The state of dedicated film scanning is so bad these days, especially for larger than 35mm formats that it seems most people still shooting film are resorting to using a digital camera and a macro lens with something to hold the film.
You're correct. Most people who do their own scans either use a macro lens or a telephoto with extension tubes to create a macro lens of sorts. Then you need a negative holder such as this one: https://www.negative.supply/shop-all/basic-film-carrier-35. You need a light table or LED hot light, and a way to hold the camera and you're in business.
Personally, I've been scanning my own 35mm negatives for the past few years with both a Full-frame DSLR and an APS-C Mirrorless camera, and get much much better results than I'd been getting via my local film processing lab and their scanning.
Scanning negatives with a modern digital camera still retains a lot of the filmic quality that we've come to expect via shooting film. Obviously you lose the "resolution" and ability to blow those images up once scanned, but you can 100% still see the grain, color reproduction, and analog anomalies that we're used to by shooting film.
Yeah, I have recently switched to "scanning" my 6x6cm and 35mm negatives with a macro lens and digital camera. It's probably not as flat out perfect as a really good scanner, but the end results look amazing and fit my workflow for digital images as well.
I recently got a camera with "pixel shift" technology and using that the files have 96mb and really show off the grain. Plus it's just faster than scanning ever was. I am really very happy with the setup.
If you don't use a planar lens it will never really look good.
What I do for my part is just use an enlarger to which I attach my digital camera with various tubes and adapters to get the distances right, and an enlarger lens. These lenses are infinitely cheaper than "regular" planar lenses and they're just made exactly for the job, and the whole assembly is rigid and aligned so it's easy to get the whole negative right.
You don't "need" much tbh. My setup is a cardboard tube, a lens I found in a broken scanner and some tape. I'm using an old kindle as backlight since it doesn't have pixels = no need for diffusion.
I wet scan on a flatbed, though I will admit the process is tedious and holding a negative up to a lamp and taking a snapshot with my phone is 80% as good ;)
Wet scanning was the default for drum scanners, the gold standard in the 90s. Essentially, you lay a drop of mineral oil on the glass drum, create a bond with the negative and then tape it down for good measure since the drum spun at a pretty good clip.
Wet scanning is worthwhile if you have a higher-end flatbed scanner like the Epson V700, 750 and 900(?) because they use higher-quality sensors and real lenses.
I almost bought a drum scanner a decade ago but would have to maintain a Mac Quadra for it to run. For large format film, a wet mount scan on an Epson V750 with VueScan software is great.
Scanning medium format can be a pain but I've had decent results with an Epson V700 for 4x5 film. A lot more cost effective than a Flextight and I maintain some VMs with the firmware on various operating systems that have traveled with me through various desktop builds.
I'd like to see some better FOSS options but images are a lot better quality than they were using cheap desktop film scanners in 2005.
Another one is that nobody made a feeder for movie film, for ANY of the consumer or "prosumer" scanners at any price.
If you want film scanned today to individual files per frame, it's still a bunch of money if you can find someone to do it... and good luck getting it done right.
The movie film scanners that are top notch are easily in the six figure range, and go up depending on the features. In a past life, I was researching a scanner to buy for film restoration purposes, and I can't remember the name of the company or unit, but it would move each frame into place and then "gently" press down to flatten the film, and then triple flash it to monochrome CCD with the appropriate RGB filter. It was a big machine and with the flattening and triple flashing was not real time capable, but not as slow as I would have expected. Somewhere between 12-18fps for 2K and slowed down for 4K+ scanning. The project was never funded, so it was all for naught, but I enjoy getting to research new toys like that while on the clock so to speak.
I understand (I work in the industry), but the point still stands that these companies could have added greatly to the appeal of their scanners by offering a motion-picture attachment, no matter how slow.
I was considering trying to build one out of an inspection microscope, but gave it up and just had my most-important reels scanned to TIFFs.
>I was considering trying to build one out of an inspection microscope, but gave it up
justifying the cost of the people that did not give up and made an actual product. "if this shit was easy, everyone would do it!" is something i remind myself all the time
Sure. I don't begrudge the cost of the devices. In this case I'm calling out the missed opportunity of having built most of such a device already and not following through to expand its appeal to more potential customers by offering a particular accessory.
Also the opposing viewpoint to remind yourself of is that hey, SOMEBODY was the first to make one of these things and started a company to do it! I've given up on ideas before starting because of the manufacturing technology required, only to see a company like GoPro come out of nowhere and own a market segment.
The Northlight is definitely one of the units we looked at, and was a top contender. Ultimately, it was between the Northlight and a second vendor, but funding was pulled before we ever got to make a decision. So, no new toys for me and the team =(
Cool! Thanks for the link. I checked out a few DIY projects years ago, and even bought a few broken Super-8 projectors to scrounge their film-advance mechanisms. But I haven't revisited the state of the DIY art for a long time.
This is a good point. Speaking of points, in the way a laser turntable pickup is noisier than a needle, would an electron microscope scan of a frame of film be best, or is there a sweet spot where the grain noise is at an optimal minimum?
I shoot lots of 120mm on a Mamiya 7 and am currently stuck with an Epson V600 as my best option. Unfortunately, the scanning experience sucks for 6x7 photos because of how not-long the bed is. You can only scan two 6x7 photos at a time, and then have to turn the film around. It's a tedious process that means most of my film remains unscanned.
As others have mentioned here I am also looking into trying to fauxscan with a DSLR and a macro lens, but I'm also wondering if there is enough of a market/interest in a more DIY/FOSS scanning project.
I've considered trying to build something more standard that could be printed or manufactured, that combines some small pointgrey (CV) cameras and macro lens, a light, and some film rolling mechanism to keep the film tight - the idea being that you could basically "scan" the film as a moving roll, scanning a whole 120mm roll in seconds. Pointgrey cameras also have a great api, so you could actually build software on top of it.
I've long thought about putting time to this, but the intersection of pro film shooters and programmers is so small that I'm not sure it would be worth it. If anyone else is interested in the idea, I'd be happy to collaborate on it!
Alternatively, the state of digital cameras is so good these days that they can be used to digitize film. https://www.youtube.com/watch?v=6pnqF8G_wiM highlights a setup with a Panasonic G9 that features an 80MP in-camera high resolution mode. A used body can be found as low as $500 in the USA, and it's a camera too!
Through various driver discs and archived files online I was able to compile some resources and threw them up on Archive.org (available on the FlexColor page of my site).
My biggest gripe is when people don't document fixes. This whole rabbit hole started because Hasselblad removed the software downloads from their site and the rest of the community doesn't really have a clue when it comes to software preservation. Flat CMS's are important to the process as well. It looks like Archive.org picked up on the page because of this post and made a couple snapshots:
https://web.archive.org/web/20230126040120/https://pathar.tl...
> A Hasselblad Flextight X1. Google the thing for the price if you’re curious, and no I wasn’t going to pay that much, I was going to pay less than half of that (a bargain to be fair).
Even for consumer scanners, it seems I’m helping my parents fight unmaintained proprietary firmware on Windows every few years. Eventually I just set them up with SANE.
Sadly it seems SANE has no Hasselblad drivers at all, which probably makes sense given that these are professional devices with a very small install base. Even worse, given the high end features I suspect the firmware isn't just a clone of some other device like it is so often with scanners.
Probably not too hard to reverse engineer with some skills and tools, but well beyond what a random photographer is likely to be able to do. The real shame is the company squatting on the driver source code and refusing to update it. It is supremely frustrating when companies treat the driver software--useless without the hardware for it to run--as top company secrets. The software does nothing without the hardware your company sells, so who cares if someone could "steal" it?
