Lena's Birthday Question: Neatest Optical Hacks
Phil Hobbs
birthday of Lena Sjooblom, the November 1972 Playboy model whose face is
so familiar from so many bad image processing papers. (Full story at
http://ndevilla.free.fr/lena.)
Last year on this august occasion, we talked about Image Processing
Snake Oil. To take a more positive attitude for 2006:
What are the neatest optical hacks you know about?
Cheers,
Phil Hobbs
PS: She doesn't look like that any more.
Dave Schaack
"Phil Hobbs" <pc...@SpamMeSenseless.pergamos.net> wrote in message
news:442ACAF0...@SpamMeSenseless.pergamos.net...
> Yes, folks, that's right! Once again, this Friday (March 31) is the
> birthday of Lena Sjooblom, the November 1972 Playboy model whose face is
> so familiar from so many bad image processing papers.
Are you saying that if a paper includes this image, it is bad?
Phil Hobbs
>>Yes, folks, that's right! Once again, this Friday (March 31) is the
>>birthday of Lena Sjooblom, the November 1972 Playboy model whose face is
>>so familiar from so many bad image processing papers.
>
>
> Are you saying that if a paper includes this image, it is bad?
I wish it were that simple. No, I don't mean that--only that image
processing is not in general the panacea that it's sometimes cracked up
to be--so that a great many image processing papers are in fact bad,
whether Lena and her ostrich feather are in there or not. We talked
about that on last Lena's Birthday, as you can read in Google Groups.
I'm not down on the entire field, but I'm sick of reading about how the
Latest And Greatest Technique can make blurred pictures of ugly people
into sharp pictures of pretty people, or in general how postprocessing
can be used as a substitute for good data. It can't. Like a Crescent
wrench, postprocessing has its place in the toolbox, but it isn't for
everything.
Since it's now officially Friday, and because I still want to have a
conversation about good stuff rather than bad stuff, I'll post a few of
my favourites. We're talking *hacks* here, so I'm leaving out Maxwell's
equations, the laser, and so forth.
1. The achromatic lens. Optical instruments basically wouldn't exist
without it--and it's all done by kluging together the properties of
glass made with and without lead oxide. Having three surfaces instead
of two then made possible better monochromatic aberration correction,
but it's still a hack.
2. Pitch polishing. Who'd have thought that you could sustain
nanometre precision over tens of centimetres using a potter's wheel,
tree sap and rouge?
3. The ruling engine. Long before laser interferometers, mechanical
averaging made possible the construction of grating devices accurate to
a small fraction of a wavelength of light, in the Victorian era no less.
What are some of your favourites?
Cheers,
Phil Hobbs
Richard F.L.R.Snashall
> Dave Schaack wrote:
>
> 1. The achromatic lens. Optical instruments basically wouldn't exist
> without it--and it's all done by kluging together the properties of
> glass made with and without lead oxide.
I don't quite concur. I think the kludge is that it was done with
only approximations of those properties.
Helpful person
of liquid mercury?
Please visit my web site at www.richardfisher.com
Adam Norton
tester when I was 13, and the cool stuff it could do was probably the main
thing that got me hooked on optics.
I have since amazed a couple of clients by implementing these techniques in
tooling when the client had been resigned to spending $30k on an
interferometer. Interferometers have their place, but there are many
alignments and other measurements that can get by with something much
simpler.
Adam Norton
Norton Engineered Optics
Optical design and systems engineering for Silicon Valley and beyond.
http://home.ix.netcom.com/~anorton/
(Remove antispam feature before replying)
Charles Manoras
"Adam Norton" wrote
> My favorites have to be the Foucault and Ronchi tests. I built a Foucault
> tester when I was 13, and the cool stuff it could do was probably the main
> thing that got me hooked on optics.
You are obviously trying to plagiarize my biography, my literary agent
wishes to get in touch with your lawyer and Oprah Winfrey is not amused.
