In tye end i still hold up the slide on a little light table in front of my eye & color grade on the imac 5k display side-by-side until i’m happy. Usually its a blend if tye original kodacrome colors and more naturalistic real-world color balance. But the possibilities are enormous packing all that data in one file. I trued the ZF with its pixel shift feature like others have, but found 32x 24mp scans merged into one was needlessly huge. Each file was a gig and if i exported an uncompressed tiff at half or even quarter the size, it contained the same amount of image data. Only less than that did i start to see a change in the detail of each grain. If you print large and/or hires, finding that threshold so your film grain doesnt look like square pixels is important. And unless you want to spend the time to scan all your slides twice, might as well just go big the first time and do them all right once.

[EXIF data available. Click here to show/hide.]

Camera-Specific Properties:
Camera Software	Google
Image-Specific Properties:
Horizontal Resolution	72 dpi
Vertical Resolution	72 dpi
Color Space Information	sRGB
Image Width	640
Image Height	951
Scene Capture Type	Standard

>>4302673
>Is it to do with film flatness?
yeah, keeps film flat both 35mm and 120 format as well. But I'm talking about normal film, I didn't have any trouble until tried to scan 30yo kodak stock with tube-like shape.
>Yea I'm still trying to find a way to shim it level. I think strips of aluminum cans should work fine. And you don't need two levels ya dingus.
you don't need level at all. Everything you need is a mirror to level camera with holders.

>>4302979
misstagged the comment >>4302997

File: IMG_1670.jpg (181 KB, 1800x1800)

181 KB JPG

I’ve got this pile of shit and nothing else.

[EXIF data available. Click here to show/hide.]

Camera-Specific Properties:
Image-Specific Properties:
Horizontal Resolution	72 dpi
Vertical Resolution	72 dpi
Color Space Information	sRGB
Image Width	1800
Image Height	1800
Scene Capture Type	Standard

>>4303001
I've sold this after 3 or 4 rolls scanned.

>>4303004
Yeah it’s not great. I’ve got an EOS-7D I could do DSLR scans with but I don’t think I have good enough lenses for it. I don’t have a tripod either.

>>4302997
To keep my negs flat i sandwich them between borosilicate microscope slides. They're flat and clear as technically possible. I tried the oil between them but could see no difference pixelpeeping so just run them dry.

On slidesthough, there is a pillow distortion which aside from being bi-axial is also much harder to flatten due to the thick frame. By a stroke of luck, the microscope slide covers that came with the slides & are made the same way are exactly tte right size to fill in the slide window, So I stack two underneath the slide, which makes up for the thickness of the slide and raises the film up to flat most of the time & shoot downward on it. If its going to be a surly fucker it i’ll put two on top as well, and then a big jumbo slide on top of that & weight the ends to sandwich the whole stack flat.

Jumbo borosilicate slides. I wonder if the digitaliza flattens yourstck so you dont need to do all that.

File: DSC00794.jpg (2.45 MB, 3548x2365)

2.45 MB JPG

>>4303050
>I wonder if the digitaliza flattens yourstck so you dont need to do all that.
More than enough, if it even has some kind of distortion I can't see it at f5.6. Grain seems to be in focus corner to corner. pic rel

For flattering you are overthinking as for me. I just use the negative carrier from my enlarger (durst m601) if I work with extra curly and old film stock (not very often).

[EXIF data available. Click here to show/hide.]

Camera-Specific Properties:
Equipment Make	SONY
Camera Model	ILCE-7C
Camera Software	Adobe Lightroom 7.2 (Macintosh)
Maximum Lens Aperture	f/1.0
Image-Specific Properties:
Horizontal Resolution	240 dpi
Vertical Resolution	240 dpi
Image Created	2024:04:08 16:13:22
Exposure Time	1/5 sec
Exposure Program	Manual
ISO Speed Rating	100
Brightness	-2.6 EV
Exposure Bias	0 EV
Metering Mode	Pattern
Light Source	Unknown
Flash	No Flash, Compulsory
Color Space Information	sRGB
Rendering	Normal
Exposure Mode	Manual
White Balance	Auto
Scene Capture Type	Standard
Contrast	Normal
Saturation	Normal
Sharpness	Normal

>>4300800
Just got the 100mm 2:1 Laowa, things fuckin great. Chinesium brands have come a long way.

