analog vs. digital
Batman
Perhaps this question is not "well formed", but I'll ask anyway, if only
to see its "stupidity". Well, there are two parts actually:
1. Can an analog computer be programmed?
2. If not, doen't this mean the universe itself is fundamentally digital
in
nature, because we can actually program stuff in it?
Now I'll stand back...way back :-) Thanks,
Rajarshi
Stuart L. Anderson
> Perhaps this question is not "well formed", but I'll ask anyway, if only
> to see its "stupidity". Well, there are two parts actually:
> 1. Can an analog computer be programmed?
Yes. A simple analog computer like a slide rule is mentally programmed by
its operator. The differential analysers of 60 years ago were mechanical
analog computers that were programmed by machining three-dimensional cams.
The vacuum-tube analog computers were programmed by setting potentiometers
(pots).
In the early days of digital computers, many differential equations could
be solved faster on analog computers. Thus there were digital-analog
computers where the digital part would convert the input into pot
settings.
Advanced digital-analog computers had servos on the pots so that the
digital part could set the pots automatically. Life got much simpler when
digital computers got fast enough to solve the whole problem.
I never used such contraptions myself, but I knew people who did.
--Stu
___________________________________________________________________
Stu Anderson stua...@drizzle.com Renton, Washington, USA
Anton Sherwood
> 1. Can an analog computer be programmed?
Real computers are analog approximations to digital ideals.
--
Anton Sherwood -- http://ogre.nu/ (not working just now)
franz heymann
Stuart L. Anderson <stua...@drizzle.com> wrote in message
news:10060646...@yabetcha.sttl.drizzle.com...
There was a mechanical differential analyser in the University where I
was an undergraduate. It took the best part of a day to solve a
linear second order differential equation.
Franz Heymann
Neil W Rickert
>Perhaps this question is not "well formed", but I'll ask anyway, if only
>to see its "stupidity". Well, there are two parts actually:
>1. Can an analog computer be programmed?
Several people have already commented on this (with "yes").
>2. If not, doen't this mean the universe itself is fundamentally digital
>in
> nature, because we can actually program stuff in it?
I wonder why people want the universe to be "fundamentally digital".
As far as I can tell, the universe is not made of numbers, so the
question never arises.
Richard Henry
"franz heymann" <franz....@care4free.net> wrote in message
news:3bf7ba91$0$8511$cc9e...@news.dial.pipex.com...
Through most of the 20th Century, the guns of naval warships were aimed by
analog computers that corrected for things like wind, speed and wave-induced
motion of the ship, etc. In real time.
John C. Polasek
The world that is interesting is analog. We could argue all night, but
just consider a high grade color photograph. NO DIGITAL CAMERA CAN DO
WHAT THAT PHOTOGRAPH CAN.
The digital photograph can easily be presented as a table of several
columns of data and as many lines as there are pixels. To the
untrained, the data would look like a function. But there is no Taylor
series, so you can't take a derivative. In a photograph you can scan
in an infinity of directions; in digital, you can look at the next
pixel and wonder if there's any connection. In other words, there is
absolutely no relationship between one pixel and any one surrounding
it. With computers we try to fake it using Fast Fourier Transforms to
try to blend the dots into a continuum; it' s synthetic.
Secondly the data can just as well be regarded as a one dimensional
sequence, not even 2 dimensional. That's how they are transmitted.
Bear in mind this: the solution of a digital differential equation is
not a solution to the differential equation.
The downside of analog is that the analog world has no good memory.
Digital has a perfect memory, but of what? A sequence of digits.
The analog world has uncertainty and randomness; a digital program has
a boring quality to it. Robots and clones that are programmed are
going to be unworkable.
In digital we're getting away with murder just due to the astounding
efficieny of transistor switches on nanometer wires allowing us carry
megabytes of prior code to do the simplest things.
John C. Polasek
Michael Barr
No need to stand back, no one is going to bite your head off. Yes,
analog computers can be programmed. First let me mention that if you
see pictures of the Eniac, there were cables all over, so cables were
used to program even early digital computers. Now digital computers
are programmed using stored programs. The Babbage engine, which
someone finally built in London a few years ago using, it was claimed,
only technology available to Babbage, was analog computer programmed
using punch cards, but the only programmable analog computer I ever
used was programmed with cables. Also potentiometers, used to set
certain constants. The computer was built in the late 40s and was
designed to solve certainly very specialized systems of simultaneous
ODEs of the form
\dot x_i = f_i(x_1,...,x_n) where f_i was a sum of terms of the form
c_{ij}x_j and c_{ijk}x_jx_k. That is every term was of degree 1 or 2.
In practice, there could be no more than 7 variables and I doubt if
there any f_i with more than three terms. The DEs were first made
into equivalent integral equations and then each variable was assigned
to one of the 7 integrators and the outputs of the integrators were
cabled to the constant multipliers (pots) whose outputs were cabled to
the multipliers and then to the adders whose outputs were fed back to
the integrators. The whole thing operated at 5 kh and that meant that
a solution appeared 5000 times a second. You viewed the output by
hooking the output of an integrator to an oscilloscope which displayed
one variable. You could then twiddle the knobs and see the graphs
change. The machine had hundreds, if not thousands, of vacuum tubes
and they were forever burning out or otherwise going bad and operating
the machine was a continual fight against entropy. I was the main
operator for about 2 years. As soon as Remington-Rand donated Penn an
obsolete Univac I, the lab hired a bunch of programmers and switched
to that. That was a trip too, since the Univac I had only 1000 memory
locations (even if each one could hold 12 bytes, 6 bits each) and the
program that resulted took up 80,000 words about half of which were
overlay instructions to make sure that the right code was in the
machine when it was needed. Those were the days.
Paul Lutus
news:3BF740E8...@batcave.cave...
> Hi,
>
> Perhaps this question is not "well formed", but I'll ask anyway, if only
> to see its "stupidity". Well, there are two parts actually:
>
> 1. Can an analog computer be programmed?
Yes. In fact, they are programmed, daily. A cam controller in a washing
machine is an example of a present-day analog microcomputer. It exists
because it is a cheap sort of computer, cheaper than its digital equivalent.
A hydraulic-servomechanism autopilot is another example of an analog
computer, more sophisticated than the above example. These devices have
sobering responsibilities in marine and aeronautical applications where the
fragility and vulnerability of digital computers makes then unacceptably
unreliable.
>
> 2. If not, doen't this mean the universe itself is fundamentally digital
> in
> nature, because we can actually program stuff in it?
What connection do you think you see between programmability, digital
systems and nature?
--
Paul Lutus
www.arachnoid.com
Tom Potter
In sci.math Batman <bat...@batcave.cave> wrote:
> Perhaps this question is not "well formed", but I'll ask anyway, if only
> to see its "stupidity". Well, there are two parts actually:
> 1. Can an analog computer be programmed?
Yes.
The power of a computer is a function of its' accuracy and frequency product.
The faster a computer can compute to a particular level of accuracy,
the more powerful it is.
Before about 1970, analog computers, which were called "differential analyzers"
were more powerful than digital computers, so they were use to simulate
(Compute) dynamic systems.
Analog computers were based on high gain D.C. amplifiers.
These amplifiers amplified D.C. (Zero frequency)
and frequencies up to about 1000 hertz or so.
The voltage output of a high gain D.C. amplifier
is a function of the input resistance and the feedback resistance.
(Of course, the total output voltage is limited by the power supply voltage,
so computations were scaled with this in mind. )
voltage(out) = voltage(in) * resistance(feedback) / resistance(input)
Voltage(in) would be your independent variable,
and you could multiply by a constant by setting the
feedback resistance by means of a variable resistance.
If you wanted to multiply by a variable, you could have the
variable drive a servo to adjust the value of the feedback resistance.
The real beauty of an analog computer was that a capacitor could be used in
the feedback path to allow one to integrate or differentiate. Analog computers
problems were almost always set up to differentiate, rather than integrate,
as this would eliminate the problem of small input offsets being integrated
over a period of time, and saturating the amplifier.
Trig, log, and other functions were simulated with "diode function generators".
These devices shaped the output vs. the input characteristics,
by back biasing the diodes and adjusting the "breakpoints" where the
diodes went into conduction. An array of diodes could be used to accurately
simulate almost any desired input vs. output function.
Although I covered multiplication by constants and variables,
and functions, I almost forgot to mention that addition and subtraction
was accomplished by simply combining electrical currents
at a "summing junction". In other words, you combined currents
to get a voltage drop across a "summing" resistor.
NASA used analog computers extensively,
and I might point out, that you could use an analog computer
with Newtonian physics to control a flight to the Moon or Mars.