> I have since amazed a couple of clients by implementing these techniques
> in tooling when the client had been resigned to spending $30k on an
> interferometer. Interferometers have their place, but there are many
> alignments and other measurements that can get by with something much
> simpler.
Simpler and just as smart, I absolutely agree.
Once my former employer needed two large mirrors about 50 cm in
diameter.
I managed to obtain them relatively cheaply from an established supplier
of astronomical components.
They arrived with their certifications under the form of Foucault data
(taken at two orthogonal azimuths as I had required).
I had a lot of trouble trying to convince my colleagues that as long as
we were not talking about lambda over some unrealistically fantastic
number this was perfectly adequate and that sending the mirrors to be
retested with an interferometer was counterproductive.
We had one interferometer but it could not have handled easily such
large mirrors and they would have to be sent somewhere else for this
purpose ($$$).
At another former employer my manager was Iwao Peter Adachi who
had stayed for a while with (Vasco) Ronchi at his institute in Florence,
Italy and while there had worked out wrote a diffraction theory of the
Ronchi test via Fourier transforms.
Very neat and somewhat more recent than the late Messrs Chester Moore
Hall and Peter Dollond (achromatic lenses) formerly of London.
BTW a more rigorous theory of the Foucaul test (beyond the simple
geometric interpretation found in many ATM books) is not easy.
The best one was elaborated by Linfoot in the UK in the fifties, the math is
quite daunting and goes beyond the Fourier transformation as in in this
instance you have to deal instead with Hilbert transforms.
See Linfoot's book : Recent [sic] Advances in Optics, Theory of the Foucault
Test.
In the same vein one could also mention the Hartmann test recently improved
by Shack etc., the work done in ophthalmology in particular using the
Hartmann-Shack device is quite amazing.
Used by itself (i.e. unimproved) the Hartmann test is also very cheap and
yields quite impressive results.
.
John Savard
<pc...@SpamMeSenseless.pergamos.net> wrote, in part:
>I wish it were that simple. No, I don't mean that--only that image
>processing is not in general the panacea that it's sometimes cracked up
>to be--so that a great many image processing papers are in fact bad,
>whether Lena and her ostrich feather are in there or not. We talked
>about that on last Lena's Birthday, as you can read in Google Groups.
>I'm not down on the entire field, but I'm sick of reading about how the
>Latest And Greatest Technique can make blurred pictures of ugly people
>into sharp pictures of pretty people, or in general how postprocessing
>can be used as a substitute for good data. It can't. Like a Crescent
>wrench, postprocessing has its place in the toolbox, but it isn't for
>everything.
Also, the famous picture of Lena Sjooblom - spelled Lenna in her
original Playboy appearance so that people would know how it was
pronounced - was part of a famous set of images. One web site shows the
origins of some of the other images; I'm pretty sure I once saw the
originals of the mandrill image in an old issue of National Geographic
(it was one of two side-by-side photos of baboons at the bottom of a
page, I remember, but I don't remember other details).
John Savard
http://www.quadibloc.com/index.html
_________________________________________
Usenet Zone Free Binaries Usenet Server
More than 140,000 groups
Unlimited download
http://www.usenetzone.com to open account
AES
Phil Hobbs <pc...@SpamMeSenseless.pergamos.net> wrote:
>>Yes, folks, that's right! Once again, this Friday (March 31) is the
>>birthday of Lena Sjooblom, the November 1972 Playboy model whose face
is
>>so familiar from so many bad image processing papers.
Not sure what you do and don't consider a "hack" but here are a few:
(Why don't you annotate 'em, yes or no, as to whether they make the
grade)
* Student in my lab who got the CO2 gas supply he needed for an early
home-brew CO2 laser by squirting it from the hallway fire extinguisher
into a plastic laundry bag. (Fire marshall was not happy at next safety
inspection.)
* Student in my lab who triggered the spiral flashlamp in his very
early home-brew ruby laser using a Ford Falcon spark coil.
* CO2 lasers using water-cooled distillation columns bought from
chemistry supply houses as the laser tubes -- and evaporated-gold-coated
eyeglass lenses for the mirrors.