File: IMG_1931.jpg (167 KB, 480x640)

167 KB JPG

Seeing this thread get bumped has motivated me to finally figure out how to mount my camera to this abandoned 3d printer build that I’ve been sitting on for like two years now lol. It’s a shitty Chinese corexy that never printed quite right and I was going to upgrade it all, but only did one linear rail and got bored with it lol. Lately I’ve been thinking it’s a good fine-motion system Uber copy stand and may be nice for mounting a camera and taking multi-part photos, like for true 1:1 6x9 and so on. Picrel the beast incubating kek. High likelihood this project also gets aborted but hey, didn’t cost nothin.

File: s-l1600.jpg (127 KB, 1000x1000)

127 KB JPG

>>4303238
Man I've been fantasizing about doing something like this, except I actually use my 3d printer so I was planning to cobble it together from chinkshit components like these. Though it might actually be cheaper/easier to start with the cheapest available 3d printer, these linear steppers are like $70 each and I'd need two of them
Kind of curious what would happen if you scanned MF or 4x5 at something absurd like 20k DPI and stitched it, what sort of detail you would get. By my calculations 1:1 should be comparable resolution to drum scanning, but with automated moving and stitching it should be possible to do 2:1, 4:1, etc. with tubes or adapted microscope lenses.

>>4302986
>I did all this stuff and made 1gb files
>here is a half megapixel image

>>4303250
There’s probably a curve of diminishing returns where you just get larger views of grain and not actually more details , that kind of thing

>>4303370
Well sure, but I wonder if more detailed grain would be noticeable. Like if you scan at 1 pixel = 1 grain the grains will all be little squares, but if you scan at 100 pixels = 1 grain you could see the shapes of individual grains. If you made a real big digital print maybe it would affect the texture. The 8000dpi sample drum scans I've been able to find online appear to only be like 2-5 pixels per dye cloud and presumably the grains in b&w would be smaller than dye clouds.
In theory you could even focus stack to get the entire thickness of the film in focus.

>>4303370
Jim kasson tried scanning 4x5 tmax with his gfx100s and between pixel shift and a two shot stitch he was well past the limit where details actually look like details instead of shaky dotted lines (pixel shift looked the best for grain peeping) so hell yeah there is. And tmax is the finest grained shit that isnt a rate ISO 10 film. That’s 400mp with bayer canceled out, so tmax is slightly lower resolution than that if you dont want to see individual grains instead of details (200mp probably, 400mp for cms20).

>>4303393
It would be like using an optically flawless enlarger that doesnt exist. In theory. In practice, the downscaling algorithm could get confused and it would look grainier than a lower resolution scan.
>entire film in focus
Film is thin as fuck, f2.8 macros wide open have it all in focus. They stop down for sharpness falloff and vignetting concerns.

File: Screenshot 2024-04-11 at (...).png (934 KB, 1306x1325)

934 KB PNG

>>4303399
>Jim kasson
Interesting, I checked it out and his results are only convincing me that more resolution is going to improve results. The pixel shift looks much better than without so there may still be room for improvement. Probably not at any normal print sizes for 4x5 but if you did the same for 35mm it might. Like say you wanted to print 30x40 at 300DPI, that would need a 8500DPI scan and more if you wanted to crop.
>the downscaling algorithm could get confused
only if your downscaling algorithm is shit

Found this interesting dry vs wet comparison in "Film grain, resolution and fundamental film particles". I wonder if oversampling and then downscaling (maybe with some specific algorithm to reduce the refracted grain effects) could bring dry scanning more in line with wet mounting.

File: pixel shift 4x5 tmax 2021(...).jpg (74 KB, 1000x755)

74 KB JPG

>>4303416
>More resolution is going to improve results
https://blog.kasson.com/gfx-100/stitch-scanning-4x5-tmax-100/
Individual grains are visible. There is nothing to do after this except count dotted lines on charts and shitpost about it.

