Can a Bundle of Soda Straws be a Camera?

This one is hard to classify. Is it a hack, or is it art? Perhaps it’s both. However you want to classify it, it’s pretty cool to turn a bundle of drinking straws into a camera.

Click for larger image

If you’re looking for the technology here, you won’t find much. There’s no lens, no shutter, and no electronics of any kind in [Mick Farrell] and [Cliff Haynes]’ Straw Camera. This is literally a box full of drinking straws standing on end, with a sheet of photo paper behind it. Each straw sends a spot of light that represents the average hue and luminance of its limited view of the subject directly to the film. The process of making an exposure consists of composing the scene, turning out the lights, loading the camera, and setting off a flash.

The resulting images are defocused but recognizable, like seeing familiar sights through a heavy fog. The straws make a strong texture over the ghostly image of the subject – indeed, the straws are the only thing in focus. The fact that the straws don’t form a perfect honeycomb due to settling and imperfections in the bundles is jarring at first, but as you see the images you get used to the extra texture.

When we first saw this, we wondered about the possibility of putting a simple photosensor at the bottom of each straw to capture similar images digitally. The TCS3200 would be about the right size, but given that there are about 32,000 straws in the bundle, the BOM might get a little out of hand. Still, a scaled down digital straw camera might yield some interesting images.

Thanks to [Stuart Rogers] for the tip.

50 thoughts on “Can a Bundle of Soda Straws be a Camera?

  1. There is an easier way to do that, Aluminium “Honeycomb Core”.

    I think you will get interesting results by dipping some or all of it in ultra black paint too, all, bottom or top third. You’d need to experiment to see what effects it gave you.

    1. Aluminum honeycomb is “easier”, maybe, but from what I’ve seen, it’s about 300 USD per square foot (300 mm x 300 mm) for a piece that’s only about an inch thick (26 mm) and a cell size of 6.3 mm which wouldn’t produce a good image. The straws are probably going to be drastically more cost-effective and produce a sharper image due to their superior length:diameter ratio.

      You can roughly estimate the angular resolution of a device like this by arctan[diameter / length] which for the aluminum honeycomb mentioned above would give no better than about 14 degrees angular resolution. This is bad.

      However, using drinking straws, even this very coarse one I have on hand, from Wendy’s that’s designed for attempting to suck melted Frostys (yeah, just try sucking one fully frozen) is 7 mm x 230 mm, which gives an angular resolution of about 1.7 degrees. Way better. And you could do even better by buying slenderer straws.

      (This can be calculated in Frink by: )

      I use diazo (blueprint) paper and various jury-rigged homemade “cameras” to take pictures, including a pinhole camera (incredibly slow) and a lens jammed into a toilet paper tube that can be focused (by sliding it in another toiler paper tube) to take pictures inside a cardboard box “camera” (maybe 15 minutes exposure). Since the time it takes to expose the diazo paper is a function of the aperture size, a drinking straw camera would be incredibly faster and more efficient, and you could easily make a “camera” that would expose a giant piece of blueprint paper in seconds or minutes. (I use sheets that are 42 inches by 30 inches and they’re dead cheap. Also, all you need to do to “develop” them is expose them to ammonia fumes for a few minutes.) In direct sunlight, diazo paper gets exposed within a few seconds.

      I think I should build a giant soda straw camera for use with blueprint paper. It’s drastically simpler and less expensive than using photo paper for a first attempt.

      1. I think you will find that reading the info on the page I provided the link for would give you a much better idea of what is actually available. The range covered is from 0.6 mm right up to 30 mm. You don’t need aerospace grade material, cheap architectural panels from China will do.

          1. So, can you quote me an actual price on a piece of aluminum honeycomb with these specifications (.6 mm honeycomb, 26 mm thickness) that is the size of my diazo paper: 42″ x 30″ (106 cm x 76 cm)?

            I have a feeling that it costs hundreds of times more than drinking straws for the same resolution.