* Revitalizing early Spectra Physics and other He-Ne lasers that had
lost their He due to out-diffusion through the glass envelope by putting
'em in a plastic bag filled with 1 atm of He overnight.
* The very early low-cost internal-mirror He-Ne lasers made by
University Labs in Berkeley using chem lab beakers (with the "100 ml,
200 ml" markings still visible on them) as the bells for the laser
structure. (Marketing manager -- or one of the management positions --
for the firm was Tom Perkins, later co-founder of mega-billion venture
capital firm Kleiner Perkins.)
* Art Schawlow and Ted Haensch's edible dye laser: ". . . The laser
material was colored Knox gelatine, prepared according to the directions
on the package . . .as such, the material was a bit soft for optical
work."
* Gouy demonstrating the 180 degree Gouy phase shift through a focus
experimentally in 1890, at a time when knowledge of wave propagation was
still in very early stages.
* Israeli CO2 gasdynamic laser made by igniting a gasoline mixture
inside a heavy-walled tank (using an ordinary spark plug to do so);
blowing open a heavy spring-loaded exit door; and letting the gas exit
through a supersonic noxxle.
* Russian very high power quasi cw CO2 gasdynamic laser, apparently
used for heat treatment of steel, made by filling an ring of big
(house-sized) heavy-walled tanks with CO2; heating them up to just under
bursting pressure using gas flames (huge Bunsen burners!); and
sequentially valving them, one after another, into a supersonic
gasdynamic laser, while previously used tanks are refilled and re-heated.
* And also
* As a follow-on to rotating mercury mirrors (which I would have cited
also), Roger Angel at UoA making deeply curved, near optically perfect,
3 or 4 meter diameter mirrors by electrically melting several tons of
glass in a backyard swimming pool sized tank supported 8 feet up in the
air on huge rotating megatwatt-capacity electrical contact bearings;
then rotating the tank for days on end while the glass cooled.
* And, in same project, making the mirror lightweight by having
sandstone posts which don't melt forming a honeycomb structure inside
the meter-thick (or thereabouts) mirror blank after the glass melts;
then blasting these out with high-pressure water after the mirror cools.
* Using ink-jet printing technology to fabricate nanotech parts or
printed circuits.
* The guy at Berkeley (??) who floats tiny GaAs tetrahedrons into
pre-etched, properly shaped pits on silicon wafers using water, then
makes 'em operate as EO devices.
* The early fiber optics guy described in Jeff Hecht's excellent book
"City of Light: The Story of Fiber Optics" who drew glass fibers many
tens of meters long by attaching the tip of a softened glass rod to a
metal bolt and firing it down the length of a hallway using a cross bow.
* British researches described in same book who demonstrated how to
fabricate clad fibers using a central nozzle connected to one crucible
and a concentric nozzle connected to another -- with colored and clear
molten sugars as the (much more easily melted) core and cladding
materials.
* Early Bell Labs White cell experiment where they broke the x,y
symmetry of the cell, and thus were able to get many more bounces
stuffed into the cell, by mechncially bending one of the end mirrors.
* Allied Chemical, who found it very difficult to separate the mixture
of rare earths used to add luminosity to Coleman lantern mantels, so
just packaged them all in one powder and marketed it as previously
unknown rare earth "dydimium".
* The old idea of compensating thermal expansion by using a
differential structure where short section of high-expansion material
compensates longer section of lower-expanson material.
* Which leads to whole idea of precision differential screw techniques
using slightly different threads per inch.
Enough for today -- I'll see what occurs to me in the midnight hours
tonight.
Dave Schaack
> I wish it were that simple. No, I don't mean that--only that image
> processing is not in general the panacea that it's sometimes cracked up to
> be--so that a great many image processing papers are in fact bad, whether
> Lena and her ostrich feather are in there or not. We talked about that on
> last Lena's Birthday, as you can read in Google Groups.
Very interesting stuff. Thanks.