File: scaled.png (1.12 MB, 266x461)

1.12 MB PNG

>>4303417
Visible grain doesn't mean all the detail is captured. Unlike pixels, grains have size and shape that take more than one pixel to be measured.
Look, here is the pixel-shift and non-pixel-shift images scaled 1/4 to match the non-pixel-shift bayer groups. The pixel shift is clearly sharper and has more detail. So the non-pixel-shift was not capturing all the information. Without an even higher resolution to do the same comparison, I don't think we can say whether the pixel-shift is capturing all the details or not.

>>4303424
The finest details its capturing look like faint broken lines. The gradations between shades look like spilled pepper. There's nothing worthwhile past this, nothing that would make it to a print unless it was a high resolution billboard (and a very moody lofi one), nothing worth looking at in a crop. It's like enlarging half frame, a ton. Or zooming in on a digital photo and saying "those 3 pixels are my reflection in the eye cool huh".

>>4303424
top looks better

>>4303426
Sorry buddy, I spent too long listening to Nikon's gaslighting with their "12MP is all you need" and "muh big photosites" to fall for the "let me guess, you need more" meme
I DEMAND 400MP 35mm scans and won't be placated with anything less

>>4303430
Big photosites on digital cameras ARE better in low light, for high ISO color accuracy, not for sharp little details. Guess what people notice first - faded colors, or soft eyebrows?

Scanning 35mm with 400mp is going to just show you the voids between individual grains. Unlike digital noise (signal errors) visible grain is the image tearing apart. You dont actually want to see grain. Large format film is for grain free enlargements not jerking off the number of megapixels or magnification factor on a microscope needed to count the smallest lines on a test chart.

>>4303439
they will hate him because he's telling the truth

>>4303439
Holy based.

File: 7Q0A6911-2.jpg (3.48 MB, 4480x6720)

3.48 MB JPG

>>4303250
>>4303238
got around to finishing a mounting print and throwing a magic arm on the thing and it was actually quite smooth. was able to shoot a full 2:1 on a 6x8 negative fairly painlessly lol. one thing i discovered is that as sturdy as magic arms can be, the single small mounting point makes them whole thing susceptible to the camera's moment of inertia. gotta find a sturdier way. it just meant waiting 3-4 seconds for the thing to stop wobbling and using a remote trigger. with some more refinement it might be a pretty wicked way to scan things, if tedious.
problem is it seems i hit those diminishing returns. and an unexpected issue is the workflow for processing 32+ images. Photoshop twisted itself into knots merging it and took forever. and doing any editing or work at all dragged, and died trying to export. affinity photo did a merge quite impressively quickly, but it is locked into some kind of spherical projection for its panorama merge which is useless. seems aside from fun experiment, not really worth the trouble.
>>4303575
picrel one chunk of image compared to the full thing.

[EXIF data available. Click here to show/hide.]

Camera-Specific Properties:
Equipment Make	Canon
Camera Model	Canon EOS R
Camera Software	Adobe Photoshop Lightroom Classic 13.2 (Macintosh)
Image-Specific Properties:
Horizontal Resolution	240 dpi
Vertical Resolution	240 dpi
Image Created	2024:04:11 23:07:11
Exposure Time	1.6 sec
F-Number	f/8.0
Exposure Program	Aperture Priority
ISO Speed Rating	100
Lens Aperture	f/8.0
Exposure Bias	0 EV
Metering Mode	Pattern
Flash	No Flash
Focal Length	100.00 mm
Color Space Information	sRGB
Exposure Mode	Auto
White Balance	Auto
Scene Capture Type	Standard
Contrast	Normal
Saturation	Normal
Sharpness	Hard
Unique Image ID	A4E23521918A58A3DEFB6F9562DFE8FA

>>4303417


>>4302986
>large and/or hires, finding that threshold so your film grain doesnt look like square pixels is important.

Hey ken rockwell, go ahead and scan your crappy snapshits of your suburb or your fat kid in your crappy 12mp jpg. Save en as 8bit too, bc you cant see that difference either. Its really cool that you have no idea the difference between an archival scan and whatever crap you think is “good enough” for your snapshits, but the thread full of people building rigs for archival scanning is not going to be the target audience for your worn-out shtick.

