Recently I resumed taking pictures with this camera, so I made this video. Sorry for my poor English speech. Twitter: @scanner_camera 1200 dpi full resolution sample : flic.kr/p/2nREwRX Thanks for watching!
Wow! I can't believe this doesn't have more views! You, sir, are amazing, and your project is too. Well done! Can't believe this is 11 years old yet performs so well also!
By the way, don't worry about your English skills! As a native (American) English speaker, I can understand you fine. Even if you do make some mistakes (which is fine and understandable), you get your point across!
Thanks so much as well for the information at the end of the video! 😁 (It looks complicated, but if all the parts together are under $150 (US dollars) I will try to make it!
Hi, I got many views when I made the 1st gen of the camera (One in the video is the 2nd gen). This video is just a recap of what I did, sometimes I brought this to take landscape photos.. It's not an easy project, I don't want to make it again 😅
Well done on making one!! Scanback cameras have been around longer than digital CMOS or CCD cameras. I wish Foveon would mature and develop. I think Sigma sold Foveon, no? Anyway, one of the 1st was Better Light and stil make scanbacks upto 384 megapixel. Sadly,this very cheap tech is packaged and made very expensive. But today that has dropped a lot. For example you can get one of the 137mpixel backs for around $600 to 1800 used. . Much respect for this guy making his own. What is great about that is he can swap lens mounts as he wishes. On 4x5 you can also by making a plate to the mount of choice. 4x5 allows for swings shift and tilt.
Thank you for watching. I might make a video a bit more about the camera soon or later. I just don't have enough time to deal with the slow camera for now.
The most important part of your experiment, is that you found that the scaner is capable of capture (almost) the original dynamic range like our eyes can. I believe that this is the beginning of many things. This is marvelous. Great material.
This project is amazing! I am blown away by the depth of colour and light achieved. I would like to understand how you made this, but I understand that explaining everything is hard also. I see that you have a post on your blog with more detailed photos. If you would be so kind as to take more detailed photos of all of the parts, such as the control board and the lens, I think I could start to figure out what you did. What an amazing project and very beautiful photos!
I would guess the most tricky part is to get the scanner to work torn down to the needed parts and to achieve focus. But in overall... its pretty "easy", yet a genious idea!
@@harrison00xXx focus I bet is the easy part because I'm pretty sure that it's fixed. scanners use a compressing lens to capture wider than the sensor so as long as you know the focusing distance of your lens and make it so that it's landing on the sensor you should be good (obviously who have to be very precise). The initialization step must be the real tricky part.
This is what Phase One did decades ago (PowerPhase FX). Those Toshiba linear CCD is not very expensive, however, the real tricky part is that you need quite a lot electrical engineering knowledge to drive them. That is why most of the DIY projects are only restricted to disassembled scanners.
HDR is a sideeffect of how scaners work. They capture intensity of each color for each pixel as it is, independently of others, while a digicams capture information for all pixels at the same moment. digicams do this to solve other problems (like speed) and the easiest approach is to estimate optimal params, capture everything, then figure out what to with that information. With scaner you obviously dont care for speed (forget about shoiting series of action) but want to have as much detail and fidelity as possible - they are also capturing like 10bit (at least AFAIR) color information 😁
@@Omaryllo well, true common digital cameras use rolling shutter, but that's not exactly the same as scanners. With scanners nobody is concerned about moving subject. while with cameras, action is obviously a thing, so some compromises are made to have fast capture. Then miniaturization to have as much features in-camera as possible also plays a role in what the sensor is designed to do.
Regular digital cameras don't have color sensors, but rather a color mask overlaid on the ccd. They guess what the other colors are that that pixel can't actually see by averaging the neighboring pixels. So in comparison, I guess you could call it 'hdr', simply because the color information is by nature higher resolution.
Incredible work, I'd love to see more about this camera and what went into development
2 года назад
After seeing some of these scanner cameras about 7 years ago I built one and was able to get some images out of it I ran into a lot of issues with initializing and eventually some electrical issues. This one you made looks very nicely done.
Can we have more information on this project please? Would be nice to build such thing by myself someday. Looks like a great tool for the landscape photography.
Amazing concept! What's your approach to focusing the lens? Because, as I can imagine, there's no way to check the focus before making the scan, right?
2:28 I would consider monochrome full spectrum cameras used with RGB filter wheels capable of also capturing true per-pixel color. This is very common in astrophotography. Thank you for sharing your project. It's really neat!
I have seen a few project similar to this before. But yours is surprising compact. But missing the front standard means you are missing a lot of fun with tilt, rise, swing or shift. But that will make it much more difficult as you has to grab some bellows figure out and interface. I do believe that exploring older lenses with massive image circles is a lot of fun. And it's more affordable than getting a Large Sense camera, which I believe uses a CCD sensor that is normally used with a phosphorus screen for medial x-rays.
this is incredible! the photos on your flickr are like works of art I want to know more, please! and your English is great, thank you for making a video in your non-native language, that is so generous of you to do
That's great photos, its ingenious those pictures are really crisp the image produced is just awesome, i am interested to know more about this project please make detailed video of this project.