You don't need, extremely precise computers nor theories
to do this. The only advantage of more precise computers and theories,
would be to minimize the number of mid course corrections,
and these are a function of technology (Burn accuracy, gyro accuracy,
fuel consistency, etc.) more so than computation. In other words,
even if you had a perfect theory and a perfect computer,
you would end up making the same number of midcourse corrections
--
Tom Potter http://home.earthlink.net/~tdp
-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 90,000 Newsgroups - 16 Different Servers! =-----
Richard Herring
> The world that is interesting is analog. We could argue all night, but
> just consider a high grade color photograph. NO DIGITAL CAMERA CAN DO
> WHAT THAT PHOTOGRAPH CAN.
Conventional photography, colour or monochrome, is based on a process
which is essentially discrete. When developed, each given grain of
silver halide either does or does not reduce to metallic silver,
depending on its photochemical history. The photograph's "pixels"
are randomly distributed and overlapping, but they are still there.
--
Richard Herring | <richard...@baesystems.com>
John C. Polasek
wrote:
I beg your pardon. A digital photograph is a less than even a 1-
dimensional continuum. It is not even a line. It is a sequence of
dots. To that extent it is not even in the class with a true
photograph.
Most of the pixel-hardware is made lithographically from photographic
film.
The socalled halide elements are in nowise insular like the well
defined pixels. They are random and from most standpoints comprise a 2
dimensional continuum. Given a certain pixel, to get the next bit of
information do you want to move left-right or up-down? With a
photograph, all directions are isotropic.
John C. Polasek
Richard Herring
> On 19 Nov 2001 12:08:34 GMT, r...@gmrc.gecm.com (Richard Herring)
> wrote:
> >John C. Polasek (jpol...@cfl.rr.com) wrote:
> >
> >> The world that is interesting is analog. We could argue all night, but
> >> just consider a high grade color photograph. NO DIGITAL CAMERA CAN DO
> >> WHAT THAT PHOTOGRAPH CAN.
> >
> >Conventional photography, colour or monochrome, is based on a process
> >which is essentially discrete. When developed, each given grain of
> >silver halide either does or does not reduce to metallic silver,
> >depending on its photochemical history. The photograph's "pixels"
> >are randomly distributed and overlapping, but they are still there.
> >
> I beg your pardon. A digital photograph is a less than even a 1-
> dimensional continuum. It is not even a line. It is a sequence of
> dots. To that extent it is not even in the class with a true
> photograph.
It's a regular 2-dimensional array of dots. The photograph is an
irregular 2-dimensional array of dots.
> Most of the pixel-hardware is made lithographically from photographic
> film.
> The socalled halide elements are in nowise insular like the well
> defined pixels. They are random and from most standpoints comprise a 2
> dimensional continuum. Given a certain pixel, to get the next bit of
> information do you want to move left-right or up-down? With a
> photograph, all directions are isotropic.
Sure. But none of that makes conventional photography "analog", nor
does it explain in what respect "no digital camera can do what that
photograph can".
--
Richard Herring | <richard...@baesystems.com>
John C. Polasek
wrote:
How would you store a photograph in a digital computer? It can't be
done. It has first to be cooky-cuttered into your famous dots, each
representable by a binary number.
An ordered collection of binary numbers is not the photograph. Even a
binary number is not a real number. It's really useful only for
ciphering.
The thing is totally granular, as a result of the mandatory dissection
process. A colored photo is as good an analogue of 2 dimensions as
there is and there aren't any for 3 dimensions. Of course a photo is
not content searchable, but then neither is a digital 'photo'. By what
rule of logic would you be impelled to search for the value 1273 or
any other integer? That means any individual pixel has approximately
zero value. It takes FFT over a group of pixels.
Debates like this can get quite sickening so consider this my last
volley.
John C. Polasek
Randy Poe
>
> Hi,
>
> Perhaps this question is not "well formed", but I'll ask anyway, if only
> to see its "stupidity". Well, there are two parts actually:
>
> 1. Can an analog computer be programmed?
Yes.
Proof by existence: We are programmable analog computers.
- Randy
Paul Lutus
news:3bf94f4f.14564244@news-server...
> How would you store a photograph in a digital computer? It can't be
> done. It has first to be cooky-cuttered into your famous dots, each
> representable by a binary number.
>
> An ordered collection of binary numbers is not the photograph. Even a
> binary number is not a real number. It's really useful only for
> ciphering.
People as ignorant as you are should be prevented from owning computers. The
results from the latest crop of digital cameras are simply beautiful. I
wouldn't think of owning a film camera any more. Not to mention the
revolution in astronomy created by CCD cameras. Now the astronomical film
archives are mostly regarded as out-of-date and soon to be updated with
better digital exposures (except that the film archives still have great
historical and research significance).
> Debates like this can get quite sickening so consider this my last
> volley.
Last or not, it is by far your dumbest. If you print a film photograph on
high-quality photo paper, and print a digital picture the same way for a
side-by-side comparison, all you will notice is the digital image has a much
greater contrast ratio and better color rendering.
--
Paul Lutus
www.arachnoid.com
Ioannis
[snip]
> > 1. Can an analog computer be programmed?
>
> Yes.
>
> Proof by existence: We are programmable analog computers.
Ho, ho! But we are NOT! We are digital as well. All our sensory organs
have a "minimum threshold resolution", including eyes (cones & rods),
tactile nerves, hearing (we cannot distinguish between two tones closer
than, say, 1/100th of a semitone) and finally and most importantly, our
brain is a very complex array of 1/0 neurons.
Our brain resolution is precicely that: 1. The very basis of our logic:
1: neuron firing for T and 0, neuron idle for F. Wonder why our math is
based on the naturals?
I've been trying to think of a thought with a resolution less than one,
but have been unsuccessful. Any takers? :*)))))
> - Randy
--
Ioannis
http://users.forthnet.gr/ath/jgal/
_______________________________________________
"The traits of the good scientist: good command
of logic and _excellent_ command of insanity."
Russell Easterly
"Richard Herring" <r...@gmrc.gecm.com> wrote in message
news:9taso2$l57$2...@miranda.gmrc.gecm.com...
But the pixel density is incredible.
I once consulted with a hospital on the practicality
of digitally storing x-rays.
The grain size on x-rays is very small.
Even part of an x-ray would require gigabytes
if we tried to keep the same resolution.
(The law requires the hospital to keep the x-rays on file for 5 years.)
Russell
- 2 many 2 count
Russell Easterly
"Ioannis" <morp...@olympus.mons> wrote in message
news:3BF811...@olympus.mons...
> Randy Poe wrote:
> [snip]
> Our brain resolution is precicely that: 1. The very basis of our logic:
> 1: neuron firing for T and 0, neuron idle for F.
This isn't exactly true.
A lot of things determine when a neuron fires.
Like, how many of its neighbors are firing,
how long has it been firing,
are there any neuro-transmitters present,
etc.
Not only that, but the signal the neuron puts out
can change as well.
I wouldn't say that our brains are digital or analog.
G=EMC^2 Glazier
photo-electric experiment is. Photography is my hobby,and I'm
thinking of getting a digital camera for Christmas.Would like to ask
what is the best one for $350 ? Getting back to
computers in the 80;s computers were put to work how the string theory
could relate to us the universe.It did not do so well.I wonder if they
are still trying? Best regards Herb
Paul Lutus
news:dTdK7.46955$XJ4.27...@news1.sttln1.wa.home.com...
Image compression can solve this problem. Very soon, the cost of the X-ray
films will be greater than the cost of computer archiving. After that, the
imaging will be done without film, completely eliminating a rather costly
part of the X-ray diagnostic system.
--
Paul Lutus
www.arachnoid.com
Ioannis
[snip]
> > Our brain resolution is precicely that: 1. The very basis of our logic:
> > 1: neuron firing for T and 0, neuron idle for F.
>
> This isn't exactly true.
> A lot of things determine when a neuron fires.
> Like, how many of its neighbors are firing,
> how long has it been firing,
> are there any neuro-transmitters present,
> etc.
> Not only that, but the signal the neuron puts out
> can change as well.
Great. So all you are saying is that the resolution then is not 1,
rather, greater than OR equal to 1. No problem with that.
> I wouldn't say that our brains are digital or analog.
Well, "logically" [*] they have to be either or, unless there is an
inbetween state which we are not aware of. If it's either/or though,
they certainly aren't analog, for the reasons given above, in
particular, for the fact that all our senses have minimum resolution
thresholds. And that's a fact in any good psychology book.
[*] quotes on "logically", because even if this is an either/or case,
the argument may, in fact, be quite moot: I am using the very resolution
I spoke of to define the resolution of my brain. Heh, Heh :*)
> Russell
> - 2 many 2 count
John C. Polasek
wrote:
It just so happens my specialty is stereo photography. The results
when using digital images can be quite horrible.