> It seems the Firewire ports on these things start to go bad after about a decade. If that happens you’re looking at a 3,000 Euro repair bill if you can find someone with the parts capable of replacing them. Hasselblad will still repair them but it’s a pain to get the scanner to them and they charge almost twice as much as third part service shops.
...
> This thing is remarkable. The bigger brother (the X5) even more so. Imacon/Hasselblad had a load of patents on the technology that means no other manufacturer can replicate it. There are other high spec scanning solutions, of course, but none that come close to this form factor.
...
> 15 years passed, at which point Hasselblad discontinued the scanners. The cost of modernising the interfaces was not worth it. In fact, only 7 years passed before Hasselblad effectively discontinued them as that was when they stopped updating the software.
This is the dark side of patents.
A company that has used the patent system to run everyone else out of business despite not being particularly innovative (the author describes a "simple" system for assuring the film is perfectly flat), refuses to keep their product line up-to-date or properly support it, charges a fortune for a service that from the sounds of it doesn't actually fix the problem with the interfaces, which may have been purposefully designed to fail anyway...and an entire market segment just dies.
It's sad that those $3000 bills for repair will probably be going to organizations like museums trying to preserve their collections or make them more accessible. Or be unaffordable to such organizations.
They might have a patent on this particular setup, but the FlexTight system produces worse results compared to traditional drum scanning. Due to the way it uses a lens between the CCD sensor and the drum also means that you get some quality issues around the edges of larger frames. It also means you can't scan large formats like 8x10. It's cool tech and it is a shame that it's stuck behind patents, but it's not like there's not other options out there.
Also the hardware is funny. IIRC they didn't iterate on the actual components much, and the move to FireWire basically just integrated a SCSI->FireWire adapter into the case. I haven't seen the inside of an X1/X5, but I'd be interested if these could be repaired by using a different SCSI adapter of some type.
Yep, I deal with this all the time in industrial automation. It's even worse than the consumer space there, because the cost of breaking stuff is so high that the default is to invent something and then coast on your laurels for a few decades. Infrastructure and tooling that somebody's dad bought 30 years ago is now too expensive to update, so you keep it limping along well past its normal service life...
In particular, the author laments the high cost of scans:
> Here in Switzerland I am looking at a cost of around 30.- CHF a frame minimum if I get a third party to scan them for me.
And worries that you have to have something like the 2007 iMac to run it:
> So you need a dedicated old rig to run the proprietary software to drive the scanner.
But doesn't seem to quite connect 2 and 2 together: Those third party shops all have ancient dedicated rigs that they bought a decade or two ago. The use of that equipment is what your $30 per scan is buying! There are label printers and laser markers and time clocks and press brakes and CNCs and inventory management systems at shops all over the world, running some PLC RTOS from the 90s, or DOS or Windows 98 with PXI cards or printer parallel ports or even completely proprietary logic boards, that are decades old and you can't get parts or service or updates for them.
It's ubiquitous in any less profitable industries that aren't on the cutting edge.
It's hard to maintain and gradually update this equipment. I fear it's going to get harder as IIoT services die off and engineered lifetimes shrink. It's hard to debug a machine built before the Internet, with coffee-stained schematics and only enough IDE hard drive space for 8-character variable names, but I worry it will be harder still to debug a machine built well after the Internet with ubiquitous documentation available only behind a login to a server that's no longer online and enough hard drive space for gigabytes of third-party libraries...
I mentioned this in another comment, but sometimes people who are dealing with these systems just don't dive in far enough either. I was able to get my FlexTight scanner working on Windows 11 64-bit by modifying the INF for some existing 64-bit drivers. For some reason nobody had tried this before, or even just documented it. Even if there were no 64-bit drivers I've gotten it to work on a Windows 10 32-bit install.
Especially in the photo world there's so much tech being used, but not a lot of people that understand the implementation/engineering of it. Not faulting them, but I can't tell you how many people I know that just work off external drives. No NAS, no backups, just originals spread across fragile external hard drives. Hell, I knew some people that up until ~2011 still stored originals on ZIP disks and never managed to move them to something a bit more stable.
Going back and working on older Macs is an absolute pain. Now there's four separate architectures that may or may not have the right software/driver at varying levels of implementation. It's amazing how often the answer is "just use Windows". Not that it's perfect, but at least you weren't dumped like last week's spaghetti.
I dove in pretty far - to the point of running dtrace on the flexcolor software, and it seemed to get stuck in a thread that was receiving data from the scanner. The Firewire port is dying (well, now it seems to be dead) so this is not a software issue and more a "i'm not prepared to drop 5,000 Euro on something that may not be fixable".
Sounds like maintaining these high end medium format cameras and scanners can be a real hassel.
I have a Nikon Coolscan V and nothing is wrong with it but the switch, but it’s hardly worth repairing or trying to sell on eBay. Working, it might be worth more than I paid for in c2005.
Maintaining high end medium format cameras is really not bad at all -- most are very simple and many have no electronics, so as long as the shutter mechanism continues to work, everything else is a piece of cake.
A C>oolScan V is worth a fair amount of money already, and will only increase in price. Send it to Frank Philips, who repairs CoolScans more or less full time. You can find him on the FaceBook CoolScan group
I don't have to create 160cm wide prints but I have scanned medium format frames on an Epson v600 flatbed and the results with the right software is more than serviceable. In fact, Nick Carver uses a v800 to scan fairly high quality images he creates big prints from. Although he does go to a drum scanner shop for the mammoth 6x17 prints but it simply goes to show that the return on value is very high on the right consumer grade scanner.
someone shooting on a vintage hasselblad that still costs used near as much as many modern high tech cameras and printing 5'x2' finished pictures isn't interested in serviceable. serviceable isn't the result they're looking for.
The camera is an optical device that along with hasselblad lenses has (many) detectable differences in the pictures taken, along with film choice. Someone shooting with a particular camera aiming for large prints is aiming for high excellence not serviceability.
there are bad cables (shielding etc) but the camera+lens+film is literally a sensor/recording device and not really comparable. Whether you like the thing they're doing is a separate consideration, but the person in the article is not looking for ok or even good results. They want everything the best it can be.
Scanning is almost always the weakest link when it comes to film photography these days and sometimes it's really bad, it's not a "copper vs gold plated cable" scenario, more like a "20$ logitech speakers vs "$2000 semi pro setup" to listen to .flac scenario
I mean, yes, I did qualify that I don't want the same fidelity from my scans but I also shoot on a Bronica SQ, which is no Hasselblad but was still professional equipment when it was released and is still extremely well respected and gives me great results.
Prints are also extremely forgiving, especially viewed from a normal viewing distance, from personal experience of having made high quality prints of my photos.
But again, you can choose to spend however much you wish to to get the last bit of detail out of things.
My dad used to produce prints in the 90's with a hybrid digital/film workflow. Scan with a FlexTight -> post in Photoshop -> write back to 4x5 with a LaserGraphics film recorder -> darkroom printing. Using this process he did up to I want to say 40x48 prints.
I used to work in a photo studio in Germany, we had 2 of these and the predecessor to the Hasselblad/Imacon X1. The difference between them was amazing. I'd still love to have one of those Coolscan 9000s, they are great machines. Even back then (around 2004) we used Vuescan for the Nikons. It was already an issue to get up to date drivers for our Macs.
The X1 was probably better, as it was a [almost] drum scanner.
