Enter.... the DigHole

doitashimash1te

doitashimash1te

Well-known
Local time
9:43 PM
Joined
Oct 8, 2006
Messages
479
Location
Amsterdam NL
View My Gallery
Dear brothers & sisters RFF'ers.

Lately, I've been thinking about pinhole cameras. They are great, challenging yet wonderfully simple machines. So I wondered why there is no such thing as a real digital pinhole. Not a modified bodycap on an existing dslr, but a dedicated digital pinhole camera, designed & built from scratch with just one thing in mind: digital pinhole photography.

I know. Such a camera will not be commercially viable and therefore it is not very likely it will see see the light of day any time soon. That is, if we leave it to the camera manufacturers.

But how difficult would it be to design (& evt. build) one ourselves? Of course, we'll need specialists in many fields, as well as creative thinkers... But with 11,000+ RFF members, I'm sure we have the necessary know-how already in our midst. I'm just curious to see what could happen if we joined forces and tried to realize this DigHole by open collaborative design.

I'll kick off here with what I got so far, which is (no more than) a basic concept.

The DigHole
DIY kit w/ as little parts as possible:
(1) body
(2) aperture/shutter mechanism
(3) electronics (image sensor/main circuit/memory card/battery).

(1) body
Plywood. Six parts. Glued on joints, black painted inside and out, and duck taped on the inside.
The body will probably not turn out to be very thick, as a result of the focal distance of the (affordable) digicam sensors.
Standard threaded hole for tripod mount.
No LCD. You won't see the results until you empty the memory card.
Viewfinder: lo-tech solution, I'm thinking of 2 aligned rectangles on top to view through.
"Busy led". This will be on for as long as the camera is "busy" taking a picture. It will be blinking when the memory card is full.
(Why this "busy led"? Because at really long exposure times, the image processor will need a lot of time to process the image. You will need to know when the camera is ready for the next exposure.)

(2) aperture/shutter mechanism
Very simple aperture/shutter mechanism that triggers electronics.
Consists of two elements: horizontal slider (aperture), and vertical slider (T shutter).
Aperture: many f/stops from f/22 and darker, selectable through a horizontal slider, one pinhole per f/stop.
Shutter: T. As lo-tech as it gets! It is a vertical slider, placed on top of the horizontal aperture slider.
-Shutter Slider Up: pinhole open, sensor and main circuit electronics triggered, "busy led" on, sensor gathers light.
-Shutter Slider Down: pinhole closed, image written to memory card, "busy led" off, electronics shut down.

(3) electronics (image sensor/main circuit/memory card/battery)
(This will be the hardest part)
CMOS or CCD image sensor, connected to a main circuit with preset image parameters in (ep)rom (like: pre-def white balance, low saturation, shut off noise filter, no sharpening, raw processing if possible - if not then highest jpeg quality, lowest ISO: 100 or 50..)
By pulling up the shutter slider and thus opening the (T-) shutter, the "busy led" will be on, selected aperture/pinhole is released, light starts falling on the sensor, and at the same time the main circuit is activated, with the preset image parameters (see above). By pushing down the slider, the pinhole will be covered, the light stops falling on the image sensor, the main circuit will assemble the picture data and send them to the memory card, and "busy led" off.
We will have to get the sensor working together with the shutter slider, memory card, battery, busy led, and our (self designed) main circuit board.

Some questions:

- Any ideas as to what image processor would be best to start with, considering cost, availability, openness to reverse engineering?

- Supposed we would know "enough" of a given sensor -by reverse engineering or otherwise- would it be possible to design the main circuit?

See this as a fun project. Any input welcome.
 
Nice idea, though I doubt it'll be economically feasible. Especially since a pinhole cap doesn't cost much. The better sensors cost thousands. And developing the software, and programming the camera will add too.

Also, sensors aren't very good in handling exposure times of minutes. Then the noise will become intolerable.

But I like the idea. When such dighole becomes available at a low enough price I might be interested. In the meantime... http://www.zeroimage.com/web2003/EntryPage/entryFrameset.htm . 🙂
 
RML said:
Nice idea, though I doubt it'll be economically feasible. Especially since a pinhole cap doesn't cost much. The better sensors cost thousands. And developing the software, and programming the camera will add too.

Also, sensors aren't very good in handling exposure times of minutes. Then the noise will become intolerable.