            Again, the important factor is price/resolution. Anyone can pay anyone to make smaller light pipes or narrower honeycomb channels (down to fiber optic light pipe dimensions), but can you make it cheaper than the drinking straws in the article for the same resolution?

            I could pay thousands of dollars for aluminum honeycomb. But why would I, if I can get the same or better effects from a few dollars of drinking straws?

    2. The paint is the interesting question here, and that could be tested with the cheaper straws, too.

      I’m actually shopping for honeycomb material for another project (catching a fountain spray without splashing), and searching around the usual suspects, I find pre-cut pieces with edge treatment, intended to go over camera flashes and studio lights, to “make the light more directional”. Most of them look like they’re either painted or anodized black, so testing the whole-depth-painted idea might be as simple as ordering an existing product and waiting.

    1. If you could create a circuit that could scan such low voltages, you could scan basic silicon diodes in the mV range to detect light. It’s $221 from digikey to buy 30000 1n4148 silicon diodes. It’s true that they are not ideal light sensors, however, if you can create a circuit that can scan them and read the mV range generated when light hits them (you’d need bright lights), it might actually work. Maybe. I don’t know what wavelengths they react to. You could possibly get color using lots of little filters, but again, not sure what wavelenghts are reactive, or how much losses you’d get. I think you’re only talknig 4-5 mV with a flashlight shining right on the thing… So maybe you’d even need to be able to sense int he microvolt range…

      If there’s a will though…

      1. You can use led’s to detect light too, If you want to go further you can use RGB LED’s to detect colour also, I’m sure if you connect enough together into a matrix you could “take a photo” so long as you have a circuit to detect low mV and perhaps a microcontroller for timing and recording voltages so you can later develop photo on laptop or might be able to to it in the microcontroller (not sure I could though).

    1. Not really. This recieves all of the light that’s travelling in one direction. Lytro-style LFCs capture every ray that goes into the aperture from *any* direction, recording brightness and direction.

      You could write some simple software that would turn a Lytro into this, though.

    1. I almost wonder is a thin sheet of translucent white plastic on the back of this coupled with a DSLR with a remote flash could be used to test the camera to make sure a given straw arrangement worked. I know it would essentially be a camera taking a picture of a camera but for testing purposes it should work.

      1. Lol, now I’m thing of making a straw “lens” for my DSLR with a Fresnel lens immediately behind the straws to focus the image onto the DSLR sensor. Would be utterly ridiculous as a lens but might be fun for artistic shots.

  2. Thank you Stuart for sending this in, and Dan for publishing it. It’s a refreshing diversion from the usual Ardu-berry-8266 fare, and while those projects often hold tips and ideas for small optimizations to existing designs, this one has inspired a whole bunch of readers to consider a whole new way of thinking about imaging. The comments section here is smaller than most but dense with ideas, and that’s simply awesome.

    1. Stars are separated by angular direction, and the system detects light from only single angle. Stars are far enough that if you direct it at some star, all straws will see the same light intensity (maybe excluding atmospheric scintillation). But with rotating the device and/or using Earth rotation, imaging is possible.

  3. Interestingly enough, with a complete lack of normal perspective this camera is perfect for portraits. You can put the camera next to the subject and nobody’s nose is going to end up gigantic.

    1. It would bounce all diffuse light to the sensor.
      The trick is to use the straws as filter for light that does not go straight in to the straws. That’s why they use black straws. But matte black would even be better.

    2. Actually just recently there was such a thing on “Play with junk”:

      I tried to do an approximation of this when I was little, but I could never arrange more than some 10 fibers. This thing has a huge amount of them.

  4. I think the image is “unfocused” not “defocused”, which isn’t a word and should be hyphenated. And implies it is focused at some point in the optical path. Believe your spellchecker.

    And it isn’t an image, as in an optical image. The brain can estimate an image from the array of colored circles. It isn’t even a convolution of light from all the straws since there are boarders and it does not spread beyond the straw. With some space between the ends of the straws and the paper, you can describe the results as a convolution. You could analyze with the mathematical tools of image processing, which has very likely been done for glass fiber bundles used in all kinds of imaging.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s