John C. Polasek
>
>
>
John C. Polasek
Xrays can't be focussed, so you find the patient standing in front of
a 12" x 15" piece of film stock. You need a vast CD array to cover the
same territory. Otherwise, use fluoroscopy and focus that image down
to the available CCD arrays.
>
Randy Poe
> Ho, ho! But we are NOT! We are digital as well. All our sensory organs
> have a "minimum threshold resolution", including eyes (cones & rods),
> tactile nerves, hearing (we cannot distinguish between two tones closer
> than, say, 1/100th of a semitone) and finally and most importantly, our
> brain is a very complex array of 1/0 neurons.
A common misconception. The signal consists of a series of
pulses. But the information content is in the pulse rate,
which is analog.
The fact that there is a threshold, that the response is
nonlinear, does not make it digital.
- Randy
John C. Polasek
wrote:
>"Russell Easterly" <logi...@home.com> wrote in message
No image compression will not do that. I pointed out that Xrays cannot
be focussed so you're stuck with 1:1. They should be able to rig a
vertical windshiled wiper array that mechanically scanned just like a
focal plane camera in a fraction of a second. Otherwise, tell me how
this image compression works.
John C. Polasek
>
>
>
Paul Lutus
> It just so happens my specialty is stereo photography. The results
> when using digital images can be quite horrible.
A world-class non sequitur. Some sexual experiences are also horrible. But
this doesn't seem to discourage the practice in general.
You might as well argue that all Internet photographic images are not worth
having, because they are all digital. Or that the numerous computer archives
of great works of art are a waste, because they are also all digital.
You do realize, don't you, that wandering away from the original topic, as
you have done, sends a clear signal that you have no confidence in your own
position?
--
Paul Lutus
www.arachnoid.com
Paul Lutus
> Xrays can't be focussed, so you find the patient standing in front of
> a 12" x 15" piece of film stock. You need a vast CD array to cover the
> same territory. Otherwise, use fluoroscopy and focus that image down
> to the available CCD arrays.
You know, every time you post, it gets worse.
1. X-rays can be focused, not easily, but it can be done.
2. Standing in front of an X-ray film, and standing in front of a sensor
panel, are equivalent situations as far as imaging is concerned. No focusing
is required in either case.
There are all kinds of ways to detect X-rays. A giant CCD would not be the
preferred approach.
--
Paul Lutus
www.arachnoid.com
James Hunter
Randy Poe wrote:
It does make it make digital, given that if there
is anything in the brain that depends on a rate,
that would immediately imply that "scientists"
are at least infinitely more stupid than they originally
thought they were.
Paul Lutus
> >Image compression can solve this problem. Very soon, the cost of the
X-ray
> >films will be greater than the cost of computer archiving. After that,
the
> >imaging will be done without film, completely eliminating a rather costly
> >part of the X-ray diagnostic system.
> >
False. Image compression can easily produce 10:1 compressions in cases where
there is a lot of redundancy, such as an X-ray image with a lot of complete
black surrounding a diagnostically useful patch.
You just don't know anything about this topic, and you should stop trying to
pretend differently.
> I pointed out that Xrays cannot
> be focussed so you're stuck with 1:1.
And you are wrong. X-rays can be focused. Not easy, limited applicability to
the present topic, but they can. 1:1 imaging is still the default approach,
but it isn't the only one, as CAT scanning and related technologies show.
> They should be able to rig a
> vertical windshiled wiper array that mechanically scanned just like a
> focal plane camera in a fraction of a second. Otherwise, tell me how
> this image compression works.
What? I should tell you how image compression works? You've already taken
the position that it doesn't. The ball is in your court to find out that you
are wrong.
There is only one required condition for image compression -- the image must
be digital.
http://www.google.com/search?q=x-ray+%22image+compression%22
--
Paul Lutus
www.arachnoid.com
me...@cars3.uchicago.edu
and quick (digital) readout.
Mati Meron | "When you argue with a fool,
me...@cars.uchicago.edu | chances are he is doing just the same"
John C. Polasek
wrote:
>"John C. Polasek" <jpol...@cfl.rr.com> wrote in message
>news:3bf9795d.25332335@news-server...
>
>> Xrays can't be focussed, so you find the patient standing in front of
>> a 12" x 15" piece of film stock. You need a vast CD array to cover the
>> same territory. Otherwise, use fluoroscopy and focus that image down
>> to the available CCD arrays.
>
>You know, every time you post, it gets worse.
>
>1. X-rays can be focused, not easily, but it can be done.
No it can't practically be done for the big chest Xrays. Bouncing them
at a negligible glancing angle from some suitable substance will not
do the trick.
>
>2. Standing in front of an X-ray film, and standing in front of a sensor
>panel, are equivalent situations as far as imaging is concerned. No focusing
>is required in either case.
Yes of course it's the same. But considering an average CCD is about
1" square, you would need 180 of them to cover the territory and a way
to stitch up the gaps between the chips. Then at 2 megabits per chip
you need to store 360 megabits in a 30th of a second or so. I would
advise a bus 180x16 wide to get the job done. And you'd need more than
a Z80!
John C. Polasek
wrote:
>"John C. Polasek" <jpol...@cfl.rr.com> wrote in message
No there's a second requirement: acquiring the data. Please tell me
how.
John C. Polasek
>
>
Paul Lutus
> >There is only one required condition for image compression -- the image
must
> >be digital.
> No there's a second requirement: acquiring the data. Please tell me
> how.
No, you can find out -- I already know. Look up "CAT scanner" and related
technologies. They go directly from X-rays to digital, computer-readable
data streams.
--
Paul Lutus
www.arachnoid.com
Paul Lutus
> >1. X-rays can be focused, not easily, but it can be done.
> No it can't practically be done for the big chest Xrays. Bouncing them
> at a negligible glancing angle from some suitable substance will not
> do the trick.
I never made that claim. I said you were wrong about focusing X-rays.
> >2. Standing in front of an X-ray film, and standing in front of a sensor
> >panel, are equivalent situations as far as imaging is concerned. No
focusing
> >is required in either case.
> Yes of course it's the same. But considering an average CCD is about
> 1" square, you would need 180 of them to cover the territory and a way
> to stitch up the gaps between the chips. Then at 2 megabits per chip
> you need to store 360 megabits in a 30th of a second or so. I would
> advise a bus 180x16 wide to get the job done. And you'd need more than
> a Z80!
Honestly. This wouldn't be the approach, but, since I already said this,
John C. Polasek
wrote:
>"John C. Polasek" <jpol...@cfl.rr.com> wrote in message
When you say you can focus Xrays, there's nothing to be gained by
focussing them out of the Xray tube.
So you are saying that after the rays leave the spot on the target of
the tube and diverge on their way through the subjects chest, 12"x15",
then you are going to focus the exit rays onto a smaller CCD, right?
(Maybe there's something I am missing about data compression that you
would introduce here). You would need a giant lens. It's not possible
to do it by reflection.
If they could do it they would be doing it.
John C. Polasek
>
Gordon D. Pusch
> So you are saying that after the rays leave the spot on the target of
> the tube and diverge on their way through the subjects chest, 12"x15",
> then you are going to focus the exit rays onto a smaller CCD, right?
[...]
> You would need a giant lens. It's not possible to do it by reflection.
You are clearly ignorant of the field known as ``grazing incidence
X-ray optics,'' which uses curved grazing-incidence mirrors to focus
X-ray beams at synchrotron light-sources all over the world, as well
as in space-based imaging X-ray telescopes such as the one aboard
the Chandra X-ray Observatory Satellite <http://chandra.harvard.edu/>,
<http://imagine.gsfc.nasa.gov/docs/science/how_l1/xray_telescopes.html>.
> If they could do it they would be doing it.
You have confused ``physically impossible'' with ``not yet cost-effective
in a hospital environment.''
-- Gordon D. Pusch
perl -e '$_ = "gdpusch\@NO.xnet.SPAM.com\n"; s/NO\.//; s/SPAM\.//; print;'
Paul Lutus
> When you say you can focus Xrays, there's nothing to be gained by
> focussing them out of the Xray tube.
I am not saying this is feasible, but of course it would confer an
advantage. It would allow magnification, as just one possibility.
>
> So you are saying that after the rays leave the spot on the target of
> the tube and diverge on their way through the subjects chest, 12"x15",
> then you are going to focus the exit rays onto a smaller CCD, right?
That would be one approach, but it isn't needed, There are better ways to
image X-rays.
>
> (Maybe there's something I am missing about data compression that you
> would introduce here). You would need a giant lens. It's not possible
> to do it by reflection.