But I like the idea. When such dighole becomes available at a low enough price I might be interested. In the meantime... http://www.zeroimage.com/web2003/EntryPage/entryFrameset.htm . 🙂
Click to expand...

Yes tell me about it - did you receive yours already?
 
Here are a few thoughts on the pinhole digital-
1) Digital sensors, as mentioned, don't handle extended exposures very well.

2) Sensors are either very small or very expensive. The trouble with a small sensor is that pinhole cameras don't scale down very well; the resolution becomes less and less as the pinhole shrinks due to diffraction effects.

Here is a possible solution that would allow a digital pinhole camera to work a little better: place a p and s digital in a full sized (maybe 8x10 format) pinhole camera, but facing the film plane from the front. In place of film, paint the 'film plane' in flat white and use it as a projection screen. Don't bother with a shutter, just use the p and s's shutter mechanism.
Am I making sense? The resulting file would be reversed left/ right, so a quick 'flip' in photoshop would be necessary for correct image orientation.
 
The trouble with a small sensor is that pinhole cameras don't scale down very well; the resolution becomes less and less as the pinhole shrinks due to diffraction effects.
Click to expand...

At any given focal length, the pinhole has an optimum size - too large and the circles of confusion degrade the image, too small and diffraction will do the same. As far as I know, the sensor size should not affect that limit. The problem may simply be making the pinhole small enough. A pin for sewing would be too big.

The optimum pinhole in mm can be calculated using this:

pinhole diameter = square root of the focal length / 28

The focal length is the pinhole-to-sensor distance.
 
I think Bryce's idea would work best. Just have all the electronics in the box, with the sensor and lens in the front, pointed towards the white screen in the back. That could be done with any old crappy sensor, although if it was built from scratch it would probably be good to have the lens permanently tilted to have the white screen in the focus plane.

Downside, it would be pretty big, although if they used 4x5 that's not too bad.

Building one of those with a Fuji F30 (because of the better-than-average high ISO performance) would be a pretty cool project.
 
Last edited:
Actually, a removable lens digicam (some have gotten pretty cheap) with a 'pinhole body cap' might be an easy way to check the idea out. A RF would be best since there's no mirror and because the flange to sensor plane distance is less than an SLR. So you could get a wide angle of view out of it.
SLRs would be limited to something like a 50mm focal length at minimum.

Finder-
Yikes, your formula calls for a .004" (0.1mm) pinhole for a 35mm equivalent focal length on a 2/3" sensor. That might be just doable, but what if you wanted to go really wide? Sounds like a difficult hole to make.
I'll have to root around for the formulas, but as I understand it the total resolving power of a lens (or pinhole in this case) is limited by the diameter of the hole. So unless I'm mistaken, scaling the design down, i.e same field of view but with shorter focal length, smaller sensor, and smaller iris, would yield less total resolution.
I've got some books laying around that cover optics theory, so later I'll try running some numbers and see what I can come up with.
Regardless what the numbers indicate, an actual test would be much, much more interesting!
Numbers as we pronounce it is wrong, by the way- numbers are things that numb the mind, so aptly named. Right?
 
Another thought- pinhole cameras are usually made to cover extremely wide fields of view. Just like the wide lenses on RF cameras, they take advantage of film's inherent indifference to light striking it at angles well away from perpendicular to the film's surface.
Digital sensors are notoriously insensitive to rays like this; so for a really wide angle of view, vignetting may well become a problem.
 
Bryce said:
Actually, a removable lens digicam (some have gotten pretty cheap) with a 'pinhole body cap' might be an easy way to check the idea out. A RF would be best since there's no mirror and because the flange to sensor plane distance is less than an SLR. So you could get a wide angle of view out of it.
SLRs would be limited to something like a 50mm focal length at minimum.

Finder-
Yikes, your formula calls for a .004" (0.1mm) pinhole for a 35mm equivalent focal length on a 2/3" sensor. That might be just doable, but what if you wanted to go really wide? Sounds like a difficult hole to make.
I'll have to root around for the formulas, but as I understand it the total resolving power of a lens (or pinhole in this case) is limited by the diameter of the hole. So unless I'm mistaken, scaling the design down, i.e same field of view but with shorter focal length, smaller sensor, and smaller iris, would yield less total resolution.
I've got some books laying around that cover optics theory, so later I'll try running some numbers and see what I can come up with.
Regardless what the numbers indicate, an actual test would be much, much more interesting!
Numbers as we pronounce it is wrong, by the way- numbers are things that numb the mind, so aptly named. Right?
Click to expand...