But scanner tech seems to have basically "hit a wall," in the last decade, or so. Not many advancements in the imaging. I think digicams and pure digital images/documents have made it difficult to justify the cost of developing them.
The Flextight is definitely better than the CoolScan, but not approaching drum scan quality. A very good CCD scanner with a great enlarger-quality lens, and the transport is very good at keeping the film flat along the line scanned by the CCD, by the expedient of bending it cylindrically, but it's not the same dynamic range as the photomultipler tube on a drum scanner, or even what the old $50K prepress X-Y flatbed scanners like a Creo Scitex or Fuji Lanovia could do.
Oh I know/agree. It is literally described as a virtual drum scanner by B&H for example though, as inaccurate as that is. To me the main difference is CCD vs. photomultipler.
VueScan did not support Digital ICE4, which was the only way to scan Kodachromes with automatic dust removal using the infrared channel (Kodachrome is opaque to the infrared LED used on the Coolscan V ED or 5000 ED). Only Nikon Scan did, and possibly SilverFast (Digital ICE was invented by Applied Science Fiction, a subsidiary of Kodak).
This is nice-sounding, but ultimately there is almost no incentive for hardware manufacturers to open source their drivers. Linux isn't a large market and macOS and Windows users generally aren't concerned with it. On top of that, many of these products (though maybe not the one in particular) is basically disposable.
The way this is handled in industry is to have 20-year-old computer running Windows XP that is just for a specific product that only has drivers for one version of Windows.
If more people committed to open source, there would be more pressure on hardware vendors.
I suppose all that's needed is for a good chunk of Windows and Mac users to stop thinking of open source as only Vegas flings of opportunity, and more like a life partner.
Has this ever worked? Even now it seems that hardly any vendors put out linux software and when they do it's garbage that tries to import shared libraries that existed on ubuntu 13.04 or something. Almost all the hardware support on linux came from people in the linux space reverse engineering and developing those drivers which are then in a state that can be upstreamed.
The wine/proton project has had far far more impact than decades of posting on game forums asking for a linux build.
> Almost all the hardware support on linux came from people in the linux space reverse engineering and developing those drivers which are then in a state that can be upstreamed.
Almost all sounds like exaggeration, Intel&AMD have done a lot, and even ARM vendors have contributed lot of drivers. See for example LWN author stats for employers, plenty of HW vendors there: https://lwn.net/Articles/902854/
Open source (OSI definition) is exactly the same thing as Free Software (FSF definition) so your comment makes no sense without clarifying further. Are you perhaps referring to Source Available software vs OS/FS?
Stallman made up his own definition of open source which is different than the common definition. The OSI best sums it up here https://opensource.org/osd their definition is essentially the same as the FSF essential freedoms. Without classifying "Stallman's definition of Open Source" people will assume something much like the OSI one.
What Stallman calls Open Source is what the rest of the world calls Source Available. Stallman shows his lack of communication skills again here by endlessly fighting to redefine words for no purpose.
Stallman appears to be considering "source available" licenses which forbid modification (and sometimes deployment) as "open source." This is not the definition of open source that everyone else uses. Certainly the OSI would not agree. There is more information here: https://opensource.org/osd
RMS was around in the 1980s when "open source" was still being defined. So maybe he thinks he can define it however he likes. But to the rest of us it doesn't make sense, like saying "software that has a free software license, like GNU Emacs and Microsoft Word." One of those two doesn't belong in that sentence, and saying that I'm using my own custom definition of "free software" is a pretty shitty way to argue.
The problem is in messaging. The whole point of the OSI was that the FSF was too ideological to be appealing to most businesses. The problem, is that the ideological angle is the point. The OSI would have never pitched the GPL, and the FSF would have never pitched the MIT license.
When you release software under a license that makes approximately zero demands of the other party, the other party won't respect your software. When you release software under a license that makes even low effort demands of the user, like reciprocation, they will.
The MIT license is a FSF certified "Free Software" license, it provides all of the "Essential Freedoms". The MIT is not "copyleft" like GPL is but copyleft is not the same thing as free software.
Once again, that's not the point. The ideology required to promote copyleft licenses at all is the point. The OSI's standpoint is that you can distribute software that provides the user with their freedom. The FSF's standpoint is that you are ethically obligated to.
Please consider not calling it "free software". I've tried to explain to RMS that it's terrible communication, but I think he likes the wordplay.
The term "free" product already meant something to people, and continues to mean that.
There is generally no opportunity to (as RMS might imagine) say "I'm glad you asked about the name. We mean free as in freedom. Let me assemble my lecture podium and explain..."
Instead, people who hear the term have been hearing the opposite of the intent.
Which leads us back ot the current problem, of people thinking it's all random hookups with open source, rather than meaningful relationships.
I don't think it's a naming problem, I think libre software has been under direct attack on a similar angle to "embrace, extend and extinguish".
But yeah I agree naming doesn't help. I'm biased because in my language libre and gratis are distinct. I'll keep that in mind, even though I'm a bit pessimistic and don't think it will help, I guess it won't hurt either.
Agreed that naming isn't the only problem. It's a complex environment of different understandings and goals, including adversarial dynamics (and sometimes very underhanded behavior, including EEE).
I think the foot-dragging by some on the name, after it's been pointed out to them, is kinda symptomatic of their difficulty operating in this environment. And they often end up discrediting action that could be viable, but needs to win buy-in from others.
Stallman is an odd character. He's deeply passionate about the FOSS movement, and its success means more to him than anything except his jargon choices. He'd rather see the whole thing burn than change the words he demands it be described with.
That's a bit unfair. He passionately advocates for his choice of language because he thinks it's important to the future of free software. You might disagree with him, but he's not exactly shy about justifying his opinions.
Software freedom didn't really disappear from public discourse, it never appeared in the first place, unless you count amongst a bunch of nerds nobody in the real world cares about or listens to. It may sound harsh, but that's reality.
Open source appeared in the public discourse because it had tangible benefits and people could understand what it meant.
I don't think open source killed the concept of software freedom, I think the concept of software freedom as a concept literally never caught on to begin with.
Honestly, I'm a bit tired of the Stallman worship at this point. A lot of people were pessimistic about the future of software and the internet, and they're not all prophets. He observed what was happening when other people chose to naively ignore it and hope for the best, and that's not nothing, but frankly, if he was so smart, then you'd think he'd have a solution. Nonetheless, it seems like everything Stallman actually proposes is impractical and out-of-touch with the current state-of-affairs, which makes me wonder what we're supposed to do other than marvel at the fact that someone was able to reason that this would become an issue some day.
In this case I don't think it would really matter. An MIT-licensed driver/app to run this scanner would be perfectly acceptable and useful to the community.
Copyleft just ensures that other parties can't take someone's open-source work and close it up and profit off of it. I don't think copyleft is really necessary for when a corporation open-sources some software essential to the use of one of their products.
The reason is probably that there are way way less devices in use now. Graphics cards are made by 2+1 companies (nvidia, amd and somewhat intel), webcams are uvc compatible, mice and keyboards are "driverless", there are less and less printers and scanners in use, and even those are wireless, noone uses digital cameras with need for gphoto2 anymore (either mobile phones or pro equipment and a card reader), flash drivers are all the same, soundcards are mostly onboard now, screens are all the same, etc... so no critical mass to bother companies anymore.
Back in the time you had a bunch of soundcars, palm pilots with activesync, lpt printers, usb printers, scsi printers, zip, jazz drives, etc.
The only pain currently are unstable wifi drivers, because there are still quite a few chipset makers) and well.. custom equipment used by a handful of people, such as op here.