> If they could do it they would be doing it.
They are doing it. As I said, CAT scanners and related technologies have
ways of producing magnification and sensing X-rays without film. Among other
things.
The data compression issue is about saving storage space. That is a
different topic.
--
Paul Lutus
www.arachnoid.com
me...@cars3.uchicago.edu
>jpol...@cfl.rr.com (John C. Polasek) writes:
>
>> the tube and diverge on their way through the subjects chest, 12"x15",
>> then you are going to focus the exit rays onto a smaller CCD, right?
>> You would need a giant lens. It's not possible to do it by reflection.
>
>X-ray optics,'' which uses curved grazing-incidence mirrors to focus
>X-ray beams at synchrotron light-sources all over the world, as well
>as in space-based imaging X-ray telescopes such as the one aboard
>the Chandra X-ray Observatory Satellite <http://chandra.harvard.edu/>,
><http://imagine.gsfc.nasa.gov/docs/science/how_l1/xray_telescopes.html>.
>
small angular acceptance and very, very long focal lengths. Perfectly
adequate for synchrotron environment, tatally inadequate for medical
imaging.
>
>> If they could do it they would be doing it.
>
>in a hospital environment.''
No, it is not "not cost effective in hospital environment", just "not
*doable* in hospital environment."
me...@cars3.uchicago.edu
>"John C. Polasek" <jpol...@cfl.rr.com> wrote in message
>
>> When you say you can focus Xrays, there's nothing to be gained by
>> focussing them out of the Xray tube.
>
>advantage. It would allow magnification, as just one possibility.
>
acceptance. Works just fine with a small emittance source, such as a
synchrotron, but not with an x-ray tube.
>>
>> So you are saying that after the rays leave the spot on the target of
>> the tube and diverge on their way through the subjects chest, 12"x15",
>> then you are going to focus the exit rays onto a smaller CCD, right?
>
>That would be one approach, but it isn't needed, There are better ways to
>image X-rays.
Image plates. Routine stuff.
Richard Herring
> >> wrote:
> >
> >> >John C. Polasek (jpol...@cfl.rr.com) wrote:
> >> >
> >> >> The world that is interesting is analog. We could argue all night, but
> >> >> just consider a high grade color photograph. NO DIGITAL CAMERA CAN DO
> >> >> WHAT THAT PHOTOGRAPH CAN.
...]
> >
> >> The socalled halide elements are in nowise insular like the well
> >> defined pixels. They are random and from most standpoints comprise a 2
> >> dimensional continuum. Given a certain pixel, to get the next bit of
> >> information do you want to move left-right or up-down? With a
> >> photograph, all directions are isotropic.
> >
> >Sure. But none of that makes conventional photography "analog", nor
> >does it explain in what respect "no digital camera can do what that
> >photograph can".
> How would you store a photograph in a digital computer? It can't be
> done. It has first to be cooky-cuttered into your famous dots, each
> representable by a binary number.
Sure. What's the problem? Analysis 101 will explain to you that
you can perform that process to any degree of precision required.
At the other end of the process, your eyes are doing the same thing. Each
discrete rod and cone on your retina either fires or does not fire.
> An ordered collection of binary numbers is not the photograph.
It's sufficient information to make an indistinguishable reproduction.
> Even a binary number is not a real number.
True, and so what?
> It's really useful only for ciphering.
It can approximate the real number to any desired degree of
accuracy.
> The thing is totally granular, as a result of the mandatory dissection
> process. A colored photo is as good an analogue of 2 dimensions as
> there is
Look closer. It's totally granular, as a result of the mandatory
discrete development process.
> and there aren't any for 3 dimensions. Of course a photo is
> not content searchable, but then neither is a digital 'photo'. By what
> rule of logic would you be impelled to search for the value 1273 or
> any other integer? That means any individual pixel has approximately
> zero value. It takes FFT over a group of pixels.
I don't see the relevance of this. I don't want to search for 1273,
I want to know what it is you think a digital camera cannot do that
a photograph can.
> Debates like this can get quite sickening so consider this my last
> volley.
--
Richard Herring | <richard...@baesystems.com>
franz heymann
[Snip]
>
> The thing is totally granular, as a result of the mandatory
dissection
> process. A colored photo is as good an analogue of 2 dimensions as
> there is and there aren't any for 3 dimensions.
[Snip]
Photographic emulsion has been used for many decades to record the
tracks left by elementary particle interactions in three dimensions.
Franz Heymann
franz heymann
Russell Easterly <logi...@home.com> wrote in message
news:dTdK7.46955$XJ4.27...@news1.sttln1.wa.home.com...
>
Yes. X ray film does have a very high resolution. However, a typical
35 mm film has a resolution of around 40 lines per mm. This
translates into something like 1.4 megapixels per picture. A good
digital camera compares very well with this.
Franz Heymann
franz heymann
I have no idea why such a good resolution is required for medical
X-ray film. I am prepared to bet that the X-ray source is not
collimated well enough to be able to fully utilise the film
resolution. Would any reader be prepared to surprise me by saying
that in fact the overall resolution achieved is much better than 1/10
mm, if that? And how many medics are going to scan a 12" x 15" film
for a feature as small as that? If it takes 5 sec to carefully
inspect an area of 1 mm square for a feature whose size corresponds to
the limiting resolution of the film, it will take 156 hours to inspect
the whole image.
Franz Heymann
me...@cars3.uchicago.edu
>
>John C. Polasek <jpol...@cfl.rr.com> wrote in message
>news:3bf9795d.25332335@news-server...
>> Xrays can't be focussed, so you find the patient standing in front
>of
>> a 12" x 15" piece of film stock. You need a vast CD array to cover
>the
>> same territory. Otherwise, use fluoroscopy and focus that image down
>> to the available CCD arrays.
>
>I have no idea why such a good resolution is required for medical
>X-ray film.
It is not. At least there is no physical reason for this (if there is
a legal one, hey, there are enough stupid laws).
I am prepared to bet that the X-ray source is not
>collimated well enough to be able to fully utilise the film
>resolution.
Indeed. Furthermore, unless you're photographing a dead body, the
smearing caused by all the little voluntary or involuntary motions
amounts to much more than the resolution of the film.
> Would any reader be prepared to surprise me by saying
>that in fact the overall resolution achieved is much better than 1/10
>mm, if that?
Ususally it is worse that that.
Russell Easterly
<me...@cars3.uchicago.edu> wrote in message
news:aOBK7.116$P4....@news.uchicago.edu...
> In article <3bfae2c6$0$8506$cc9e...@news.dial.pipex.com>, "franz heymann"
<franz....@care4free.net> writes:
> >
> >I have no idea why such a good resolution is required for medical
> >X-ray film.
>
> It is not. At least there is no physical reason for this (if there is
> a legal one, hey, there are enough stupid laws).
I was told that some new techniques involved
looking at x-ray film with a microscope to find
"micro-fractures. That was like 15 years ago.
me...@cars3.uchicago.edu
Could do it with a synchrotron source, for sure, but not with an x-ray
tube unless I throw nearly all the flux away. If I do this, then the
exposure time, is very, very long. OK for micro-fractures in
inanimate objects, useless for living patients.
Mike Kent
>
> Russell Easterly <logi...@home.com> wrote in message
> news:dTdK7.46955$XJ4.27...@news1.sttln1.wa.home.com
...
> 35 mm film has a resolution of around 40 lines per mm. This
> translates into something like 1.4 megapixels per picture. A good
> digital camera compares very well with this.
Looking at Kodak's MTF curves
B/W films: Tri-X will resolve over 60 lines/mm.
T-Max 100 looks like it will resolve 120+.
Slides: Both Ektachrome and Kodachrome will
resolve 50-60 lines/mm.
franz heymann
Mo Bloe Bro Of Joe Bloe
<MoBloeBro...@thebarattheendoftheuniverse.org> wrote in message
news:qm1kvtk8dk2e3mf18...@4ax.com...
> On Tue, 20 Nov 2001 07:11:12 GMT, me...@cars3.uchicago.edu
> Articulated:
>
> I wonder if any of these posters here besides you know that one
> cannot focus x-ray flux with an optical lens. While an Aluminum
lens
> would work perfectly to do so.
>
> Why would we do that when we have full size plates... :-]
> This just highlights the fact that emitting x-rays through a body
> cavity, and then condensing the image data down to a low res imaging
> chip would be ridiculous.
>
> Replacing the full size plates with an electronic array would be
> (and is) the only real solution. Any reduction would be a
degradation
> of the image data.
That might not be a bad thing in the case of a diagnostic X-ray
picture
Franz Heymann
franz heymann
James Hunter <James....@Jhuapl.edu> wrote in message
news:3BF9857C...@Jhuapl.edu...