A pinhole and a lens aperture is not quite the same thing; the pinhole is proportional to wavelength. But the size of the hole is not neccessarily a problem - this is true with lenses that the aperture is relative to focal length - f/11 would have different physical diameters in different focal length lenses, but resolving power based on diffraction would be equal.

But you a right that you could not go to an infinitely small hole with and infinitely small focal length. Where the cut off is, I am not sure. But since wavelength is measured in nanometers, I would imagine it can get very small - I think the hard thing would be able to make the hole.
 
For some good info and a design program check out:
http://www.pinhole.cz/en/pinholedesigner/.
I do like the idea of a small digi inside a big pinhole camera but I also think that getting very wide angle focal lengths will be difficult. Partly because of the small sensor size. Truly small pinholes are available from Lenox Laser(for example). Link:http://www.lenoxlaser.com/opticalproducts/optical_apertures.html.
I think the main limiting factor for wide angle "lenses" is the acceptance angle required by most digi sensors. Normal and tele "lenses" shouldn't have this problem, though.
Certainly a project worth pursuing, I think!
Rob
 
Larger is Better

Larger is Better

Finder said:
At any given focal length, the pinhole has an optimum size - too large and the circles of confusion degrade the image, too small and diffraction will do the same. As far as I know, the sensor size should not affect that limit. The problem may simply be making the pinhole small enough. A pin for sewing would be too big.

The optimum pinhole in mm can be calculated using this:

pinhole diameter = square root of the focal length / 28

The focal length is the pinhole-to-sensor distance.
Click to expand...

True. But if you plot a graph of optimal pinhole size verses projection length in terms of the resulting f/ratio, you'll find that for the same angle of view, the optimized smaller format cameras will have much lower f/ratios than that of optimized larger formats (i.e. since the angle of view is the same, a shorter projection distance camera equates to a smaller format size).

This means that when optimized for their projection distance, larger formats yield sharper images overall with pinhole than do smaller formats. Of course, those larger f/ratios result in longer exposure times - you can't get something for nothing; it takes more time to collect the greater amount of information present in the light path of the larger format camera.

One reason may be that there is a baseline level of diffraction evident with all pinhole apertures. This diffraction blurr is larger relative to the amount of detail present in the smaller format image.
 
Bryce said:
Another thought- pinhole cameras are usually made to cover extremely wide fields of view. Just like the wide lenses on RF cameras, they take advantage of film's inherent indifference to light striking it at angles well away from perpendicular to the film's surface.
Digital sensors are notoriously insensitive to rays like this; so for a really wide angle of view, vignetting may well become a problem.
Click to expand...

Most of my best pinhole images came from cameras with longer focal lengths (projection distance about the same as diagonal of film format).

Keeping in mind that many people who shoot silver media in pinhole cameras are using paper negatives, the light falloff caused by wide angles of view tends to exacerbate the already difficult issues with tonal range and contrast from orthochromatic, blue-sensitive media in daylight conditions. Light falls off with a cosine^4 function with pinhole apertures. Also, there are noticable issues of image blurr when light strikes film or paper negatives at a severe off-axis angle. It causes a loss of resolution near the edges and corners of wide angle negatives.

The reason many of my cameras have longer focal lengths is that they use a falling plate design, where film boards are preloaded in the camera with film or paper negatives; the camera's projection distance has to be longer than the height of the film board, in order for it to fall properly. This design, although permitting carrying multiple LF negatives out in the field in a sealed box design, precludes the use of wide angles of view. But the images are sharper, corner-to-corner than wide angle designs, and the exposure is much more even from center to edge.

I've also found that these pinhole cameras with more 'normal' projection distances, if used in landscape applications, are much less dependant on foreground objects for acquiring proper composition in the final image. Too, objects in wider angle cameras tend to recede much quicker in size with just a small increase in camera distance. Thus, for my personal style of pinhole image-making, normal angles of view have far fewer compromises, giving better results overall.
 
You must log in or register to reply here.
Back
Top Bottom