A bit of introduction

One thing that never ceases to surprise me with the rise of digital media and the unprecedented availability of information is how easy things have become. Good quality cameras, for instance, that can put early professional DSLRs to shame are now embedded in our phones. 

Processing and editing are available at the touch of a finger, we have infinite amounts of presets, filters, algorythm based processes, and there is certainly a lot more to come, with quality increasing, automation taking over and all those exciting new features we read about on our favorite websites ans blogs.


At the same time, however, I am surprised by the flipside of this technological rush which has, in a way, flattened many different fields of knowledge into a sequence of (more or less) corporate press releases, and reduced passionate users to the role of loyal customers, rather than that of pioneers.


I'm not making a judgment here, these are definitely good times for anyone with an interest in media and creation of content, but to me waiting for the unveiling of some flagship camera by one of the usual brands, or spending weeks getting hyped for the next minor update of Lightroom just does not cut it.


When I think about photography, I think about people exploring new and uncharted territory, figures that are half way between the painter and the alchemist, that are ready to tear down what they have available to make it better or learn something in the process, and, to be honest, I don't often find that kind of spirit very easily.


True, we are seeing a resurgence of old photographic techniques, but that is, again, different. As beautiful and interesting as these old techniques are, they are old techniques. 

What I like to search for is something new, something that gives me the feeling of a step ahead (or sideways) in a field that is quite saturated with corporate standardization.


So imagine my surprise and excitement when I discovered that there were people boldly stepping ahead, from the comfort of integrated systems and customer support to the misty darkness of DIY and experimental photography.

I had grown so accustomed to being a consumer, that the realization that I did have the option to modify, disassemble or create my own gear, processes and workflows was mind-blowing.

And that is what I felt after first reading about scanner camera projects.


I'm not sure how I landed on the first page that got me interested, but I do remember the two resources below were the first I found, and how I spent quite a few evenings thinking how incredibly awesome it would be to actually build my own, maybe improve it, tailor it to my needs and be able to get an image out of something that was a pile of junk before I started working on it.


Here are the first two websites I studied and based my project on:





To me, these guys are something akin to modern-day Daguerres, and pick up on a long tradition of observation and experimentation with light that goes back thousands of years, from ancient China to Aristotle's Greece, with potential roots in prehistoric times.

This rich tradition of ingenuity and creativity, which can be attempted by anyone, resonates with me much more than any industrialized innovation, as revolutionary as it may be.


So once I discovered it could be done, I decided I would do it too, and after doing quite a bit of research, I finally started having somewhat of a clear idea of what I wanted to achieve, and how to do it.





The project

So what is it that sparked my enthusiasm?


To put it very simply, it's the possibility of crating a rather large format digital camera by using a modified flatbed scanner as a sensor.

The moving beam of light you see when you use your average scanner is connected to light sensors, and while there can be different kinds of sensors (CCD or CIS, for example), the general idea is that they act in a similar way to the sensor on any digital camera, converting the impression of light they receive into data that then goes to form an image.


As you may know by comparing prices of full-frame vs.crop sensor cameras, size does matter, and the bigger the sensor, the bigger the price (and the higher the risk of construction defects), so scanners end up providing a very interesting solution for anyone interested in obtaining large digital files without having to go bankrupt.

This means that while the size of the sensor on a scanner is, indeed, much smaller than the sensor on any camera, its ability to move across a relatively large surface enables it to capture a final image that can be, by digital standards, huge!


This means it is possible, in theory, to create a large format digital back that can yield A4 format image. And it gets even better, as the modding of the scanner is nothing too complicated, and can be done fairly easily and without any special tools or skills (I will go into more detail on how to do it in another post).


So why aren't companies, photographers and speculators not flocking over this kind of project like vultures on a carcass?

That might be because of the drawbacks I still did not mention...


There are a number of issues connected to this kind of camera, and they relate to design, functionality and results, and once these issues are known, it starts to make a little more sense that this is still somewhat of an underground project, or material for a novelty article on Petapixel or any other big, photography related media outlet, but little more.