> Since it's now officially Friday, and because I still want to have a
> conversation about good stuff rather than bad stuff, I'll post a few of my
> favourites. We're talking *hacks* here, so I'm leaving out Maxwell's
> equations, the laser, and so forth.
Your examples and those by others were also very interesting.
Phil Hobbs
What a great list!
AES wrote:
> * Student in my lab who got the CO2 gas supply he needed for an early
> home-brew CO2 laser by squirting it from the hallway fire extinguisher
> into a plastic laundry bag. (Fire marshall was not happy at next safety
> inspection.)
Man after my own heart. Of course, if you weren't so cheap he could
have asked for $10 to buy a little one at the hardware store. ;) Fire
extinguishers are a useful source of dry ice in emergencies.
> * Student in my lab who triggered the spiral flashlamp in his very
> early home-brew ruby laser using a Ford Falcon spark coil.
I'd definitely hire that guy too. People like that get things done.
> * CO2 lasers using water-cooled distillation columns bought from
> chemistry supply houses as the laser tubes -- and evaporated-gold-coated
> eyeglass lenses for the mirrors.
The condenser tube is an excellent idea, although you'd really want
something a bit better than glass.
There was an outfit in my hometown, Vancouver, calle Vortek. They made
incredibly powerful arc lamps by running a huge flow of water helically
down the *inside* of the tube, held there by the centrifugal force. You
could melt 1/2 inch steel by holding it next to one of those lamps.
> * Revitalizing early Spectra Physics and other He-Ne lasers that had
> lost their He due to out-diffusion through the glass envelope by putting
> 'em in a plastic bag filled with 1 atm of He overnight.
Soda-lime glass, I gather. You can kill a soda-lime glass envelope PMT
in a few hours the same way. An orchid to you folks and an onion for
Spectra.
> * The very early low-cost internal-mirror He-Ne lasers made by
> University Labs in Berkeley using chem lab beakers (with the "100 ml,
> 200 ml" markings still visible on them) as the bells for the laser
> structure. (Marketing manager -- or one of the management positions --
> for the firm was Tom Perkins, later co-founder of mega-billion venture
> capital firm Kleiner Perkins.)
Brr. I bet that wasn't annealed too well. Don't bump it!
>
> * Art Schawlow and Ted Haensch's edible dye laser: ". . . The laser
> material was colored Knox gelatine, prepared according to the directions
> on the package . . .as such, the material was a bit soft for optical
> work."
Ted's other two brilliant hacks were the model train interferometer,
which used a corner cube, and of course the aforementioned holey fibre
comb generator. A very smart guy.
> * Gouy demonstrating the 180 degree Gouy phase shift through a focus
> experimentally in 1890, at a time when knowledge of wave propagation was
> still in very early stages.
How did he do that? Fringe counting with a mirror on a lever?
> * Israeli CO2 gasdynamic laser made by igniting a gasoline mixture
> inside a heavy-walled tank (using an ordinary spark plug to do so);
> blowing open a heavy spring-loaded exit door; and letting the gas exit
> through a supersonic noxxle.
Wow, just like a V1 rocket. It's good that CO2 lasers have lots of gain.
> * Russian very high power quasi cw CO2 gasdynamic laser, apparently
> used for heat treatment of steel, made by filling an ring of big
> (house-sized) heavy-walled tanks with CO2; heating them up to just under
> bursting pressure using gas flames (huge Bunsen burners!); and
> sequentially valving them, one after another, into a supersonic
> gasdynamic laser, while previously used tanks are refilled and re-heated.
Now *THAT's* perverse. And very Soviet. It was somebody like Illyushin
the aircraft designer who said, "You Americans make airplanes like a
fine watch, whereas we make ours like a cheap alarm clock. But when you
knock it off the night table, the alarm clock keeps working."
