Reflecting Stereoscope

Half HD Reflecting Stereoscope (aka Mirror Stereoscope) for viewing side by side pairs on computer monitor.  Mirror box widens inter-ocular distance to 10".  Carl Pisaturo, 2012

.

A New Bounce off an Old Idea
My recently completed Half-HD reflecting stereoscope (above) is a 3D viewer which bears an obvious resemblance to its 19th century cousin, the Holmes stereoscope (right).
.
The Holmes uses magnifying lenses to allow the source material, actual photographic stereo pairs about the size of a postcard, to be viewed from a close distance, thus giving a good field of view. This compact approach works well with high resolution material like a good photograph - equivalent to perhaps 300-500 PPI (pixels per inch) . 

.
The Holmes approach starts to break down if the quality of the source material drops lower than around 300 PPI. If you've used a Holmes stereoscope or similar device to view reproductions of old stereocards found in many books, you know what I'm talking about:  the magnification makes half-tone dots visible  (exaggerated example at right).  It starts looking more like modern art than a smooth photographic image.

.
And if some sort of LCD display (e.g a smartphone or computer monitor) were used as the source material for a Holmes viewer, the magnification probably will make the RGB pixels visible (exaggerated example at right).  Not good!
.
I say probably because as of 2012, such displays are well below the requisite 300 PPI, but this may change in the next several years.  
.
As one fast forwards into the digital photography age and thinks about the fundamental issues of building a quality digital stereoscope, 2 key principles are obvious... 
.
The first is image density, essentially how many PPI (pixels per inch) can be placed on a flat surface.  Higher PPI media, like slides, photographic prints and "super-displays-of-the-future", allow physically smaller stereoscopes.  Conversely, lower PPI media, like the typical HD computer monitor, necessitates physically larger stereoscopes because the eyes need to be farther away to see a smooth image.
.
The second key principle is image quality: the total number of pixels in an image.  This is obviously key to the viewing experience: the more pixels the better.  1 megapixel per image is in the "pretty good" range, and 10 megapixels per image is in the "amazing" range.
.
An HD computer monitor is the obvious candidate for a digital stereoscope as of 2012 because of its low cost and high (1920 x 1080 pixel) pixel count.  This allows a pair of 1 megapixel images (960 x 1080 pixels each).  That's pretty good image quality.  The downside is that such a 22" HD monitor has low-ish image density (100 PPI), necessitating about 20" of distance from the screen for the pixels to melt into smooth imagery.  
.
The key challenge in making such a viewer arises from the fact that human eyes are 2.5" apart and the monitor's images are 10" apart.  Unfortunately, we cannot "spread out" our vision beyond parallel.  That's why a mirror arrangement is required. The mirrors make it as if our eyes were 10" apart, and that allows 3D perception of the large images on the monitor. 
.

.

.
For good optical performance, typical mirrors can't be used because the path would be distorted passing through the supporting glass.  Instead, "front surface" mirrors are employed.  These have a very accurate reflective surface on the used side.  CNC machined slots in the polycarbonate holder plates ensure good geometric accuracy of the mirror system and its placement relative to the monitor.
.
Compared to 3D TV
The advent of reasonably priced 3D TVs has dramatically shifted the digital stereo photography viewing landscape, rendering mirror stereoscopes niche items which offer slightly higher image quality.
.
The above reflective stereoscope has effectively the same pixel count as a passive 3D TV, but since it has zero ghosting and requires the viewer to be properly aligned, it presents superior image geometric quality.  The lack of polarizing filters or shutter lenses also gives the reflective stereoscope superior brightness and sharpness.  Furthermore, as higher resolution monitors become affordable, the reflective stereoscope can readily employ them, thus increasing its quality advantages proportionally.
.
The reflective stereoscope's cons include an aspect ratio is roughly square whereas the 3D TV is "widescreen";  only one viewer at a time compared to several with 3D TVs; bulky (though interesting and readable) form;  and fragile mirrors which need to be protected from dust and fingers.       
.

Newspeak: Deregulation Lowers Prices

Ken Lay was once a high flier.  As CEO of Enron, he was paid $42 million per year.  And he was smart too - PhD in economics.  He was one of George W Bush's key contributors.  This gave him and his company what is euphemistically called "access" to the political process.  This is what access actually looks like:
.

.
Despite Lay's insistence that deregulation "drives prices down", we see below that the exact opposite happened.
.