There are less external devices you’d plug in, but internal hardware is still a huge battle. Internal laptop webcams usually don’t work on day one, finger print scanners usually don’t work, suspend and sleep is big problem for most new laptops.
Linux is even worse at maintaining compatibility with old drivers. Opening your source doesn't help when the OS doesn't offer a stable API; you can try to get your driver into the kernel tree, but that takes forever, has high code quality standards, and can still lead to you getting rejected essentially arbitrarily. So you can publish a driver that will work on linux version xyz, but that requires udev and systemd and glibc and everything else to be of versions that match.
but in linux it does not matter as much, you have the source, it is unpleasant to update it to a different api but not actually that hard, much easier than when all you have is a binary blob that you want to get running on your system.
"Should" or not, the reality is it's rare. Rejecting low-quality hardware drivers tends to mean ending up with no drivers rather than with high-quality ones.
What do you mean by “get”? I have an Epson V700. I have done plenty of side by side comparisons. The sharpness of the grain in the scan is what matters, not merely “getting” the grain. To speak of the resolution of a film exposure is incoherent. Grain != pixels. The V700 is also inconsistent because of the floppy plastic mounts and because they lack a mechanism to really keep the film flat.
If you can see individual grains you are already limited by the film itself, anything beyond that is noise. Sharper grain is just a higher spatial frequency noise.
Because Epson scanners don't actually resolve the grain. Low resolution scans tend to exaggerate grain... which can make you think you're resolving the grain.
Also, photos aren't only about resolution and sharpness. Colour and tone reproduction is also important. And consumer scanners are acceptable, but nowhere near as good as drum scanners. Even minilab scanners (Fuji Frontier, Noritsu) are no match for drum scanners.
It can't. It really can't. Even with betterscanning.com holders you can't beat dedicated film scanners or dlsr scans. Drum-scan obviously being the best. I've never wet mounted on flatbeds so can't comment.
My almost 2decades old Nikon V-ED, which I bought when it came out, still outperforms the most expensive flatbed you can buy now (V850-pro), which I also have.
Again, I will write. The job of an OS is to provide an abstracted shim between HW customers have, and the software they have. When this doesn't happen it is a giant fail. There are untold thousands of devices just like this scanner for which the choices to drop support (or change) the OS driver ABI puts people in a bad position because frequently this low volume specialty hardware costs more than the OS vendors product and a new PC/etc to run it because the HW vendor uses MS/Apple/etc's decisions as a chance to force their customers to upgrade to the latest HW. Frequently when the old stuff continues to work just fine.
So IMHO this is a giant middle finger from the OS vendors to their customers, because (as an OS developer myself) dropping support for these kinds of things are rarely done because its costing developer time to have modules that are mostly untouched for years in the tree, or infrequently provide a small shim from the new driver model to the old one. And even when it turns out to be real effort, the OS vendors show us they can provide very transparent shims as long as it benefits them and not a second longer (ex most recently doing transparent x86 emulation on arm based machines).
Apparently the Apple Silicon macs don't have the capability of running 32bit binaries. Since this software is abandoned, it's not like a longer support window would have given them time to recompile the drivers.
Yes, Apple only cares about allowing supported software to easily keep running. This means they can implement more modern and secure systems than Microsoft is able to achieve. It's also the reason Windows is plagued with 32 bit software currently in development as MS has no way to push for the removal of obsolete features.
Given the vast difference in power, it’d probably have been enough to just cascade all the various layers they already have. Arm > x86 > ppc > 68k
Even if each layer loses, say, 50% performance, the fastest 68k was only 75mhz.
(Although now that I think about it I think the handled the 68k/ppc transistion by just making everyone compile for both and slap em into the same binary. Memory is hazy.. it’s been at least 20 years since I’ve touched a classic mac)
You're confusing the "fat" binary format that allowed developers to ship the same application file for both 68k and PowerPC with the emulator that allowed for compatibility with existing 68k code. Much of the OS and apps remained 68k to the end of classic Mac OS's life.
Intel Macs do, though; until Apple removed it. Worked fine in Mojave, but me (and many others) one day rebooted our Macs to find dozens of programs, libraries and plugins that no longer worked.
Whole thing sorta smells like Apple trying to further assert the dominance of AArch64, what with their insane stake in the ISA and it's eventual dominance. Not a terrible thing inherently, but the push to depreciate 32-bit hardware and software alike was pointless. 32-bit ARM can (and should) live alongside AArch64 as an ISA. From what I understand, the M1 even has secure 32-bit emulation mode built-in, without any OS functionality to leverage it.
Whatever Apple does, it's hard for developers to treat the Mac like a regular development target now. Besides the ISA change, the push for Metal-agnostic GPU pipelines has left developers scrambling. Massive open projects like Blender didn't even have an Apple Silicon strategy until Apple sponsored one. It's a bit of a mess, and kinda tragic that Apple is pointlessly complicating cross-platform development.
For all the hate Microsoft regularly gets, they are pretty amazing about that: I have ancient software that still runs fine on Windows 11 through layers upon layers of compatibility libraries baked into the OS.
I'm sure it's costing them dearly, but there's plenty of niche and industrial hardware like this that lasts much longer than a generation of operating systems, and there's no alternative. Linux pays lip service to compatibility, but ABIs change constantly and it's easy to get stuck in dependency hell. MacOS X is... MacOS X.
Speaking of smelly: My scanner just broke down last week. The powersupply started smelling burned plastic and then gave up the ghost completely. Can't find a powersuppy fitting the special contact. It has been working since windows 95 or so, was now running windows 10 without problems and without having to install special drivers from the producer. Old was a canon, early USB model.
On the other hand, there is little that necessitates this sort of thing to be part of the OS instead of just being separate piece of software like VueScan
This isn't the MS-DOS/PC-DOS era where it was necessary for software to have full and direct access to communicate directly with the (ISA 16-bit bus attached, whatever) hardware and peripherals on a desktop workstation, because that can crash the whole multitasking OS.
In the case of the scanner in the original post here which appears to be from the era of MacOS 10.4, 10.5 and 10.6 , the operating system clearly doesn't let the scanning software talk DIRECTLY to it, there's an abtraction layer of some sort for the SCSI or (Firewire 400?) bus.
This paragraph from the blog post stuck out to me:
> The software that drives these scanners is compiled for 32bit architecture and hasn’t been updated in well over 10 years. Hasselblad refuse to update it to modern architectures, and refuse to open source it to allow others to do that, so you are stuck using an OS that can run it.
It's fine if the scanner company doesn't think it's worth putting in the work to add 64-bit support to their software themselves, but they're the ones artificially making it harder to support indefinitely by keeping the code private despite seemingly having no plans to ever sell anything that uses it again. It seems weird to me to blame OS vendors for operating in bad faith when the people who own the actual code that no longer works aren't even willing to let anyone else attempt to do the work to update it for them.
In business terms, open-sourcing a piece of software they haven't touched for ten years means spending a number of hours on doing that for no obvious gain¹. Legal has to clear this for approval (risks? liability?), someone has to decide if opening up the code won't give away company IP, and someone has to actually find the source code (not a given at all), put a licence on it (going through legal again), upload it, and provide a minimum level of documentation.
Doing nothing on the other hand, costs nothing.
1: There is the positive marketing you gain from such a move, but someone has to quantify if doing this is worth it.
That's fair, but I still don't see why that somehow means that the OS vendor should therefore be expected to keep that code working in perpetuity. I have no love for Apple or MacOS, but I just don't see how they're ultimately the ones who should be blamed for this situation.