Franz Heymann
franz heymann
How many quacks actually investigate 360 megabits to locate a tumour?
Franz Heymann
franz heymann
Mike Kent <mke...@home.com> wrote in message
news:3BFB51D7...@home.com...
Agreed, but the better figures you quote still do not invalidate my
argument.
Franz Heymann
Brian Chandler
> >
> > Russell Easterly <logi...@home.com> wrote in message
> > news:dTdK7.46955$XJ4.27...@news1.sttln1.wa.home.com
> ...
> > Yes. X ray film does have a very high resolution. However, a typical
> > 35 mm film has a resolution of around 40 lines per mm. This
> > translates into something like 1.4 megapixels per picture. A good
> > digital camera compares very well with this.
Does it? I thought "lines per mm" actually meant "line pairs per mm".
In other words, in a millimetre of film you can see 40 black lines,
with 40 white lines in between them. This implies 80 pixels/mm, for a
total of 5.5 megapixels.
(Much more like the figures I recall being quoted in photography
circles.)
me...@cars3.uchicago.edu
>On 18 Nov 2001 05:02:33 GMT, bat...@batcave.cave (Batman) wrote:
>
>>Hi,
>>
>>Perhaps this question is not "well formed", but I'll ask anyway, if only
>>to see its "stupidity". Well, there are two parts actually:
>>
>>1. Can an analog computer be programmed?
>>
>>2. If not, doen't this mean the universe itself is fundamentally digital
>>in
>> nature, because we can actually program stuff in it?
>>
>>Now I'll stand back...way back :-) Thanks,
>>
>>Rajarshi
>Fellas: Xray film doesn't have to have 120 lines/mm. The original spot
>size on the Xray tube target is an irreducible couple of mm's or so.
>To get the smallest spot, they make the anode a rotating cone like on
>a lathe so it doesn't overheat. It would not be possible to make a
>shadowgraph with resolution exceeding the mm spot size.
Something like this, yes. To go much smaller you need different sources.
>I think they may be taking in the public selling such high resolution
>film at probalby a much higher price
Probably. For any work done using x-ray tubes this is a total
overkill. Kinda like fancy hi-fi amplifiers with a flat response all
to way to 50 or 70 khz. Snobs payed big money for such:-)
James Hunter
franz heymann wrote:
Well, I guess you misunderstood what "brain" meant,
that since that's been known to happen recursively in "science",
it's not actually a faux paus of Goedelian magnitude.
John C. Polasek
>Hi,
>
>Perhaps this question is not "well formed", but I'll ask anyway, if only
>to see its "stupidity". Well, there are two parts actually:
>
>1. Can an analog computer be programmed?
>
>2. If not, doen't this mean the universe itself is fundamentally digital
>in
> nature, because we can actually program stuff in it?
>
>Now I'll stand back...way back :-) Thanks,
>
>Rajarshi
Fellas: Xray film doesn't have to have 120 lines/mm. The original spot
size on the Xray tube target is an irreducible couple of mm's or so.
To get the smallest spot, they make the anode a rotating cone like on
a lathe so it doesn't overheat. It would not be possible to make a
shadowgraph with resolution exceeding the mm spot size. I think they
may be taking in the public selling such high resolution film at
probalby a much higher price
John C. Polasek
John Popelish
You might be interested in looking at this site that discusses the
resolution aspects of digital x-rays.
http://www.uic.edu/classes/dadm/dadm396/lect4/inform4c.htm
Here is a clip:
"The spatial resolution of these digital images averages about 9-10
line pairs
per millimeter (lp/mm), while conventional films demonstrate up to
16-18
line pairs per millimeter. Some digital manufactures claim output
at 20 lp/mm.
Humans can only resolve up to 8-9 line pairs. This resolution level
is now
as good as that of film and is diagnostically acceptable."
and:
"Radiographic images are acquired at 8 bit depth and on the gray
scale monitors
they are displayed at 8 bit depth (256 shades of gray). These
images can be
extrapolated by the computer up to 12 bits (4,096 shades of gray).
We only see up to 40 shades of gray."
If an image is acquired at 20 pixels per mm (10 line pairs per mm) and
stored at an uncompressed 6 bits of gray scale (64 levels) that comes
out to 2.6 megabytes per 12 by 14 inch xray, before compression. The
digitization process might well use a much higher resolution gray
scale to collect the raw information, but an optimized 6 bit rendering
is as good as the human eye can make use of. Lots of other image
processing techniques could be applied to get the most useful
information into that final rendering (small scale contrast increase
with large scale contrast reduction, edge emphasis, sharpening, smear
reduction (from motion during exposure), emphasis of details in a
certain size range, etc.
35 mm camera films need a lot higher resolution than xray films,
because they are intended to be enlarged many times before the image
is viewed, whereas xray film is viewed directly.
--
John Popelish
John Popelish
Sorry about the math error. It seems that I converted mm to inches by
using a factor of 2.54. Also, forgot to divide by 8 to change bits to
bytes. Math corrected, below.
> You might be interested in looking at this site that discusses the
> resolution aspects of digital x-rays.
>
> http://www.uic.edu/classes/dadm/dadm396/lect4/inform4c.htm
>
> Here is a clip:
>
> "The spatial resolution of these digital images averages about 9-10
> line pairs
> per millimeter (lp/mm), while conventional films demonstrate up to
> 16-18
> line pairs per millimeter. Some digital manufactures claim output
> at 20 lp/mm.
> Humans can only resolve up to 8-9 line pairs. This resolution level
> is now
> as good as that of film and is diagnostically acceptable."
>
> and:
>
> "Radiographic images are acquired at 8 bit depth and on the gray
> scale monitors
> they are displayed at 8 bit depth (256 shades of gray). These
> images can be
> extrapolated by the computer up to 12 bits (4,096 shades of gray).
> We only see up to 40 shades of gray."
>
> If an image is acquired at 20 pixels per mm (10 line pairs per mm) and
> stored at an uncompressed 6 bits of gray scale (64 levels) that comes
> out to 2.6 megabytes per 12 by 14 inch xray, before compression. The
Make that (25.4 mm/in * 20 per mm)^2 * 12 in * 14 in * 6 bits per
pixel / 8 bits per byte = 32.5 megabytes.
> digitization process might well use a much higher resolution gray
> scale to collect the raw information, but an optimized 6 bit rendering
> is as good as the human eye can make use of. Lots of other image
> processing techniques could be applied to get the most useful
> information into that final rendering (small scale contrast increase
> with large scale contrast reduction, edge emphasis, sharpening, smear
> reduction (from motion during exposure), emphasis of details in a
> certain size range, etc.
>
> 35 mm camera films need a lot higher resolution than xray films,
> because they are intended to be enlarged many times before the image
> is viewed, whereas xray film is viewed directly.
>
> --
> John Popelish
--
John Popelish
John C. Polasek
John Popelish
> Fellas: Xray film doesn't have to have 120 lines/mm. The original spot
> size on the Xray tube target is an irreducible couple of mm's or so.
> To get the smallest spot, they make the anode a rotating cone like on
> a lathe so it doesn't overheat. It would not be possible to make a
> shadowgraph with resolution exceeding the mm spot size. I think they
> may be taking in the public selling such high resolution film at
> probalby a much higher price
The sun is a lot bigger than a few millimeters wide, but I can produce
solar shadow contact prints with a resolution a lot smaller than the
spot diameter of this light source. Hint. The ratio of the distance
from source to film and object (throwing the shadow) to film matters
too.
--
John Popelish
John C. Polasek
wrote:
Yes that's right to the extent that the tube is far away. The part of
the subject far from the film is fuzzed, that near or in contact, not.
The fuzz f = spot S x T thickness/D distance.
f = S(T/D) = T*(S/D)
For a guy 10" thick, and with D = 30", the spot size could be
reduced by 3 and yes to 0 for the subject's skin in contact. That's
why it could be important to take 2 views-do they do that?
But do you really tbink a reduced fuzz size of 1/2 mm warrants film
with 120 lpmm?
As you can see the fuzz is the effective angle of the source (S/D) x T
which for 2 mm and 30x25 = 1/375 radians for the Xray. The sun is 1/2
degree or 1/120. For your 3D shadow print, anthing 10" away, 250 mm
you will get fuzz of 250mm/120 or 2 mm fuzz.
So you will get spot size (reduced) fuzz in any case.