Again, I will go into more detail in future posts, but a few examples of issues have to do with how to connect the scanner to a camera body, how large the projection from the lens to the sensor needs to be, whether or not lenses should be interchangeable, how to implement features that can make your creation into something more than a big toy, such as focusing systems or a viewfinder, how to make the camera transportable, how to power it in the field, how to deal with long exposure times and artifacts, and many other of these big and small problems.


These issues are what makes a project like this so interesting, as it's quite unlikely that it will be taken over by the corporate Moloch and stocked in the isles of your local electronics store with cheap microwaves, crappy stereos and overpriced smartphone accessories.

There is a lot of room for innovation, experimentation and customization in a project like this, and I feel it's falls more in field of craftsmanship than that of industrial design, which is a definite plus in my eyes.



Blueprints, anyone?

Above is a very basic depiction of how the scanner camera is meant to work.

The necessary elements are (starting from the right):

  1. A lens that can project light onto the sensor
  2. A modified scanner that can act as a sensor
  3. A computer that is able to connect to the scanner and process the information it receives (and possibly power the scanner itself)

Obviously, the space between the lens and the sensor needs to be enclosed by a lightproof (ideally) casing, but it theory these three elements are all you need to get started, which is exactly what I a couple of years ago, when I built my first prototype.


My Scanner Camera 1.0

Ingredients for this first build included:

  1. A Canon Lide 110 scanner (CIS sensor)
  2. A large cardboard box
  3. A lens (Auto-Takumar 55mm f1.8 m42 mount, should you be interested)
  4. A laptop

A rough assembly is really not too complicated, as you can see in the photos below.


I mounted the lens to the front of the box by sandwiching some excess cardboard between the lens and an m42 to k-mount adapter, and used some duct tape to block holes and gaps and reduce light leakage.

The U-shaped duct tape pattern at the bottom right of the box is to cover a flap I cut out to easily access the front of the camera body, should I need to move stuff around or change the lens.


I cut out the back of the box, so that I could fit the scanner vertically inside it, and used the scanner cover as a back plate, to keep light out while still having easy access to the scanner and the inside of the camera.


I needed access to the scanner because I attempted to control focusing, or better, to find the correct focal length of the lens, by moving the scanner closer or further from the lens, as needed.


Once the laptop was connected to the scanner I ran preview scans until I had an image I was relatively satisfied with, but, as I was using a freeware version of a scanning software (I don't remember which, anymore), I could not save any of the actual scans, so what I was left with were cropped screenshots of the preview scans, which are still very fun to see and be able to produce, but definitely lack any quality and resolution.


So what were the results like?


First results and considerations

Above are some of the first images I was able to achieve.

As you can see in the first one, some light was definitely coming through the lens and hitting the sensor, but the scanner was obviously not placed correctly, because it was very far from being in focus (or even an actual image...).

The second one shows how framing can be an issue when not using a viewfinder, as you are forced to predict what is going to be framed, then scan to see if you got it right, and if not scan again, and repeat until you get it right.

Also, while there is definitely some kind of embryonic image, it's still quite out of focus.

The third image, finally, has a decent amount of focus, and I am actually in it, posing like a boss.




But there are still some issues, which are quite evident, namely, the fact that the image is so small. As you can see, the circle the lens projected is merely a fraction of the actual scan, and this is due to me using a lens designed for 35mm cameras, meaning that the distance between the lens and the focal plane is just a few centimeters, and the lens is designed to cover a 35mm negative, not a massive A4 scan, but still, the camera does work.


Below are a few cropped versions of some of the screenshots I took, just to see the magic of light in action.

As I mentioned before, these are not the actual scans, but low quality screenshots, so the quality of these crops is definitely quite degraded if compared to what I was looking at on the scanning software.


These pictures amazed me, to be honest. 

Just the fact that I built my own camera, and it worked, was enough to make my day, even if the quality of the photos was lower than that of the old Game Boy toy camera.


These crops, however, give us some insights into some of the technical issues that affect the quality of photos taken with a scanner camera.

First are the horizontal lines, which show the movement of the sensor across the width of the frame. these are very evident in these images because of their low quality, but are definitely present in all images, and act somewhat as the grain would on film.

Like it or hate it, it's something to take into account.