I'm going to have to look into building one. While a lot of my subjects move a lot, like cats, I also like to do still subjects, like landscapes. I have a photo scanner that stopped working--probably the light tube burned out, so it should work for this.
@@babaarcuszatir Because of the high dynamic range, you'd need an ND filter and have to bracket several stops. Otherwise, with most sensors, you'll end up with blown out pixels and unrecoverable information in the highlights
I’m sorry, but you only need an ND filter if you cannot raise the shutter speed and/or cannot/don’t want to decrease the aperture to reach ETTR. In any other scenarios ND will not help, not a bit.
Congratulations my friend. This is a great achievement. I have been thinking in the same direction also. But haven’t been able to finalise it. Anyways …I think I will try to work on making a video camera out of a scanner next.
I just recognized my home village and the house you somehow used from WikipediaTo explain the linear CCD! That‘s so funny! It‘s the Glockenkelter in Kernen im Remstal, it‘s a very small village 18km outside of stuttgart. This is at the border to the hills/forest, a big portion of my childhood happend around that house. This was such a nice coincidence, thank you! How did you come across the image on Wikipedia/why did you choose this image? Just curious! Also of course: it‘s an amazing camera, keep up the work, I’m looking forward to leaning more from you!
This is incredibly cool. If I was any good at DIY I'd make my own, as some of the photos you showed are absolutely incredible! Any future ideas for this camera?
Awesome Idea and work! Have you ever though about tearing down an old, probably even broken APS-C or full frame DSLR for the mirror mechanism to have something to check focus?
@@fusseldieb yeah that’s what I’m talking about! I’ve got the star adventurer 2i but there are much better mounts out there that would be really steady
How do you manage exposure time? And what is the reallife dynamic rage achieved? For example shooting indoor with a window in frame. Would there be some details outside and inside the room?
There is no exposure control other than aperture. I can change ISO by tricking initialization process but I don't see much benefit compare to photoshop adjustment. I set it to the lowest noise.
Ryan, this is a great project! I would love to find out more about your basic approach on this and basic controller. I think I see a Raspberry Pi in there but not sure. If you have any technical details that you would be willing to share I would love to see them. これは本当にスゲーです!感動しました!
so, youve rigged the scan head to 'scan' across the focused frame of the lens as if it were scanning a document...... Thats pretty friggin neat! I was thinking you were using moving mirrors or something, but ultimately same idea. Awesome stuff, the detail is insane. The palm tree photo looks like you were picking up the detail limit of the lens, there was some chromatic aberration as it zoomed in. But resolving the flaws in the lens its using is pretty impressive. Nice work. Ive been (slowly) looking into using the sensor from an old hp scanner as a spectrometer sensor (since all I need of the diffracted light is the brightness values across a given distance), any tips on cracking the firmware open on these things? My current plan is to just tap into the ICs data rails and more or less bit bang and directly read the registers, but if I could leave it mostly intact and utilize the USB that would be preferred. Also, no worries on the pronunciation. Ive heard much worse, unapologetic, English as second language speakers on youtube. You're decently adept. But thanks for the subtitles so we dont have to second guess a word here or there(Im actually terrible at picking up accents, and I got yours pretty quick, especially with the subs, so Id say you're golden). Insert old 'English is my first language and I can barely speak it, so you're fine' joke here lol. Cool work man, I want to see more of what you can get that rig to resolve!
The main difficulty I see with this, which other people aren't considering, is a complete lack of focus preview. You'd have to take a preview scan, a single frame, in order to check your focus. There is no way to check it live, no viewfinder, etc. So taking an image would require a lot of preparation, unless you've found a workaround for that.
maybe do a fast scan to preview focus and then a slow scan to capture the image. Or use traditional mirror focusing and when you are ready, retract the mirror to scan.
RU-vid's compression made the pictures less appealing, this is soo awesome! You might want to put a link for the raw/original files for us to see it better!
Looks like a fun project for those with the aptitude and skills to do it. I didn't realize that scanners have such a large sensor. It would be a good camera for real estate and architectural photography.
amazing, so all electronics from actual scanner but with some changes to make it camera? do you have to make your own software to capture such large image ?
the math just doesn't work out. 60x45mm area scanned at 1200 dpi is about 6 megapixels worth of data. even if you multiply it 3x to compensate for the difference in fine details compared to a Bayer-filter sensor, it's still just about equivalent of a 20 megapixel digital camera. think about scanning medium format film with the same epson scanner that is used to make this camera. otherwise, great job, looks fun and all, and it can be used for special movement effects.
EPSON scanner's scan area is A4 paper. Actual CCD is about 1/5 length of A4 paper, the scanner uses a lens to shrink the image. So when I say 1200 dpi, this is about the scanning software setting. So the real resolution is 1200dpi x 5= about 6000dpi.