John C. Polasek
Ken Muldrew
>>I think they may be taking in the public selling such high resolution
>>film at probalby a much higher price
>
>Probably. For any work done using x-ray tubes this is a total
>overkill. Kinda like fancy hi-fi amplifiers with a flat response all
>to way to 50 or 70 khz. Snobs payed big money for such:-)
Radiologists certainly see things that I don't see. It seems more like
a pattern recognition task where the pattern that they recognize is
not merely the parts of the image that are tightly focussed. Things
that are too small to be resolved can still change the image, though
without the resolution you have no certainty about what it was that
changed the image. If the radiologists are recognizing patterns then
they may have to maintain the resolution that they were trained with.
I would guess that early x-ray setups just used stock film processes
without bothering to develop new techniques for low resolution film.
After all, you shouldn't lose anything by having high resolution (at
least not back then when film was cheap).
I used to think that radiologists just had overactive imaginations
(and perhaps they do) but one time when I called a guy on some MR
pictures, he showed me the original tapes. The still pictures were
absolutely unconvincing but when you cycled through the images, the
structures were clearly there. We have a pattern recognition and
reconstruction system in our brains that is truly remarkable.
Ken Muldrew
kmul...@ucalgary.ca
Jonathan G Campbell
>
> Hi,
>
> Perhaps this question is not "well formed", but I'll ask anyway, if only
> to see its "stupidity". Well, there are two parts actually:
>
> 1. Can an analog computer be programmed?
>
> 2. If not, doen't this mean the universe itself is fundamentally digital
> in
> nature, because we can actually program stuff in it?
>
> Now I'll stand back...way back :-) Thanks,
>
I guess 'analog computer' could cover a multitude of machines -- not
just
the differential equation simulators of textbooks published in the 1950s
and
60s.
After graduating in 1972, I was interviewed for a job in an offshoot of
a
milling company. Their product was an analog computation of a
linear programming optimisation problem with constraints for animal
feeds: p
possible input ingredients (barley, wheat, soya, linseed oil ..., each
providing some amount of an output content (fiber, oil, protein ...)), q
characteristics of the output feed; problem optimise cost. From what I
remember, the computation had to be done daily according to input costs,
and
current general purpose machinery took a significant part of a day to
get
the answer.
AFAIK, the machine sold for ~$70,000. That was 1972 -- then came the
PDP-11,
and extinction for the analog machine.
Then in a later job, I worked on an analog sort of nearest mean
classifier
for remotely sensed image data. Distance calculations and comparisons
were
done in analog. Fortunately, not long later (1975) digital (small 'd')
again
came to the rescue.
Oh yes, programming in each case was by potentiometer selection of
parameters.
Best regards,
Jon C.
--
Jonathan G Campbell BT48 7PG jg.ca...@ntlworld.com 028 7126 6125
http://homepage.ntlworld.com/jg.campbell/
Steve Harris
news:3bfd3a1d....@news.ucalgary.ca...
> I used to think that radiologists just had overactive imaginations
> (and perhaps they do) but one time when I called a guy on some MR
> pictures, he showed me the original tapes. The still pictures were
> absolutely unconvincing but when you cycled through the images, the
> structures were clearly there. We have a pattern recognition and
> reconstruction system in our brains that is truly remarkable.
COMMENT:
No doubt about it. It just takes a few experiences to show you this. For
example, at a cardiac "cath conference" you can look at a cardiac
cine-angiogram (basically, an X-ray movie), where the cardiac vessels show
up as a nest of spaghetti filled with contrast, moving with every heart
beat. Is this bit, a narrowing? How big is it? It's hard to tell from any
given still. But when you watch the motion picture, your brain subtracts a
lot of crap somehow, and there the answer is, clear as day. It's very
frustrating to try to publish, however!
I had an experience of this kind look at the Zapruder film as it was
included in _JFK_ the movie. You can look at individual color stills from
that thing all day, and it's still not quite clear what is happening. Lots
of grain and magnification problems. But if you see the movie, it's
horrifyingly clear. The right side of the president's head blows out above
the ear, a great dollop of brain comes out, his head whips back and forth,
and he slumps over with a big flap of scalp hanging down. Neither his face
nor the back of his head shows much damage, in contradiction to numerous
conspiracy theories. So that's it.
Insofar as interpolating between "stills," there are really good people have
some kind of visual buffer in which they can abstract data from a series of
stills, and that's where the talent lies in radiology, ultrasound, CIA-type
ground photo-reconnaissance interpretation, and so on. If you have this
ability, it's very difficult to explain to someone who doesn't, what you see
and why. Your brain has done a lot of image processing that even computers
can't do, quite yet. I'm sure that the experts can pick out stuff from
bubble chamber events that even computers cannot see. The problem is that
nobody has the time to let anything but a computer look at most of that
stuff. So the computer has to be used as primary screen, with human back up
for the interesting events.
It's worth digressing a little here to note that this is where the border
between science and art lies. Knowledge transmission has been reduced to a
"science" when it can be written up or put in any kind of recordable medium,
and any trained person reading or watching it, can be taught the technique.
There are rules for what you're doing, and native talent plays a smaller
role. You get to "art" when neither the teacher or student has been able to
quite abstract (all of) the rules, but learning is still accomplished by
showing examples, until inductively the rules are transmitted somehow
anyway, without being named or quantified. It is an art (a so-called "black
art" in engineering) when one neural net teaches another by repeated
example.
Now, the opportunity here for voodoo and shamanism, of course, is large
(especially when there's no good quantifiable feedback, as in the fine
arts). However, voodoo or not, the problem is that (as in radiology and
other kinds of medicine, and engineering) it is evident that something
concrete CAN be transmitted this way, so it's not ALL voodoo. Some talent
can be transmitted whose end product ability and utility might well be
partly measurable by the methods of science, even if the algorithm for
transmitting it, isn't. That's odd, but completely understandable from the
standpoint of neural net theory. Science hasn't even quite known
historically what to do with such stuff. So it remains there on the
borderlands, sometimes called "scientific intuition" or "scientific taste"
or "scientific talent." It's real, but hard to precisely quantify. You run
across it all the time when you meet great scientists, and all you can do is
shake your head. As you did with the radiologists.
SBH
--
I welcome Email from strangers with the minimal cleverness to fix my address
(it's an open-book test). I strongly recommend recipients of unsolicited
bulk Email ad spam use "http://combat.uxn.com" to get the true corporate
name of the last ISP address on the viewsource header, then forward message
& headers to "abuse@[offendingISP]."
Ken Muldrew
>"Ken Muldrew" <kmul...@acs.ucalgary.ca> wrote in message
>> I used to think that radiologists just had overactive imaginations
>> (and perhaps they do) but one time when I called a guy on some MR
>> pictures, he showed me the original tapes. The still pictures were
>> absolutely unconvincing but when you cycled through the images, the
>> structures were clearly there. We have a pattern recognition and
>> reconstruction system in our brains that is truly remarkable.
>
>COMMENT:
>
>No doubt about it. It just takes a few experiences to show you this.
That's for sure. I once spent some time trying to get a computer to
recognize the boundary of a cell as ice grows into the picture (the
cell boundary and the ice boundary look pretty much the same). At
about the same time I recall riding my bike on a road that cut through
a parking lot at an angle and looking for moving cars (so I wouldn't
get creamed) and thinking what an enormous task machine vision will
be. When I was describing some of my problems to a more wise person
than myself he simply said, "Don't be an idiot. Just draw a line
around the damn thing yourself". That saved me a lot of time.
>It's worth digressing a little here to note that this is where the border
>between science and art lies. Knowledge transmission has been reduced to a
>"science" when it can be written up or put in any kind of recordable medium,
>and any trained person reading or watching it, can be taught the technique.
>There are rules for what you're doing, and native talent plays a smaller
>role. You get to "art" when neither the teacher or student has been able to
>quite abstract (all of) the rules, but learning is still accomplished by
>showing examples, until inductively the rules are transmitted somehow
>anyway, without being named or quantified. It is an art (a so-called "black
>art" in engineering) when one neural net teaches another by repeated
>example.
In that case, I nominate mathematics as the queen of the arts.
>Now, the opportunity here for voodoo and shamanism, of course, is large
>(especially when there's no good quantifiable feedback, as in the fine
>arts). However, voodoo or not, the problem is that (as in radiology and
>other kinds of medicine, and engineering) it is evident that something
>concrete CAN be transmitted this way, so it's not ALL voodoo. Some talent
>can be transmitted whose end product ability and utility might well be
>partly measurable by the methods of science, even if the algorithm for
>transmitting it, isn't.
That pretty much sums up the dilemma that scientists face in trying to
get funding. You need to prove that you have talent and potential but
that can only be evaluated from a body of work and its impact on the
rest of the community. I suppose the magnitude of the problem is a
function of how long it takes to evaluate work once it has been
completed. It's pretty easy to examine an athlete's ability but hard
to examine a teacher. That must be the reason for the difference in
their salaries (the teacher is too risky since it might turn out that
you gave all that money to a flim-flam artist).