There are also vertical traces, which I assume might be a point in the slide that controls the sensor movement, where the sensor slightly moves, or slightly jumps

The central image, was taken out of the window, and shows another issue to be taken into account, which is the lack of shutter speed control.

Since the image is not a single frame being captured through a shutter, there is no easy way to control exposure, so intense light variations will cause intense variations in the overall exposure of the image.



While limitations in the optics I had available and in the software I was using did not help the quality of the images I took, I have to say that the amount of things I learned from setting this camera up, and the satisfaction that resulted from actually seeing images created through a device of my making were more than enough to compensate for the low resolution, so I definitely consider this first prototype a success. 






It took me over a year to get back to this project, the main reason being I had other things to do, but also the fact that I kind of got lost in trying to think about how to make the scanner camera a better and more durable tool, and with all the information available online it's quite easy to get lost in small details, and postpone action because there will always be something you think can be done better, or some new idea you hope to be able to implement, and while all these things are important and can contribute to a better result, there does come a point when these ideas need to be put to the test, otherwise they will remain ideas.

So I recently decided to build the box setup again, and try out some other glass elements, mostly taken from a few disassembled lenses, and see if I could get better pictures, before moving to the fine-tuning of the camera itself, and the optimization of the body and other "cosmetic" features.


So here is the new version...



My Scanner Camera 1.2

Ingredients for the second build are:

  1. A Canon Lide 110 scanner (CIS sensor) 
  2. A large cardboard box
  3. A lens (the front element of an old Sigma 100-200mm f4.5 zoom)
  4. A laptop

I thought I would use just the one element, because I was hoping it would project a larger image on the sensor than a small 35mm one, as for the rest, not much changed, if not a few details.



In this case I fixed the scanner to the back of the box, initially, hoping that by moving the lens I would be able to determine the focal length of the lens and obtain sharp images.

This, however, did not prove to be a very good idea, as the lens had limited movement, and the box way way too big for me to obtain a decent image, so I ended up cutting away the back half of the box, making the scanner placing slightly adjustable and using a black trash bag to cover the back and avoid too much light leaking in.


The lens proved to be quite interesting, because by using it in one direction it gave a "normal" field of view, with somewhat of a shallow depth of field, while flipping it made it slightly more of a wide angle, with a more generous focus area.


After trying a bunch of different free software, I initially settled for the Windows 10 scanning app, which is incredibly simple to use, and has a very good user interface, however, after saving the first image correctly, the app stopped saving my scans, and even when I set a different destination folder, or changed settings, I was unable to get an image to save. 

I'm not sure what issue is, and I might look into it, but ultimately I settled for Easy Scan, a freeware software that has a less user-friendly interface, but also a lot more options and tweaks that can be applied to scans.

One of the reasons I had some trouble finding a good, free, scanning software to use, is that I wanted something that would allow me to scan TIFF files, as I hoped the better quality of the file would make it easier to edit the images than if they were saved as JPEG or PDF.

This is a matter of choice and of experimentation. TIFF images are by no means a necessity, just what I decided to go with.

Scan settings are another important part of the process, and while I am still testing things out, another variable which seems to be quite important to the quality of your scans is what resolution you set. Each software will have different ways do alter these settings, and I will post some comparisons of the same scan at different settings in another post.




Results and considerations

The fist two images are screenshots of scans, before I found Easy Scan and was able to save TIFF files, while the third is an actual scan I edited to increase some contrast and to get an idea of what can be dove with some levels adjustment.

By zooming into the three images you should be able to see why the actual scan is much better that a screenshot.


The first consideration that came to mind once the scans were made was that there was some issue with the magnification effect given from the distance between the lens and the sensor, which is why I decided to cut off a large portion from the back of the box, and bring the scanner closer to the lens, hoping I would be able to get the full projected circle onto the sensor, rather than just a magnified image.


Focus, as you can see, if really good, for being something that has not fixed setting and was improvised by moving the subject closer or further from the lens, a few centimeters at a time.

This is obviously not an effective method, nor is it a quick one, so that will be a big part of the improvements I am hoping to make in further versions of the camera.


Still, these scans are definitely a step ahead when compared to the previous ones.