> The software that drives these scanners is compiled for 32bit architecture and hasn’t been updated in well over 10 years
Might be worth it to try using Wine to run the windows software if they have windows drivers. Wine should run on mac if that's your poison.
> It seems the Firewire ports on these things start to go bad after about a decade. If that happens you’re looking at a 3,000 Euro repair bill
A soldering iron is cheap, and so are firewire ports on digikey. I doubt they invented their own proprietary firewire connection. Ferrari's use volvo parts, you can probably fix it for like $3.
Does userland driver software, which probably wants to make all sorts of weird syscalls to communicate with hardware, tend to actually work under Wine?
I'll second this. I bought some ridiculous behemoth of a Fujitsu scanner that could do 60ppm 20 years ago when it was manufactured and got it working under SANE in Linux. I reported a bug that they fixed to make it work (it didn't initially) I think because it had really old firmware they'd never seen before.
People have done it with vmware ESXi and passthrough. But you can't do it with say, Virtualbox. So with Firewire, real OS on the metal is usually the answer.
I have so far successfully: used eeprom programmers, flashed a multi of android software, interfaced with a PLC from under wine with no problems or workarounds whatsoever
Could you use a 35mm condenser enlarger to project the image directly onto a scanner bed? You would probably need to disable the light in the scanner, remove the glass, and upgrade the light in the enlarger.
I have a Nikon Coolscan that I keep running. On Windows there's a "compatibility mode" that I think means "pretend you have a 32-bit address space" that I need to use to run the drivers.
Do modern equivalents of these scanners exist, but they're too expensive for the author? Or has the market for such things declined to the point where you just can't buy them anymore? (If so, why? Everything's just end to end digital now, so nothing to scan, I guess?)
Feels like the kind of thing you’d be better off renting than buying. I wonder if there are any facilities you can buy a day pass or something and use expensive gear like this. I know there are for woodworking and related hobby gear.
Ok, so the old Hasselblad scanners are breaking down and his big box store scanner isn't good enough. The unanswered question I have is why he doesn't buy a modern professional scanner. Modern pro scanners blow past his 8000 dpi benchmark at less than a quarter the cost of the old Hasselblad.
For example, the Epson Perfection V850 Pro scanner has an optical resolution of 6400dpi for $1300 while doing the full bed, greatly exceeding the X1's 3200dpi in that mode. If you want to match the film scanning performance, the Plustek OpticFilm 8200i SE film scanner has a resolution of 7200dpi for just $400.
PlusTek and Epson are absolutely nowhere near the quality of discontinued scanners like FlexTight, Nikon CoolScan and Fuji Frontier. Modern scanners claim high DPIs, but the real resolution is far below. What good is scanning at 6400 dpi if the lens only resolves at 3000 dpi. I've scanned multiple tests on an Epson 700 myself, and found that there was no visible difference between scanning at full and half resolution.
I have that scanner you're talking about and a Nikon V ED. The Nikon with 'only' 4000dpi massively out-resolves the V850. It's true resolution is somewhere just over 2000dpi and that's after adjusting the trays height.
Because of that I only use the V850 for 4x5 and 120. It's a waste of time for 35mm even if it's quicker to scan multiple frames.
1) optical resolution is not the same as DPI (which I think you know, but its worth pointing out for those people chatting at the back of the class)
2) the way they film is placed on the scanner makes a huge difference to the quality.
the hassblad scanner effectively sucks the film down onto some sort of curved plate, which means that the optics can assume that the film is going to be at position x +- 0.05mm. This is critical because the closer the sensor is to the film, the more the placement affects the quality.
This means that in most highend cases, the film holder has a greater bearing on quality than the sensor/optics.
I have the advantage that at work I have access to a medium format digital camera. This means that I can take really good macro pictures of film negatives. but because the film isn't flat, its a challenge to accurately capture the entire frame. I need to get a proper film gate that slightly tensions the film to make it straight. then I need to worry about getting the camera at 90 degrees to the film.
I wonder if there is an alternate optical solution using a macro lens. Use an even backlight and a dedicated macro lens to take a digital picture of the negative and postprocess. This is essentially what the Hasselblad scanner seems to be doing.
I spent months scanning about 2500 family negatives and slides on an Epson Perfection V600 photo scanner.
While it's no FlexTight, I am happy with the results, especially because I had no plans to crop.
In hindsight, I wished I had used SilverFast rather than the Epson scanning software. SilverFast offers Multi-Exposure which does two scans for maximum dynamic range and then merges them into one.
Also, the Epson default film holders have no ability to flatten the film strips so I probably ended up with softer images in many cases. I believe there are 3rd party adapters that address this.
I have the same scanner. Would you mind sharing your setup? I've only used mine for flat-bed document scanning, but do have a bunch of old family slides.
Do you have the illuminated back? If so, your cover is maybe 4" (10cm) thick.
I have the v750 and scan medium and large format film with excellent results.
You can buy Viewscan software from https://www.hamrick.com, made by a NASA JPL engineer, if I remember correctly.
You get a perpetual license (I reinstalled after 12 years and the updated software ran on my old license code). The software can pull detail out of even the worst negative. I cannot recommend this highly enough! I do have SilverFast as well, and have tried the Mac-native scanning, but Viewscan is the best.
If you have slides and you haven't kept the original film adapters, you can find them on ebay or even Amazon. They are simple plastic holders, nothing special.
You can usually select a particular film type if you scan color negatives that will automatically color correct for the film quirks. All in all, a very easy process once you get set up.
I'm flabbergasted, yeah, it does seem to have the illuminated back. I never looked far enough to see that the mat covering the inside of the lid was removable. I don't recall film holders at all, but will definitely have to look into getting some.
I bought this thing when we moved, to get set up ASAP for a remote home office, so documents and some photos have always been the purpose. I had taken the slides in a box to a photo place, where they used some setup with a DSLR to get JPEGs, but now I'll need to dig the box back out. Thanks!
The V600 is better than most and has a CCD sensor instead of an inferior CIS, but the limiting factor is the optics. The V700/V750/V800/V850 have proper lenses.
The old-school way to scan film is to lay the film directly on glass with a thin layer of mineral oil. Works terrific for medium and large format. We used to do this for drum scanners.
I also use an Epson Perfection (V750 Photo). These machines produce very good results if you are careful with film positioning. I still haven't figured out a way to scan really old negatives in rolls: despite buying several magic holders, I've yet to find one that can oppose the force of a nearly 80-year old film roll.
I would not recommend third-party software, though. The problem with scanner software is that every developer seems to think that I have unlimited time to tweak the settings for every scan, and that scanning those 5 negatives is my only job for the next month. That might be the case for some people, but trust me, if you're looking at several thousand scans, you do not want to tweak each one individually. You want software that works with you. And so far every third party program I tried did not have this approach.
I wonder what kind of results modern upscaling software could achieve with that less detailed scan. I've seen much more impressive improvements than what it would take in this case, and I reckon the end result could look even better than the Hasselblad scan.
I'm sure film purists would scoff at this, but there's a large market that would prefer cheaper solutions that do this in software. We're already used to these enhancements in modern cameras, so why not scanners?
Those upscalers never looks remotely correct to the trained eye beyond minor scaling increases, even with state of the art (I work in post production).
That's just making up stuff, though. I don't think only "film purists" scoff at this. Anything that even remotely has the chance to be used for archival purposes must not be upscaled that way. Not only do you effectively fill in details that simply aren't there, you also don't even have any foolproof way to know which details are made up by upscaling. Worse, some super fancy upscaling might falsify parts of the image that are still above the threshold of where you would need upscaling.