>That's odd, but completely understandable from the
>standpoint of neural net theory. Science hasn't even quite known
>historically what to do with such stuff. So it remains there on the
>borderlands, sometimes called "scientific intuition" or "scientific taste"
>or "scientific talent." It's real, but hard to precisely quantify. You run
>across it all the time when you meet great scientists, and all you can do is
>shake your head. As you did with the radiologists.
I like the way Mark Kac put it when describing Feynman. "There are two
kinds of geniuses: the 'ordinary' and the 'magicians'. An ordinary
genius is a fellow whom you and I would be just as good as, if we were
only many times better. There is no mystery as to how his mind works.
Once we understand what they've done, we feel certain that we, too,
could have done it. It is different with the magicians. Even after we
understand what they have done it is completely dark. Richard Feynman
is a magician of the highest calibre."
It's strange how the number of magicians seems to decline as you get
older, though. I think my first year as an undergrad is probably when
I met the most people who seemed to have an utterly incomprehensible
genius. Now I don't meet anyone like that (although I do interact with
a lot of people who are brighter than I am). I wonder if it's just an
overdeveloped cynicism.
Ken Muldrew
kmul...@ucalgary.ca
Tom Potter
"Jonathan G Campbell" <jg.ca...@ntlworld.com> wrote in message
news:3BFD8C3A...@ntlworld.com...
In the late 60's, the head of the Ralston Purina Company data processing
department
told me that they had a DIGITAL program, that manipulated the contents
of their various products (Foods for pets, cows, pigs, humans, etc. )
based on spot prices, futures, cost of storage, transportation,
and of course, the particular nutrients needed in the various products,
and that this program was responsible for ALL of the profit
that was generated by Ralston Purina.
The program advised them of when and what to buy and sell,
when and where to ship, rent storage space,
adjust product formulas, etc.
I suppose that they have a much better program today.
--
Tom Potter http://home.earthlink.net/~tdp
-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 90,000 Newsgroups - 16 Different Servers! =-----
me...@cars3.uchicago.edu
>me...@cars3.uchicago.edu wrote:
>
>>>I think they may be taking in the public selling such high resolution
>>>film at probalby a much higher price
>>
>>Probably. For any work done using x-ray tubes this is a total
>>overkill. Kinda like fancy hi-fi amplifiers with a flat response all
>>to way to 50 or 70 khz. Snobs payed big money for such:-)
>
>Radiologists certainly see things that I don't see. It seems more like
>a pattern recognition task where the pattern that they recognize is
>not merely the parts of the image that are tightly focussed. Things
>that are too small to be resolved can still change the image, though
>without the resolution you have no certainty about what it was that
>changed the image.
Not jsut this. You may "see" things which are not there at all. Any
decent experimentalist is aware of this problem.
If the radiologists are recognizing patterns then
>they may have to maintain the resolution that they were trained with.
>I would guess that early x-ray setups just used stock film processes
>without bothering to develop new techniques for low resolution film.
Of course.
>After all, you shouldn't lose anything by having high resolution (at
>least not back then when film was cheap).
>
Sure. It is not a problem having more resolution that needed, but
there is no need to pretend that it is required. Radiologists
nowadays are quickly converting to using imaging plates which
typically have a resolution of 0.1 mm but allow for fast processing
and yield directly a digital image. Then you can play with these
images superimposing them one on another and really start finding
patterns.
>I used to think that radiologists just had overactive imaginations
>(and perhaps they do) but one time when I called a guy on some MR
>pictures, he showed me the original tapes. The still pictures were
>absolutely unconvincing but when you cycled through the images, the
>structures were clearly there. We have a pattern recognition and
>reconstruction system in our brains that is truly remarkable.
>
Yes, indeed. No existing hardware-software combination comes even
close.
Steve Harris
>
> That pretty much sums up the dilemma that scientists face in trying to
> get funding. You need to prove that you have talent and potential but
> that can only be evaluated from a body of work and its impact on the
> rest of the community. I suppose the magnitude of the problem is a
> function of how long it takes to evaluate work once it has been
> completed.
That's the problem. And it's why scientists and various kinds of
sociology-of-science types seem to talk past each other. It's only if you
take the very long and historical view that science lifts itself above the
noise of fashion and other kinds of stuff that social anthropologists study.
In the short term, science lives by all of the rules of ordinary advertising
and self-promotion. It's great to make a solid contribution to the knowledge
of mankind, but you'd like to have it acknowledged by being spread and used
and cited within your professional lifetime. Would like to get grants,
promotions, tenure, prizes, fame. Even a better shot at getting your next
paper published in the journal of your choice. And that all takes more than
just good ideas and work, alas.
>
> I like the way Mark Kac put it when describing Feynman. "There are two
> kinds of geniuses: the 'ordinary' and the 'magicians'. An ordinary
> genius is a fellow whom you and I would be just as good as, if we were
> only many times better. There is no mystery as to how his mind works.
> Once we understand what they've done, we feel certain that we, too,
> could have done it. It is different with the magicians. Even after we
> understand what they have done it is completely dark. Richard Feynman
> is a magician of the highest calibre."
A good quote from Gleik's bio. As also Gell-Mann's description of Feynman's
methods, which are generally not taught at Cal Tech: 1) you write down the
problem, 2) think very hard 3) write down the answer.
> It's strange how the number of magicians seems to decline as you get
> older, though. I think my first year as an undergrad is probably when
> I met the most people who seemed to have an utterly incomprehensible
> genius. Now I don't meet anyone like that (although I do interact with
> a lot of people who are brighter than I am). I wonder if it's just an
> overdeveloped cynicism.
Dunno. I wish I'd met Feynman. I did meet Gell-Mann once, and he was indeed
goddamn smart.
jmfb...@aol.com
kmul...@acs.ucalgary.ca (Ken Muldrew) wrote:
>"Steve Harris" <sbha...@ix.RETICULATEDOBJECTcom.com> wrote:
>
>>"Ken Muldrew" <kmul...@acs.ucalgary.ca> wrote in message
>
>>> I used to think that radiologists just had overactive imaginations
>>> (and perhaps they do) but one time when I called a guy on some MR
>>> pictures, he showed me the original tapes. The still pictures were
>>> absolutely unconvincing but when you cycled through the images, the
>>> structures were clearly there. We have a pattern recognition and
>>> reconstruction system in our brains that is truly remarkable.
<snip>
>>It's worth digressing a little here to note that this is where the border
>>between science and art lies. Knowledge transmission has been reduced to
a
>>"science" when it can be written up or put in any kind of recordable
medium,
>>and any trained person reading or watching it, can be taught the
technique.
>>There are rules for what you're doing, and native talent plays a smaller
>>role. You get to "art" when neither the teacher or student has been able
to
>>quite abstract (all of) the rules, but learning is still accomplished by
>>showing examples, until inductively the rules are transmitted somehow
>>anyway, without being named or quantified. It is an art (a so-called
"black
>>art" in engineering) when one neural net teaches another by repeated
>>example.
>
>In that case, I nominate mathematics as the queen of the arts.
We were able to "tell" if the computer system was in trouble first
by the sound pattern [or lack] of the clickety clack of the
terminals. When the sound pattern changed, all of us (programmers
and grunts) would turn around and look at the light pattern
(also known as blinkenlights). Just about everybody figured out
how to tell if the system was going to crash and if there was
enough time to give the exit command to the editor.
I know the art existed, but I was never able to describe how
to do it.
>
>>Now, the opportunity here for voodoo and shamanism, of course, is large
>>(especially when there's no good quantifiable feedback, as in the fine
>>arts). However, voodoo or not, the problem is that (as in radiology and
>>other kinds of medicine, and engineering) it is evident that something
>>concrete CAN be transmitted this way, so it's not ALL voodoo. Some talent
>>can be transmitted whose end product ability and utility might well be
>>partly measurable by the methods of science, even if the algorithm for
>>transmitting it, isn't.
>
>That pretty much sums up the dilemma that scientists face in trying to
>get funding. You need to prove that you have talent and potential but
>that can only be evaluated from a body of work and its impact on the
>rest of the community. I suppose the magnitude of the problem is a
>function of how long it takes to evaluate work once it has been
>completed. It's pretty easy to examine an athlete's ability but hard
>to examine a teacher. That must be the reason for the difference in
>their salaries (the teacher is too risky since it might turn out that
>you gave all that money to a flim-flam artist).
Don't underestimate the notion of feedback. These days, an almost
instant feedback is required for funding. I don't know how science
and other research is going to advance in this day and age when
some of it should take years of careful study.