 After deducing the distance between the sensor and the lens I started getting results like these:


The three images above are uncropped JPEG images exported from the original TIFFs. Unfortunately I cannot upload TIFF images to this platform, so there is probably some overall loss of quality.


What you can see, though, if you enlarge them, if that there is definitely quite a bit of detail, and while the projected circle doesn't really cover all the space it ideally would, it's already quite an improvement compared to the scans I got from my first attempt, and the quality of the image if much, much, better.


The first image is taken with the glass element oriented the way it was designed, and has been slightly edited for improved contrast. 

Those weird, distorted, lines you can see in the background, on the left, are actually me moving around. 

These distortions are a peculiarity of scanner photography, and are caused by the fact that each pixel is scanned at a differnt moment from the others, so movement is not recorded as the conventional blur we see in "traditional" photos, but more as a decomposed set of horizontal lines, which can yield very interesting and creative effects. I will be talking about this in more detail in another post.

Depth of field here is very shallow, as you can see, with the body of the camera being sharp, but the front of the lens being already quite blurred.


The second image was obtained by flipping the lens, without really moving the camera, if not to frame the background.

Depth of field here is a little more broad, and the angle is more of a wide one.

Focus here is not as sharp, but that's due to the imprecise and approximate method of focusing I am currently using.


The third is a self-portrait, again with the lens in it's correct positioning (though missing most of its glass elements).


Both the second and third image are unedited.


Now we are finally getting some decent results, and even when enlarged quite a bit, these scans retain a good amount of detail and quality, as can be seen below.


Now, if you enlarge these cropped images, the level of detail that a scanner can achieve become quite evident, especially in the second image, which is the one with the best focus.

Considering the impresses portion of the scan is already less than a third of the original, and these are very close crops, the fact that the images can still maintain detail even when enlarged further is quite amazing, and opens a lot of interesting possibilities and challenges.

Suddenly the scanner camera start to seem less of an interesting experiment and more of a tool I want to use more and better.


But there is more, I tried to edit some of the images, to see how they would hold up once I ran them through my usual workflow.

The images below are cropped and edited versions of other scans taken with the same settings as the ones above.



Unfortunately I am out of focus in the first, but if you look at the background, it's quite sharp, and I'm quite happy with the result.


The second and third are images I am happiest with, so far. The amount of different shades, and how light fall off into darker spots is really full and rich. 

Focus is good, and again, these images can take quite an enlargement before losing serious quality.


Obviously, the horizontal lines tracing the path of the sensor are still present, especially when zooming in, but that can be addressed in post-processing to a decent extent, and, to be honest, I find it to be part of what makes the photo unique.


And while I was having fun, I decided to take my wobbly and not too sturdy rig out by the water, and see what I would be able to photograph outdoors. 

I have to apologize for how boring the subject is in most of these photos, but I was the only model available to spend most of the day in front of a buzzing cardboard box with a laptop over it.


Focusing outdoors was not very easy, and the artificial light created some issues too, as well as some interesting light effects, unique to a scan, which you can see in the first image, but the experiment was most definitely successful, and I will be working on fine-tuning the camera better, and making it easier to setup and bring outdoors.


These images were edited a little, mostly due to heavy overexposure on the left side of the second two images, due to a lamp post right on my left



While my first experiment was more of a tentative approach, to see if I could actually get all the elements to work, version 1.2 definitely produced good images, and helped me get a clear idea of what elements I need to work on in order to improve this camera and make it into something I can use in different environments and with good control over its features.


I hope you enjoyed this trip through my experience with the scanner camera, and I hope you check in again for more detailed posts regarding specific issues and improvements, and if you have any questions, suggestions or thoughts, please get in touch, I will be happy to listen or try and clear any doubts.


As for this project, it could never have taken off without the great information I found on the websites I listed earlier, so I feel a great amount of gratitude for the work others did before me and the time they put into making this information available, which is why I am now trying to spread the knowledge as well, and hopefully get more people to discover how there is a lot that can be done with just some basic information and the will to try stuff out.


I will be making some videos detailing any progress and work I do on my scanner camera, so if videos are your medium of choice, please take a look at my youtube channel and subscribe, all support means a lot!


Thanks for reading!