>>4301075
When you build that time machine let me know so i can go back in time and give my photographer grandfather this z7 before he shot 50 years of work. It’ll make my life 1000x easier.
>just choose one or the other, i cant handle two kinds of cameras existing, one must be superior
No, the kind of aspie nerds that fill the internet think photography is about the gear. They are fucking hopeless & need to imagine everything is a black & white, all-or-nothing binary choice. These people are idiots. The world didn't stop painting landscapes just because the Sony A7ivxrS3 shoots it at 600megapixels. Dont be so fucking german.

>>4303591
Neat. Are you controlling it just with gcode over serial? I am curious what the accuracy/repeatability of the positioning is. Best case is it would be accurate enough that the images would just line up and you could combine them with an imagemagick script or somesuch. I doubt it is that accurate but maybe good enough to use something less computationally intensive than full pano stitching.
With gcode scripting + tethering + automatic stitching it shouldn't matter so much how tedious it is, 'cause you'd just kick it off and go shitpost for a while like you'd do with an old fashioned scanner.

>>4303631
I used commands over serial with pronterface for this. It has manual controls that can override the end stops in the firmware and I couldn’t figure out which g codes do that. I ripped the tool head off and this printer had 2/3 homing end stops on the tool head so I couldn’t use typical g-code cause it can’t home lol. I feel like with a solid sturdy mount, it would be fairly repeatable, it’s on a sturdy linear rail and the steppers resolve to like 0.1mm or something like that. I think it would be more useful for 4x5 or larger formats maybe. Even at 1:1 would still take a few shots.

Is PEC12 still the good film cleaner, or is there something better now?
I have have a whole stack of absolutely filthy slides from the 70's.

File: IMG_5596.jpg (32 KB, 640x427)

32 KB JPG

I can’t decide between proper Epson V850 scanner setup with original Epson fluid mount
or
camera scan method with Negative Supply light source (99cri) and 4x5 sheet film holder with ANR glass

[EXIF data available. Click here to show/hide.]

Camera-Specific Properties:
Image-Specific Properties:
Horizontal Resolution	72 dpi
Vertical Resolution	72 dpi
Color Space Information	sRGB
Image Width	640
Image Height	427
Scene Capture Type	Standard

File: IMG_5598.jpg (253 KB, 800x600)

253 KB JPG

>>4308275
Or Durst enlarger and use is with camera scan metod.
For the price of Negative supply light source and film holder I can almost buy Durst 1200 enlarger

[EXIF data available. Click here to show/hide.]

Camera-Specific Properties:
Image-Specific Properties:
Horizontal Resolution	72 dpi
Vertical Resolution	72 dpi
Color Space Information	sRGB
Image Width	800
Image Height	600
Scene Capture Type	Standard

>>4308275
From the samples I've seen wet mounting is probably not worth it with a flatbed, though you need something to keep the negative off the glass to avoid newton rings. With 4x5 you should get loads of resolution, more than DSLR unless you stitch like >>4303591 That's a big difference between DSLR and scanner; with a scanner resolution increases with negative size, while with DSLR every negative gets the same resolution.
I myself have purchased an "unrepair" (returned) ender 3 with the goal of stitching DSLR scans of my 6x17s like that other anon, but haven't got around to mounting the DSLR on it yet. These maybe-broken printers are shockingly cheap and I am impressed with how sturdy it is. Stay tuned.

>>4308275
>99cri
Also btw I'm pretty convinced this is pointless wank for scanning, with the possible exception of slide film but maybe even then. CRI is about how close all the different colors in a scene match what you would see with your eyes under daylight. But negative film is never meant to be viewed with your eyes under daylight and when you're scanning film there are only 3 colors (cyan, magenta, yellow dyes) that you're trying to capture for digital processing. A lower CRI might even improve channel separation. I'll be using a $13 light box and see how that goes.

File: 20240429_192118.jpg (1.21 MB, 2000x1125)

1.21 MB JPG

It ain't much but it gets the job done. I need a extension tube so I can use more than 1/4 of the sensor.