Imagine a historian going through old photos (potentially ones that were previously deemed uninteresting, even) and is led astray by made up details. Imagine a criminal investigation doing the same.
In short, unless you have very specific purposes in mind, and always be careful to clearly mark upscaled images as such and to keep the originals, upscaling film photos that way is a terrible idea.
Before anyone asks, I do think there is some value to upscaling of old photos and videos, like the popular HD/4k 60fps upscaling of videos from the 1800s: It gives even casual viewers a better sense of what the actual scene, and the life depicted within, might have been at that period. It's easier to accept this as "more real" (even though it's ironically less real).
But I really wish any such upscaling was very clearly marked as such, and pointed out to even casual viewers with disclaimers: Unlike the old choppy, blurry, low-res video, this is not a sample of reality back then anymore, but a speculative reconstruction.
To be fair upscaling with videos isn't necessarily making information up, but depending on the method used may use information from other frames, wich I think isn't the same as speculative reconstruction for photos.
For increasing the frame rate for videos this depends on the model used, but in general most of them don't add any new information, but instead calculate how something looked in-between two frames. https://medium.com/axinc-ai/flavr-a-machine-learning-model-t...
Definitely! I was explicitly talking about things like "AI upscaling". Not simple interpolation (which keeps the image spectrum the same and is effectively just anti-aliasing), for example.
To be honest though, the method you linked does seem to fall in the bucket I am warning against: It does make information up, it does add that made up information. Those frames in between are not simply interpolated, or you would not need an ML model.
Case in point, individual droplets in that example animation may not have existed in reality, or moved the way they did here. But hopefully that's fixed by just dropping the interpolated frames, at least.
It's taking the next step in line with what all digital cameras already do. They all do post-processing and make approximations to what the eye actually sees. Just because upscaling technology is still in its infancy and it's not entirely foolproof yet, is not an indication that it won't become seamless and undetectable in the future.
I agree that such features should be optional and all processing should be clearly marked, but to dismiss the technology altogether is a mistake.
Yep, it really makes me curious what the best DPI you can get on a consumer negative scanner is. It's pretty obvious from that comparison that it's a low-dpi scan upscaled to the same size for the sake of comparing them. But surely there are a few high end models that can do reasonable looking 3200 or 6400 DPI scans?
330 comments
[ 0.31 ms ] story [ 357 ms ] threadSeems a pity.
Personally, I've been scanning my own 35mm negatives for the past few years with both a Full-frame DSLR and an APS-C Mirrorless camera, and get much much better results than I'd been getting via my local film processing lab and their scanning.
Scanning negatives with a modern digital camera still retains a lot of the filmic quality that we've come to expect via shooting film. Obviously you lose the "resolution" and ability to blow those images up once scanned, but you can 100% still see the grain, color reproduction, and analog anomalies that we're used to by shooting film.
I recently got a camera with "pixel shift" technology and using that the files have 96mb and really show off the grain. Plus it's just faster than scanning ever was. I am really very happy with the setup.
What I do for my part is just use an enlarger to which I attach my digital camera with various tubes and adapters to get the distances right, and an enlarger lens. These lenses are infinitely cheaper than "regular" planar lenses and they're just made exactly for the job, and the whole assembly is rigid and aligned so it's easy to get the whole negative right.
https://petapixel.com/2017/02/14/wet-mount-scanning-get-high...
Can you really get 80% of the detail like with the Hasselblad X1 with your phone?
For the wet scanning is it a regular consumer grade scanner or something specific / special? How do the results compare to the X1?
Wet scanning is worthwhile if you have a higher-end flatbed scanner like the Epson V700, 750 and 900(?) because they use higher-quality sensors and real lenses.
I almost bought a drum scanner a decade ago but would have to maintain a Mac Quadra for it to run. For large format film, a wet mount scan on an Epson V750 with VueScan software is great.
I'd like to see some better FOSS options but images are a lot better quality than they were using cheap desktop film scanners in 2005.
If you want film scanned today to individual files per frame, it's still a bunch of money if you can find someone to do it... and good luck getting it done right.
I was considering trying to build one out of an inspection microscope, but gave it up and just had my most-important reels scanned to TIFFs.
justifying the cost of the people that did not give up and made an actual product. "if this shit was easy, everyone would do it!" is something i remind myself all the time
Also the opposing viewpoint to remind yourself of is that hey, SOMEBODY was the first to make one of these things and started a company to do it! I've given up on ideas before starting because of the manufacturing technology required, only to see a company like GoPro come out of nowhere and own a market segment.
It used an 8K line array CCD, and pulled a film frame across by locking the frame in a gate and moving the gate with a servo.
Indeed, 6 figures for one of these machines.
https://www.filmlight.ltd.uk/support/documents/other/legacy_...
As others have mentioned here I am also looking into trying to fauxscan with a DSLR and a macro lens, but I'm also wondering if there is enough of a market/interest in a more DIY/FOSS scanning project.
I've considered trying to build something more standard that could be printed or manufactured, that combines some small pointgrey (CV) cameras and macro lens, a light, and some film rolling mechanism to keep the film tight - the idea being that you could basically "scan" the film as a moving roll, scanning a whole 120mm roll in seconds. Pointgrey cameras also have a great api, so you could actually build software on top of it.
I've long thought about putting time to this, but the intersection of pro film shooters and programmers is so small that I'm not sure it would be worth it. If anyone else is interested in the idea, I'd be happy to collaborate on it!
Through various driver discs and archived files online I was able to compile some resources and threw them up on Archive.org (available on the FlexColor page of my site).
Let me know if you have any questions!
I hate the author.
EDIT: I found a price, it's $10k.
Probably not too hard to reverse engineer with some skills and tools, but well beyond what a random photographer is likely to be able to do. The real shame is the company squatting on the driver source code and refusing to update it. It is supremely frustrating when companies treat the driver software--useless without the hardware for it to run--as top company secrets. The software does nothing without the hardware your company sells, so who cares if someone could "steal" it?
...
> This thing is remarkable. The bigger brother (the X5) even more so. Imacon/Hasselblad had a load of patents on the technology that means no other manufacturer can replicate it. There are other high spec scanning solutions, of course, but none that come close to this form factor.
...
> 15 years passed, at which point Hasselblad discontinued the scanners. The cost of modernising the interfaces was not worth it. In fact, only 7 years passed before Hasselblad effectively discontinued them as that was when they stopped updating the software.
This is the dark side of patents.
A company that has used the patent system to run everyone else out of business despite not being particularly innovative (the author describes a "simple" system for assuring the film is perfectly flat), refuses to keep their product line up-to-date or properly support it, charges a fortune for a service that from the sounds of it doesn't actually fix the problem with the interfaces, which may have been purposefully designed to fail anyway...and an entire market segment just dies.
It's sad that those $3000 bills for repair will probably be going to organizations like museums trying to preserve their collections or make them more accessible. Or be unaffordable to such organizations.
Also the hardware is funny. IIRC they didn't iterate on the actual components much, and the move to FireWire basically just integrated a SCSI->FireWire adapter into the case. I haven't seen the inside of an X1/X5, but I'd be interested if these could be repaired by using a different SCSI adapter of some type.
In particular, the author laments the high cost of scans:
> Here in Switzerland I am looking at a cost of around 30.- CHF a frame minimum if I get a third party to scan them for me.