> * As a follow-on to rotating mercury mirrors (which I would have cited
> also), Roger Angel at UoA making deeply curved, near optically perfect,
> 3 or 4 meter diameter mirrors by electrically melting several tons of
> glass in a backyard swimming pool sized tank supported 8 feet up in the
> air on huge rotating megatwatt-capacity electrical contact bearings;
> then rotating the tank for days on end while the glass cooled.
>
>
> * And, in same project, making the mirror lightweight by having
> sandstone posts which don't melt forming a honeycomb structure inside
> the meter-thick (or thereabouts) mirror blank after the glass melts;
> then blasting these out with high-pressure water after the mirror cools.
It's amazing what you can get done by really believing in Newton's Laws.
My friend Ed Yarmchuk, who invented the self-servowriting hard disk,
did away with interferometrically-controlled spin stands in disk
manufacturing, basically by believing Newton. Accuracy improved by
about 10x and throughput by much larger factors. It happened in the lab
next door to me and I didn't find out till afterwards. :(
> * Using ink-jet printing technology to fabricate nanotech parts or
> printed circuits.
>
> * The guy at Berkeley (??) who floats tiny GaAs tetrahedrons into
> pre-etched, properly shaped pits on silicon wafers using water, then
> makes 'em operate as EO devices.
Self-assembly is pretty cool, but definitely counts as a hack because it
only works statistically.
>
> * The early fiber optics guy described in Jeff Hecht's excellent book
> "City of Light: The Story of Fiber Optics" who drew glass fibers many
> tens of meters long by attaching the tip of a softened glass rod to a
> metal bolt and firing it down the length of a hallway using a cross bow.
A crossbow. Nice--I hope there was nobody in the way. Dropping it down
the stairwell I'd believe. When I was an undergraduate at UBC, we had
an annual egg-dropping contest in the glass-enclosed stairwell of the
physics lab building (four stories tall, and originally intended for a
Foucault pendulum). The winner was the one whose egg got there fastest
while remaining intact upon hitting the solid granite floor. They had a
nice setup with crisscross laser beams at the top and bottom of the
shaft, so they got good timing.
People did things like aerodynamically shaped styrofoam nosecones with
drinking-straw fletching, and it was all very fun. Then some bright
soul wrapped the egg in bubble wrap in a foam Big Mac container,
duct-taped it to a nylon rucksack full of rocks, and heaved it over the
fourth-story railing.
Result:
Bright soul: 1
Glass stairwell: 0.
> * British researches described in same book who demonstrated how to
> fabricate clad fibers using a central nozzle connected to one crucible
> and a concentric nozzle connected to another -- with colored and clear
> molten sugars as the (much more easily melted) core and cladding
> materials.
>
Optical cotton candy. Nice.
>
> * Early Bell Labs White cell experiment where they broke the x,y
> symmetry of the cell, and thus were able to get many more bounces
> stuffed into the cell, by mechncially bending one of the end mirrors.
Could you say more about that? What's a White cell?
> * Allied Chemical, who found it very difficult to separate the mixture
> of rare earths used to add luminosity to Coleman lantern mantels, so
> just packaged them all in one powder and marketed it as previously
> unknown rare earth "dydimium".
Hmm. Didymium is a kind of filter glass used for torch welding
glasses--praseodymium and neodymium, I think. I wonder which came first?
>
> * The old idea of compensating thermal expansion by using a
> differential structure where short section of high-expansion material
> compensates longer section of lower-expanson material.
That's the Huyghens pendulum (or was it Hooke?). The transient response
is sufficiently horrible that this definitely qualifies as a hack,
though we have to cut the guy some slack, considering he improved
timekeeping by something like two orders of magnitude, and was working
in the 17th century! (Another very smart guy.)
>
> * Which leads to whole idea of precision differential screw techniques
> using slightly different threads per inch.
>
Differential screws are especially nice because they're pretty solid
mechanically, if you get the keyways really snug. Victorian mechanical
engineers really were pretty amazing. Mechanics was where the action
was, of course, so you got the very best people, and that's enough to
make any field exciting.
Thanks again for a great list!
Cheers,
Phil Hobbs