>>4308278
The 3D printer is definitely an easier way to do stitching, I’m the other anon you mentioned. I haven’t made progress yet since I decided to drop the magic arm mount for something more stiff, but haven’t come up with something yet. Should be easy enough to do on an ender 3, remove the hotend and bolt a clamp/ball head into it. The only issue I can foresee for you is the much heavier weight of the camera setup chewing up your POM wheels, but I’m not certain. Little tip I discovered from my 3D printer dicking around, don’t remove the homing end stops unless you want to have it smash and grind on itself before you pull the plug lol. Not printing but still need them if your printer is setup to home before you can move it

File: enderscan_rig.jpg (696 KB, 1608x1330)

696 KB JPG

>>4308524
Here's what I came up with. No ballhead, I just drilled some holes in a board and attached it where the hotend used to be with the pre-tapped m3 threads and the camera with a 1/4" bolt. I shifted the Z stop up so that the lens can't crash into the plate (had to make a little extension wire) and homing works fine.
I did some test shots where I took a picture, moved 5mm in x and 5mm in y, then back to the old position and take another shot. It's just a few px off which I figure is good enough. The red dot is the same pixel coordinates to show how close it is getting the scratch at the same position. I calculated this as equivalent 3300 DPI. The light source is a $12 tracing table with two layers of ground glass on top to smooth it out. Honestly hardware-wise this was super easy.

To make this good I figure I need to
- get a better lens (currently have an old chinon 50mm with tubes)
- brighter light source (currently at 1s exposure which I'm sure will cause motion blur)
- power cable for the camera (if the battery goes dead I'll have to remove the camera which will mess up any manual alignment/focus I've done)
- buy some film holder instead of just taping it to the ground glass
- write an "autofocus" script that takes pictures at various Z offsets and picks out the sharpest one
- write a script to take all the pictures, calling pktriggercord and writing the gcode commands to the serial port
- write a script to merge all the pictures (maybe using some of the panotools components?) and puts it in a format I can feed into darktable/rawtherapee

Ideally those last three would be some general purpose utility that you just run and it does the scan. Might not get to it for a while but this is definitely showing promise.

File: yzot1poqk8ub1.jpg (30 KB, 461x461)

30 KB JPG

>scanner stops enumerating over USB (again)
I hate scanners so much bros
guess I'll try rebooting

>>4308921
With how much boomers charge for decent scanners i'd rather buy a macro lens

File: raw0013.jpg (1.26 MB, 1500x1017)

1.26 MB JPG

>>4308921
get an old scuzzy piece of shit, i know u wont regret it

[EXIF data available. Click here to show/hide.]

Camera-Specific Properties:
Equipment Make	Nikon
Camera Model	LS-2000
Camera Software	Adobe Photoshop CC 2018 (Windows)
Image-Specific Properties:
Image Orientation	Top, Left-Hand
Horizontal Resolution	300 dpi
Vertical Resolution	300 dpi
Image Created	2024:03:10 18:34:19
Color Space Information	sRGB
Image Width	1500
Image Height	1017

>>4308517
The camera is nice and all but I’m getting moist at that beautiful test gear in the background.

File: hugin_stitched.jpg (185 KB, 1500x1139)

185 KB JPG

well it "works" by some description
The hugin auto-align scatters the images randomly. It's probably confused by the heavy vignetting.
Also the image was taken on a shitty box camera so there's not much detail for it to latch onto.

[EXIF data available. Click here to show/hide.]

Camera-Specific Properties:
Equipment Make	Pentax
Camera Model	K-7
Camera Software	GIMP 2.10.36
Image-Specific Properties:
Image Orientation	Top, Left-Hand
Horizontal Resolution	300 dpi
Vertical Resolution	300 dpi
Image Created	2024:04:30 16:34:31
Color Space Information	sRGB

Do any of you make money scanning film for others?

File: magicked.jpg (2.19 MB, 3000x2260)

2.19 MB JPG

The naive imagemagick strategy would almost work, if it weren't for that awful vignetting
seems like exposure is a bit uneven too, not sure what's up with that

[EXIF data available. Click here to show/hide.]