And worries that you have to have something like the 2007 iMac to run it:
> So you need a dedicated old rig to run the proprietary software to drive the scanner.
But doesn't seem to quite connect 2 and 2 together: Those third party shops all have ancient dedicated rigs that they bought a decade or two ago. The use of that equipment is what your $30 per scan is buying! There are label printers and laser markers and time clocks and press brakes and CNCs and inventory management systems at shops all over the world, running some PLC RTOS from the 90s, or DOS or Windows 98 with PXI cards or printer parallel ports or even completely proprietary logic boards, that are decades old and you can't get parts or service or updates for them.
It's ubiquitous in any less profitable industries that aren't on the cutting edge.
It's hard to maintain and gradually update this equipment. I fear it's going to get harder as IIoT services die off and engineered lifetimes shrink. It's hard to debug a machine built before the Internet, with coffee-stained schematics and only enough IDE hard drive space for 8-character variable names, but I worry it will be harder still to debug a machine built well after the Internet with ubiquitous documentation available only behind a login to a server that's no longer online and enough hard drive space for gigabytes of third-party libraries...
Especially in the photo world there's so much tech being used, but not a lot of people that understand the implementation/engineering of it. Not faulting them, but I can't tell you how many people I know that just work off external drives. No NAS, no backups, just originals spread across fragile external hard drives. Hell, I knew some people that up until ~2011 still stored originals on ZIP disks and never managed to move them to something a bit more stable.
Going back and working on older Macs is an absolute pain. Now there's four separate architectures that may or may not have the right software/driver at varying levels of implementation. It's amazing how often the answer is "just use Windows". Not that it's perfect, but at least you weren't dumped like last week's spaghetti.
I have a Nikon Coolscan V and nothing is wrong with it but the switch, but it’s hardly worth repairing or trying to sell on eBay. Working, it might be worth more than I paid for in c2005.
*summicron being a line of leica lenses. really really nice lenses
haha, i see what you did there
PS: It helps when you use it with 3rd party drivers - VueScan.
The camera is an optical device that along with hasselblad lenses has (many) detectable differences in the pictures taken, along with film choice. Someone shooting with a particular camera aiming for large prints is aiming for high excellence not serviceability.
there are bad cables (shielding etc) but the camera+lens+film is literally a sensor/recording device and not really comparable. Whether you like the thing they're doing is a separate consideration, but the person in the article is not looking for ok or even good results. They want everything the best it can be.
Prints are also extremely forgiving, especially viewed from a normal viewing distance, from personal experience of having made high quality prints of my photos.
But again, you can choose to spend however much you wish to to get the last bit of detail out of things.
That's supported by VueScan[1].
[0] https://www.filmscanner.info/en/NikonSuperCoolscan9000ED.htm...
[1] https://www.hamrick.com/vuescan/nikon_coolscan_9000_ed.html#...
But scanner tech seems to have basically "hit a wall," in the last decade, or so. Not many advancements in the imaging. I think digicams and pure digital images/documents have made it difficult to justify the cost of developing them.
Linux was gaining ground on open source drivers this for a while, but then it seemed like we started to backslide.
Open source works best if more people treat it like a marriage commitment, not the occasional random holiday fling.
The way this is handled in industry is to have 20-year-old computer running Windows XP that is just for a specific product that only has drivers for one version of Windows.
I suppose all that's needed is for a good chunk of Windows and Mac users to stop thinking of open source as only Vegas flings of opportunity, and more like a life partner.
The wine/proton project has had far far more impact than decades of posting on game forums asking for a linux build.
Almost all sounds like exaggeration, Intel&AMD have done a lot, and even ARM vendors have contributed lot of drivers. See for example LWN author stats for employers, plenty of HW vendors there: https://lwn.net/Articles/902854/
It's a meaningful distinction and the fact that open source mindshare won is what effectively killed any momentum that free software had.
Software freedom completely disappeared from the public discourse and it was not fortuitous.
https://en.wikipedia.org/wiki/Source-available_software
I'll let Stallman speak for me:
https://www.gnu.org/philosophy/open-source-misses-the-point....
What Stallman calls Open Source is what the rest of the world calls Source Available. Stallman shows his lack of communication skills again here by endlessly fighting to redefine words for no purpose.
That's absolutely wrong, as clearly explained in the article I linked to.
RMS was around in the 1980s when "open source" was still being defined. So maybe he thinks he can define it however he likes. But to the rest of us it doesn't make sense, like saying "software that has a free software license, like GNU Emacs and Microsoft Word." One of those two doesn't belong in that sentence, and saying that I'm using my own custom definition of "free software" is a pretty shitty way to argue.
When you release software under a license that makes approximately zero demands of the other party, the other party won't respect your software. When you release software under a license that makes even low effort demands of the user, like reciprocation, they will.
The term "free" product already meant something to people, and continues to mean that.
There is generally no opportunity to (as RMS might imagine) say "I'm glad you asked about the name. We mean free as in freedom. Let me assemble my lecture podium and explain..."
Instead, people who hear the term have been hearing the opposite of the intent.
Which leads us back ot the current problem, of people thinking it's all random hookups with open source, rather than meaningful relationships.
But yeah I agree naming doesn't help. I'm biased because in my language libre and gratis are distinct. I'll keep that in mind, even though I'm a bit pessimistic and don't think it will help, I guess it won't hurt either.
I think the foot-dragging by some on the name, after it's been pointed out to them, is kinda symptomatic of their difficulty operating in this environment. And they often end up discrediting action that could be viable, but needs to win buy-in from others.
Open source appeared in the public discourse because it had tangible benefits and people could understand what it meant.
I don't think open source killed the concept of software freedom, I think the concept of software freedom as a concept literally never caught on to begin with.
Honestly, I'm a bit tired of the Stallman worship at this point. A lot of people were pessimistic about the future of software and the internet, and they're not all prophets. He observed what was happening when other people chose to naively ignore it and hope for the best, and that's not nothing, but frankly, if he was so smart, then you'd think he'd have a solution. Nonetheless, it seems like everything Stallman actually proposes is impractical and out-of-touch with the current state-of-affairs, which makes me wonder what we're supposed to do other than marvel at the fact that someone was able to reason that this would become an issue some day.
Go back to Reddit.
Copyleft just ensures that other parties can't take someone's open-source work and close it up and profit off of it. I don't think copyleft is really necessary for when a corporation open-sources some software essential to the use of one of their products.
Back in the time you had a bunch of soundcars, palm pilots with activesync, lpt printers, usb printers, scsi printers, zip, jazz drives, etc.
The only pain currently are unstable wifi drivers, because there are still quite a few chipset makers) and well.. custom equipment used by a handful of people, such as op here.
Also, photos aren't only about resolution and sharpness. Colour and tone reproduction is also important. And consumer scanners are acceptable, but nowhere near as good as drum scanners. Even minilab scanners (Fuji Frontier, Noritsu) are no match for drum scanners.
My almost 2decades old Nikon V-ED, which I bought when it came out, still outperforms the most expensive flatbed you can buy now (V850-pro), which I also have.
If you're interested https://www.filmscanner.info/en/FilmscannerTestberichte.html... educate you without having to buy anything to find out for yourself.
So IMHO this is a giant middle finger from the OS vendors to their customers, because (as an OS developer myself) dropping support for these kinds of things are rarely done because its costing developer time to have modules that are mostly untouched for years in the tree, or infrequently provide a small shim from the new driver model to the old one. And even when it turns out to be real effort, the OS vendors show us they can provide very transparent shims as long as it benefits them and not a second longer (ex most recently doing transparent x86 emulation on arm based machines).