Old News

A New Twist on Photography

.
This is what 1% of my lifetime output looks like.  If one is fortunate enough to get a productive lifespan of 50 years, spending 6 months on a project is 1% of it.  Life is finite... you only get 100 6-month projects, that's it.  Choose them well. 

This purpose of this device can best be described as "scene capture", a photographic recording of a scene, including changes over time, with the highest practical fidelity:  high resolution 3D panoramic time lapse.  More information here.

Another way of thinking about this is that it creates gifts to the future: the ability to accurately see the past.
.

.
I love old photographs, the more detailed and busy, the better.  They are as close as it gets to time travel.  This is the scene outside my front door, 121 years ago: it's the Old West for real - San Francisco 1891. Pretty fascinating.  But the engineer/optimist in me wants better: it sure would be great to be able to zoom in to little details, look left and right, maybe hang out for an hour and see who goes by in what sort of conveyance, who goes into the store and what they come out with, and seeing it in color and 3D would be a big help. 

Of course we can't do anything about the shortcomings of past photographic techniques, but we can create new ones with such features and use them to record scenes of our current environment.  Then, in 10, 100, 1000 years, these photosets will be the gift of frozen time, preserved history.

That's what the "History Preserver" project is all about.  The camera rig pictured above is the first finished part of the project.  Next comes the viewing apparatus, then 2 non-trivial software components, one for file processing, one for playback.  And then of course the actual photographic field work.  It's a daunting project for one person without funding who has also to earn a living... full of sub-projects, sub-sub-projects and subtle gotchas.  But despite occasional doubts and motivation difficulties, I believe it will eventually be worth the effort, especially in 121 years.
.

from the PATENTLY ABSURD files ...

GO PRIMARY: check out an actual patent claim

We hear about patents a lot, but who ever looks at the things?  Well, give it a try.  Warning: you may find it difficult going, because they are practically unreadable.

The claim below is the key piece of recently granted US patent  8,090,532.  The patent title is "Pedestrian Route Production" and it belongs to Microsoft.  The application was submitted to the US patent office in 2007, and after 4 years (!) of mysterious percolation through that bureaucracy, was accepted in 2012.

This wording is now, in effect, a part of United States law.  It, along with 14 additional claims, defines the extent of the protection conferred by the patent.

The patent excludes anyone but Microsoft from making, using, selling or offering for sale the subject matter defined by the claims.  In short, for 20 years, you aren't legally allowed to do this, (if you can understand what THIS is) :

What is claimed is:

1. Computer storage media having embodied thereon computer-useable instructions that, when executed, implement a system, the system comprising: a search component that locates at least one information source, retains pedestrian history from a plurality of pedestrians and addresses of at least one information source that has a history of providing reliable information, identifies low quality information sources that do not provide information used in route generation, and blocks information obtainment for the low quality information sources; a gather component that obtains information related to pedestrian travel including security information, weather information, and terrain information, wherein the gather component obtains the information from the at least one located information source; an artificial intelligence component that makes at least one inference regarding a route based on a previous pedestrian behavior; a filter component that determines, based on the at least one inference, the information that is likely relevant and deletes information that is commonly of little value in part through examination of previously produced routes; an analysis component that determines an importance of the information to a user, estimates how likely the information is to change, and chooses if the user should reach a destination through a pedestrian route and/or through a conventional route; a generation component that obtains the information from the gather component and produces a direction set for use by a pedestrian based at least part upon the obtained information; and a resolution component that resolves a conflict between an information source with a financial interest and an information source without a financial interest and instructs the generation component to produce the direction set based upon the information source that does not have a financial interest in providing the direction set.

This claim is ONE SENTENCE.  It has 282 words.

The understandability of this sentence is a matter of some importance:  the validity of the patent and whether or not a party is infringing upon it hinge upon its meaning.  Millions or billions of dollars may depend upon its meaning in a patent lawsuit.  A potential innovation may lay dormant because of fears that it may fall under this wording. 

To me, this 282 word sentence is confusing - oddly inhuman and way too long.  What do the experts say about sentence length?  According to a guidebook for business and government people writing manuals:

Sentence Length and Structure
Keep your sentences short. Longer sentences demand greater concentration from the reader. If your sentence is longer than 25 or 30 words, have a second look at it. While we were often encouraged in school to write more complex and usually longer sentences to demonstrate our growing mastery of syntax, in business and technical writing, shorter is better. The best way to keep sentences short is by limiting them to expressing one idea. If you have two ideas, use two sentences. 