Come back to this video to see there are photo samples, the color is stunningly good. I guess if you pair with some high-end modern lens you can create that is on other level
Not much info here, but lower down in the comments Ryan posted that the scanner used was a Epson Perfection v370. I'm thinking of trying to hack this together for studio photography, in order to pre-visualize 4x5 film photos (in order to save film). For setting the distance to the lens, an old microscope could be used as it has very rigid and excellent gearing for making tiny & repeatable movements.
The resolution can go very high? I'm sure that you know better than I do that this is not an accurate statement (subtitles at 1:35). You are talking not about the resolution but merely about the size of the image in pixels. Pixels themselves don't resolve anything. Resolution is the ability of the system to resolve some distinct object features as separate, so it is measured in angular or linear distance units. Nevertheless, it does not reduce the value of your excellent work. It is excellent!
Ryan! This is a super interesting project. I do have a question for you about the scanner. In similar projects people tend to use a ground glass in front of the scanner, but this destroys the resolution of the camera! But from your camera it looks like a ground glass wouldn't be needed if the focal plane is lined up correctly-- is this true?
Probably the other projects use CIS scanner (If you are seeing monochrome in the result it's CIS). I use CDD scanner and removed lens to shrink down the image.
Wow, the resolution picture it takes is better than I thought it would be. Interesting project. There’s a lot of scientific use for it. From my understanding, the linear CCD sensor is from a paper scanner?
What Epson flat bed scanner has so small linear CCD, this looks more like film scanner CCD ? Also how are you getting 140MP with 1200 DPI on 6x4.5? it is more like 2824x2123 pixels which is about 6Mp ? In order to capture 140Mp on 6x4.5cm frame you need something (roughly) 10000x14000 pixels, which is (roughly) 5600 DPI ! First of all that has to be darn good linear CCD to support up 6400 dpi Also and probably the most important your device linear movement has to be super precise to be able to scan 6400 lines per inch !
Ryan you absolute GENIUS!!!! is this a project your working on that needs backing cos if so open up a Kickstarter man and your bound to get massive donations. Is it possible to buy one ?
Awesome. This looks so fascinating. Would you please make a video or write an instruction so we can also learn how to build the same nice thing? Thank you.
WOW! Great job! I think it would WICKED if one or both of two things were to happen. Firstly, if this can be made to be smaller, allowing for even more storage capacity. Secondly, if a few of these were made, imagine putting right next to each other allowing for the added benefit of multi-angle picture capturing. Speaking of which, imagine combining this camera with a light field camera! the beauty of both worlds! :D
Very impressive results, trying to create a scanning camera myself using a similar scanner sensor. In my attempt i try to directly interface with the sensor PCB. Using a raspberry pi pico i have been able to talk to the ADC on the sensor PCB and retrieve the image data. This should allow for manual gain and integration time settings aswell as full controll on how the stepper motor is driven. As the raspberry pi pico has a slow USB(1.1) interface i am directly writing the image data to a SD card. This seems to work quite well although i have not yet started on the mechanical part of the camera so cant show any photos yet.
That sounds fantastic. I have no idea how to interface with the linear CCD. If I don't need to rely on EPSON firmware then it might be able to do a lot of other things. One big limitation of the camera is that fixed exposure time (shutter speed).
@@raspy00135 Yes that is one of the reasons i tried to avoid the EPSON software. Luckily you don't have directly interface with the CCD as the there is already a ADC(AKM AK8419) on the back of the sensor pcb so the flat ribbon cable only contains digital signals and some power. Which does mean that the image quality should be comparable with the the output from the EPSON software. At the moment the wiring and the code is still a bit of a mess but if it seems to work well enough i will try to publish the code and designs. It seems that a lot of EPSON scanners share the same ADC and pinout on the ribbon cable so it should be quite universal. Perhaps some other configuration for the timings would be required as the actual CCD itself does differ between models. Trying now to get the sensor mounted on a old large format camera, for now i will keep the sensor fixed at the back of the camera and try to rotate the camera on a stage to get a panorama like scan.
@@sbuntinx which epson scanner did you buy? I'm going to try this with a v200 but I'm not sure how to trick the scanner into initializing. If you could publish your code that would be amazing!
@@will3346 I am using a v100 at the moment, also have a v330 that i plan on using later aswell. My code is not going to help with the initialization as i replaced the epson board containing usb and power connector with a board i made myself using a micro controller(raspberry pi pico) that has two ribbon cable connectors to talk to the sensor board. For the initialization using the epson board, i think it checks if the light bridge gets closed which normaly happens when the moving part of the scanner gets moved to the start position. There also might be some check where it tries to capture light at the start position to calibrate or check if the light is working. Like i said in my previous comment, the code and wiring is still a bit of a mess as i just got it working a few days ago. Planning to create a proper pcb and clean/improve to code so that i can publish it. For the moment i am working on the mechanics of the camera.
This idea is came to me a week ago and now I see this!!! I was thinking what if I can just use a scanner without light as a digital back for a large format
Regarding the lightbox area. Which LEDs did you use, just the ones from the scanner? How was it controlled/ turned on and off for the initialization. I am trying to build a scanner camera with the Epson v37, and it seems as this is one of the more complicated details, so I would be really happy, if you could help me.