It could be possible, with emulation, to run even 68000 binaries targetting the original Mac's operating system.
... if Apple cared. They very obviously do not.
What's wrong with 32-bit software? If you don't need more than 4GB of virtual memory, you've got half the memory footprint for pointers.
Even if each layer loses, say, 50% performance, the fastest 68k was only 75mhz.
(Although now that I think about it I think the handled the 68k/ppc transistion by just making everyone compile for both and slap em into the same binary. Memory is hazy.. it’s been at least 20 years since I’ve touched a classic mac)
Whole thing sorta smells like Apple trying to further assert the dominance of AArch64, what with their insane stake in the ISA and it's eventual dominance. Not a terrible thing inherently, but the push to depreciate 32-bit hardware and software alike was pointless. 32-bit ARM can (and should) live alongside AArch64 as an ISA. From what I understand, the M1 even has secure 32-bit emulation mode built-in, without any OS functionality to leverage it.
Whatever Apple does, it's hard for developers to treat the Mac like a regular development target now. Besides the ISA change, the push for Metal-agnostic GPU pipelines has left developers scrambling. Massive open projects like Blender didn't even have an Apple Silicon strategy until Apple sponsored one. It's a bit of a mess, and kinda tragic that Apple is pointlessly complicating cross-platform development.
I'm sure it's costing them dearly, but there's plenty of niche and industrial hardware like this that lasts much longer than a generation of operating systems, and there's no alternative. Linux pays lip service to compatibility, but ABIs change constantly and it's easy to get stuck in dependency hell. MacOS X is... MacOS X.
In the case of the scanner in the original post here which appears to be from the era of MacOS 10.4, 10.5 and 10.6 , the operating system clearly doesn't let the scanning software talk DIRECTLY to it, there's an abtraction layer of some sort for the SCSI or (Firewire 400?) bus.
> The software that drives these scanners is compiled for 32bit architecture and hasn’t been updated in well over 10 years. Hasselblad refuse to update it to modern architectures, and refuse to open source it to allow others to do that, so you are stuck using an OS that can run it.
It's fine if the scanner company doesn't think it's worth putting in the work to add 64-bit support to their software themselves, but they're the ones artificially making it harder to support indefinitely by keeping the code private despite seemingly having no plans to ever sell anything that uses it again. It seems weird to me to blame OS vendors for operating in bad faith when the people who own the actual code that no longer works aren't even willing to let anyone else attempt to do the work to update it for them.
Doing nothing on the other hand, costs nothing.
1: There is the positive marketing you gain from such a move, but someone has to quantify if doing this is worth it.
Might be worth it to try using Wine to run the windows software if they have windows drivers. Wine should run on mac if that's your poison.
> It seems the Firewire ports on these things start to go bad after about a decade. If that happens you’re looking at a 3,000 Euro repair bill
A soldering iron is cheap, and so are firewire ports on digikey. I doubt they invented their own proprietary firewire connection. Ferrari's use volvo parts, you can probably fix it for like $3.
It's also work checking out SANE itself and seeing if anyone got the scanner working in Linux natively.
Given the skills you could def DIY repair it for cheaper, the high price is likely including uncertainty of what's broken and some form of warranty.
Also not much advantage to line scans when you can get 150MP backs now for copy work.
See https://www.filmscanner.info/en/FilmscannerTestberichte.html for very detailed tests of scanners, including claimed vs real DPI
Because of that I only use the V850 for 4x5 and 120. It's a waste of time for 35mm even if it's quicker to scan multiple frames.
1) optical resolution is not the same as DPI (which I think you know, but its worth pointing out for those people chatting at the back of the class)
2) the way they film is placed on the scanner makes a huge difference to the quality.
the hassblad scanner effectively sucks the film down onto some sort of curved plate, which means that the optics can assume that the film is going to be at position x +- 0.05mm. This is critical because the closer the sensor is to the film, the more the placement affects the quality.
This means that in most highend cases, the film holder has a greater bearing on quality than the sensor/optics.
I have the advantage that at work I have access to a medium format digital camera. This means that I can take really good macro pictures of film negatives. but because the film isn't flat, its a challenge to accurately capture the entire frame. I need to get a proper film gate that slightly tensions the film to make it straight. then I need to worry about getting the camera at 90 degrees to the film.
While it's no FlexTight, I am happy with the results, especially because I had no plans to crop.
In hindsight, I wished I had used SilverFast rather than the Epson scanning software. SilverFast offers Multi-Exposure which does two scans for maximum dynamic range and then merges them into one.
Also, the Epson default film holders have no ability to flatten the film strips so I probably ended up with softer images in many cases. I believe there are 3rd party adapters that address this.
I have the v750 and scan medium and large format film with excellent results. You can buy Viewscan software from https://www.hamrick.com, made by a NASA JPL engineer, if I remember correctly.
You get a perpetual license (I reinstalled after 12 years and the updated software ran on my old license code). The software can pull detail out of even the worst negative. I cannot recommend this highly enough! I do have SilverFast as well, and have tried the Mac-native scanning, but Viewscan is the best.
If you have slides and you haven't kept the original film adapters, you can find them on ebay or even Amazon. They are simple plastic holders, nothing special.
You can usually select a particular film type if you scan color negatives that will automatically color correct for the film quirks. All in all, a very easy process once you get set up.
Hope this helps!
I bought this thing when we moved, to get set up ASAP for a remote home office, so documents and some photos have always been the purpose. I had taken the slides in a box to a photo place, where they used some setup with a DSLR to get JPEGs, but now I'll need to dig the box back out. Thanks!
https://www.analogfilm.camera/scanning/deciding-when-to-wet-...
I would not recommend third-party software, though. The problem with scanner software is that every developer seems to think that I have unlimited time to tweak the settings for every scan, and that scanning those 5 negatives is my only job for the next month. That might be the case for some people, but trust me, if you're looking at several thousand scans, you do not want to tweak each one individually. You want software that works with you. And so far every third party program I tried did not have this approach.
I'm sure film purists would scoff at this, but there's a large market that would prefer cheaper solutions that do this in software. We're already used to these enhancements in modern cameras, so why not scanners?
Imagine a historian going through old photos (potentially ones that were previously deemed uninteresting, even) and is led astray by made up details. Imagine a criminal investigation doing the same.
In short, unless you have very specific purposes in mind, and always be careful to clearly mark upscaled images as such and to keep the originals, upscaling film photos that way is a terrible idea.
But I really wish any such upscaling was very clearly marked as such, and pointed out to even casual viewers with disclaimers: Unlike the old choppy, blurry, low-res video, this is not a sample of reality back then anymore, but a speculative reconstruction.
To be honest though, the method you linked does seem to fall in the bucket I am warning against: It does make information up, it does add that made up information. Those frames in between are not simply interpolated, or you would not need an ML model.
Case in point, individual droplets in that example animation may not have existed in reality, or moved the way they did here. But hopefully that's fixed by just dropping the interpolated frames, at least.
It's taking the next step in line with what all digital cameras already do. They all do post-processing and make approximations to what the eye actually sees. Just because upscaling technology is still in its infancy and it's not entirely foolproof yet, is not an indication that it won't become seamless and undetectable in the future.
I agree that such features should be optional and all processing should be clearly marked, but to dismiss the technology altogether is a mistake.
You need a lens that's designed to reproduce flat subjects to get something decent: https://www.closeuphotography.com/lens-tests
Maybe worth a 3D printed holder like this.
https://www.thingiverse.com/thing:3645894