Camera-Specific Properties:
Camera Software	GIMP 2.10.36
Image-Specific Properties:
Image Orientation	Top, Left-Hand
Horizontal Resolution	72 dpi
Vertical Resolution	72 dpi
Image Created	2024:04:30 18:03:19
White Point Chromaticity	0.3
Color Space Information	sRGB

File: vignetting_partly_fixed.jpg (1.45 MB, 5580x3720)

1.45 MB JPG

I feel like I'm about 80% there but that final 20% is gonna be a real bitch

[EXIF data available. Click here to show/hide.]

Camera-Specific Properties:
Camera Software	GIMP 2.10.36
Image-Specific Properties:
Image Orientation	Top, Left-Hand
Horizontal Resolution	72 dpi
Vertical Resolution	72 dpi
Image Created	2024:05:01 12:06:15
Color Space Information	sRGB

>>4309319
It might honestly be easier just to reduce magnification to tame that vignetting, like damn. Might need to get a proper macro lens for it to work more smoothly

File: partial_scan_28mm.jpg (818 KB, 5170x3760)

818 KB JPG

>>4309376
yeah the lens is a problem, it would work a lot better with a real macro. I found this old 28mm lens which has less vignetting with the tubes. I also suspect some of the banding was due to ambient light so I made it a little hood which helped I think. I am cropping just the center of each image to reduce the vignetting, so it's not as much magnification as it looks.
Battery died mid-scan but it's almost usable now. I ordered an AC cable which I think I'll need to scan any real pictures, since I'm combining 391 images and that's probably more than what a full charge can handle.

[EXIF data available. Click here to show/hide.]

Camera-Specific Properties:
Camera Software	GIMP 2.10.36
Image-Specific Properties:
Image Orientation	Top, Left-Hand
Horizontal Resolution	72 dpi
Vertical Resolution	72 dpi
Image Created	2024:05:01 17:24:02
Color Space Information	sRGB

File: kit_lens_scan.jpg (1.69 MB, 4542x2358)

1.69 MB JPG

The kit lens with no tubes is the best yet, though the magnification is less.
I found that the wide angle was amplifying any non-flatness in the film, so now I'm at 55mm. You can see the ghosting at the top which I assume is caused by a bow in the film. Maybe I should buy one of those digitaliza film holders.

[EXIF data available. Click here to show/hide.]

Camera-Specific Properties:
Camera Software	GIMP 2.10.36
Image-Specific Properties:
Image Orientation	Top, Left-Hand
Horizontal Resolution	72 dpi
Vertical Resolution	72 dpi
Image Created	2024:05:02 07:33:45
Color Space Information	sRGB

File: SDC10283.jpg (476 KB, 2568x1044)

476 KB JPG

I just use a point and shoot over a light box

[EXIF data available. Click here to show/hide.]

Camera-Specific Properties:
Equipment Make	SAMSUNG
Camera Model	SAMSUNG TL100
Camera Software	Adobe Photoshop CS6 (Windows)
Maximum Lens Aperture	f/2.8
Sensing Method	One-Chip Color Area
Focal Length (35mm Equiv)	35 mm
Image-Specific Properties:
Image Width	2592
Image Height	1944
Number of Bits Per Component	8, 8, 8
Pixel Composition	RGB
Image Orientation	Top, Left-Hand
Horizontal Resolution	96 dpi
Vertical Resolution	96 dpi
Image Created	2013:05:21 18:00:14
Exposure Time	1/173 sec
F-Number	f/3.0
Exposure Program	Normal Program
ISO Speed Rating	80
Lens Aperture	f/3.0
Exposure Bias	0 EV
Metering Mode	Multi Spot
Light Source	Unknown
Flash	No Flash
Focal Length	6.30 mm
Color Space Information	sRGB
Image Width	2568
Image Height	1044
Exposure Mode	Auto
White Balance	Auto
Scene Capture Type	Standard
Contrast	Normal
Sharpness	Normal

File: 1553110018781.jpg (161 KB, 579x546)

161 KB JPG

>>4308921
bro that aint so bad
if i try to scan two strips of film in batch mode in a row using a sa-21 with coolscan 4000 and vuescan without rebooting the computer, then it crashes and i lose the scans from the second strip