This claim sentence is TEN TIMES longer than a sentence should be!  What is going on? 

Could such ultra-long sentences and arcane technical language help when dealing with such complex ideas?  Maybe lawyers have invented a BETTER way of communicating,  a way that eliminates ambiguity so that it's crystal clear what the patent really means.  This way, patent disputes are quick and easy to sort out. 

But hold on... in fact, patent disputes are the opposite of quick and easy: they are so long and complex that lawyers refer to patent litigation as "the sport of kings".  Multi-month multi-million dollar courtroom battles between corporations are the norm:  despite armies of top legal talent scrutinizing every word and diagram, they STILL can't agree on what a patent really means.  

The US patent system, in which huge numbers of vague, obvious and over-broad patents are granted, is burdensome to innovators large and small - supremely ironic, since its original intent was to encourage innovation.  

Burdensome because for the next 20 years, hardware and software creators who touch upon navigation functionality must take into account not only ambiguous software method patent 8,090,532 described above, but hundreds (who can even discover how many?) of other patents.  They must avoid stepping on all claims of all in-force patents.  They must not write code which can, in a court of law, be interpreted as being too similar to the methods vaguely described in these patents, or else risk ruinous legal costs. 

It's a tragi-comic, massive burden upon the nation's productive forces, slowing human material progress with a defacto taxation of legal and clerical man-hours where engineering hours should be.  

Don't Do This ...   One of the Thousands of Patented Obvious Software Methods  (5,278,985 Hewlett-Packard, 1994)

In conclusion, consider Polaroid Corp v. Eastman Kodak Co , a patent battle which cost over $100 million and lasted 12 years.  The case, involving 10 patents, was so complex it took the judge, Rya Zobel, 3 years after the trial to write the Memorandum of Decision.  Niether company is doing so well; perhaps the money would have been better spent on R&D.

Making a Certain Kind of Rubber Foot Happen ...

A little design story ... As designs evolve and become more refined, one often wants a component that doesn't quite exist. At this point, the options are

a) ACCEPT an existing component   ( defeat, lameness )

b) MODIFY an existing component   ( compromise, regret )

c) MAKE a component   ( nice, but a lot more work )

An unusual need ...  I wanted to copy a nice feature from an old Swiss tripod (left):  its rubber feet are mounted on threaded rods which are lathed to points at the outer ends.  The feet can be rotated up causing metal points to become exposed at the bottom.  In effect, the tripod can rest on rubber pads or metal points, depending on the situation.  E.G. for best stability in wind, you want the metal points, and on nice wood floors, the rubber is preferred.

Although there are lots of rubber feet in catalogs such as McMaster-Carr, I was unable to find a good solution, so I decided to MAKE.

Machining Hard Rubber - A Dead End ... My first inclination was to simply obtain some "machinable rubber" and with a few lathing and tapping operations be done. I read on the internet (risky) that rubber harder than 90A could be machined and got a piece of this to try. It was messy and frustrating. While barely cuttable and drillable, lathing left a terrible finish and tapping was impossible. I gave up.

The Molding and Casting Solution ... (see photo sequence at right) This simple open-mold project illustrates many aspects of technical molding and casting in general.  The objective was simply to make a rounded cylinder with a cast in thread.

First, a "master" was made out of black Delrin (an easy to work plastic) on the lathe.  The master defines the final casting shape.  The threaded rod below the master ill create a seat in the mold for the core.

A negative of the master is created in silicone rubber (Smooth-Sil 930 from Smooth-on) - this is the MOLD.  Silicone rubber has the outstanding quality of not sticking to stuff, as well as being stretchable.  A housing is required to contain the silicone pour.  Here it is an acrylic pipe on a Delrin base.

A "core" is something that sits temporarily inside a casting, here it is a length of aluminum 1/2-13 threaded rod.  The core acts as a "space holder", resulting in a defined empty space within a casting (the thread).

 The core is sprayed with mold release and inserted into the mold, then the casting pour is done. Here, an 80A hardness urethane (PMC-780 dry from Smooth-on) is used. The casting cures in a few hours if warmed and then can be extracted easily from the silicone mold. The core is a bit more difficult to remove, that's why the flats were made on it.

The castings are just what was needed.   Below they are shown in their final context.