>
>>That's odd, but completely understandable from the
>>standpoint of neural net theory. Science hasn't even quite known
>>historically what to do with such stuff. So it remains there on the
>>borderlands, sometimes called "scientific intuition" or "scientific
taste"
>>or "scientific talent." It's real, but hard to precisely quantify. You
run
>>across it all the time when you meet great scientists, and all you can do
is
>>shake your head. As you did with the radiologists.
>
>I like the way Mark Kac put it when describing Feynman. "There are two
>kinds of geniuses: the 'ordinary' and the 'magicians'. An ordinary
>genius is a fellow whom you and I would be just as good as, if we were
>only many times better. There is no mystery as to how his mind works.
>Once we understand what they've done, we feel certain that we, too,
>could have done it. It is different with the magicians. Even after we
>understand what they have done it is completely dark. Richard Feynman
>is a magician of the highest calibre."
I like that description.
>
>It's strange how the number of magicians seems to decline as you get
>older, though. I think my first year as an undergrad is probably when
>I met the most people who seemed to have an utterly incomprehensible
>genius. Now I don't meet anyone like that (although I do interact with
>a lot of people who are brighter than I am). I wonder if it's just an
>overdeveloped cynicism.
I've known a couple of the magicians. Every once in a great while
I discover another one. They're still out there but they're not
working on producing great stuff. Since they're also human, they
need more than a couple of months to get something remarkable out
the door.
/BAH
Subtract a hundred and four for e-mail.
Ken Muldrew
>Sure. It is not a problem having more resolution that needed, but
>there is no need to pretend that it is required. Radiologists
>nowadays are quickly converting to using imaging plates which
>typically have a resolution of 0.1 mm but allow for fast processing
>and yield directly a digital image.
Are these similar to the amorphous selenium plates used to see the
beam in linac x-ray treatment machines? Are they based on the
photoelectric effect?
Ken Muldrew
kmul...@ucalgary.ca
me...@cars3.uchicago.edu
>me...@cars3.uchicago.edu wrote:
>
>>Sure. It is not a problem having more resolution that needed, but
>>there is no need to pretend that it is required. Radiologists
>>nowadays are quickly converting to using imaging plates which
>>typically have a resolution of 0.1 mm but allow for fast processing
>>and yield directly a digital image.
>
>Are these similar to the amorphous selenium plates used to see the
>beam in linac x-ray treatment machines?
Hmm, I don't know, maybe. Have to check the details.
>Are they based on the photoelectric effect?
They're based on excitation of electrons to metastable states
(lifetime of the order of hours) in some ionic compound. Said states
can then be forced to deexcite using a laser beam of an appropriate
wavelength. So, after exposing the plate you "read" it by scanning it
with a laser beam (measuring the density of excited states in each
location). The scanning process takes about 1 minute with the device
we're using, but I think that there are already newer devices which
cut it down to about 10-15 sec. Resolution is 0.1 mm, but I think
that this is a matter of practicality, not an "inherent limit".
Ken Muldrew
>>Are these similar to the amorphous selenium plates used to see the
>>beam in linac x-ray treatment machines?
>
>Hmm, I don't know, maybe. Have to check the details.
>
>>Are they based on the photoelectric effect?
>
>They're based on excitation of electrons to metastable states
>(lifetime of the order of hours) in some ionic compound. Said states
>can then be forced to deexcite using a laser beam of an appropriate
>wavelength. So, after exposing the plate you "read" it by scanning it
>with a laser beam (measuring the density of excited states in each
>location). The scanning process takes about 1 minute with the device
>we're using, but I think that there are already newer devices which
>cut it down to about 10-15 sec. Resolution is 0.1 mm, but I think
>that this is a matter of practicality, not an "inherent limit".
Hah! Now that you mention it I think there was a laser in there (I was
looking at one of the linacs yesterday for the first time). Thanks for
the explanation.
Ken Muldrew
kmul...@ucalgary.ca
Gregory L. Hansen
Ioannis <morp...@olympus.mons> wrote:
>Randy Poe wrote:
>[snip]
>> > 1. Can an analog computer be programmed?
>>
>>
>> Proof by existence: We are programmable analog computers.
>
>Ho, ho! But we are NOT! We are digital as well. All our sensory organs
>have a "minimum threshold resolution", including eyes (cones & rods),
>tactile nerves, hearing (we cannot distinguish between two tones closer
>than, say, 1/100th of a semitone) and finally and most importantly, our
>brain is a very complex array of 1/0 neurons.
>
>1: neuron firing for T and 0, neuron idle for F. Wonder why our math is
>based on the naturals?
The rate at which neurons fire is continuous, up to the threshhold
frequency. A lot of things determine when a neuron fires. It fires when
the concentration of stimulating transmitter minus inhibiting transmitter
passes a threshhold. Those transmitters are constantly being destroyed by
enzymes, so the cooperative effects of two neurons on a third is weighted
by distance, which is continuous, and the time that each fires, which is
again continuous.
>
>I've been trying to think of a thought with a resolution less than one,
>but have been unsuccessful. Any takers? :*)))))
I have half a mind to tell you about voltage to frequency converters in
signal processing electronics!
--
"'No user-serviceable parts inside.' I'll be the judge of that!"
x...@erols.com
wrote:
----------------------------
snip
----------------------------
>
>> and there aren't any for 3 dimensions. Of course a photo is
>> not content searchable, but then neither is a digital 'photo'. By what
>> rule of logic would you be impelled to search for the value 1273 or
>> any other integer? That means any individual pixel has approximately
>> zero value. It takes FFT over a group of pixels.
>
>I don't see the relevance of this. I don't want to search for 1273,
>I want to know what it is you think a digital camera cannot do that
>a photograph can.
>
>> Debates like this can get quite sickening so consider this my last
>> volley.
>
>--
>Richard Herring | <richard...@baesystems.com>
Actually, they do have some reasonable digital cameras. Each picture,
uncompressed, is about 4 GigaBytes.
When the digital camera can deliver a meaningful 16 to 64 GigaBytes,
uncompressed, per picture, and do lossless compression to a reasonable
file size, then Digital cameras will have replaced film cameras for
all practical purposes. Of course, the resolution of the printers and
monitors will have to increase in like manner.
But with that resolution, the stereo pictures should look the same.
Since you are starting to press the optical limits of the glass, there
should not be any significant increases in film resolution over
current available films.
That leaves only changes in film sensitivity and range.
You cannot do lossy compression, unless you know (in advance) what
data you will need, and what future generations will look for in the
picture. If you can do 1000:1 lossless compression, giving you 64 MB
files, you would have the beginning of a workable system.
But I do not think 1000:1 lossless compression is possible, but it is
a good goal to try to achieve. Present digital cameras are good
enough for snapshots and screensavers on current monitors.
x...@erols.com
<franz....@care4free.net> wrote:
------------------
snip
------------------
>> Yes of course it's the same. But considering an average CCD is about
>> 1" square, you would need 180 of them to cover the territory and a
>way
>> to stitch up the gaps between the chips. Then at 2 megabits per chip
>> you need to store 360 megabits in a 30th of a second or so. I would
>> advise a bus 180x16 wide to get the job done. And you'd need more
>than
>> a Z80!
>
>How many quacks actually investigate 360 megabits to locate a tumour?
>
>Franz Heymann
>
>
None. All you need to do is to know in advance which part of the image
will contain the tumor, and only have high resolution on that area.
The rest of the image could be erased, as long as you maintained the
spatial cordinates of the tumour image.
I would think that most doctors would want better resolution than
360MB on a image the size of the human chest. Eight bits per pixel
should be enough for the exposure data on an x-ray, since the x-ray
film does not have a great gamma curve.
x...@erols.com
wrote:
>franz heymann wrote:
>>
--------------------------------------
>Here is a clip:
>
> "The spatial resolution of these digital images averages about 9-10
>line pairs
> per millimeter (lp/mm), while conventional films demonstrate up to
>16-18
> line pairs per millimeter. Some digital manufactures claim output
>at 20 lp/mm.
> Humans can only resolve up to 8-9 line pairs. This resolution level
>is now
> as good as that of film and is diagnostically acceptable."
>
>and:
>
> "Radiographic images are acquired at 8 bit depth and on the gray
>scale monitors
> they are displayed at 8 bit depth (256 shades of gray). These
>images can be
> extrapolated by the computer up to 12 bits (4,096 shades of gray).
> We only see up to 40 shades of gray."
>
Reasonable up to here.
Some pearl graders have been tested at over 4000 shades of gray.
you are two orders of magnitude off.
If you said the average person, never training himself to look at gray
images and interpolate them, can only see up to 40 shades of gray,--
I would probably agree.
When the doctor has suspicions about an area, he will study that one
spot very carefully. The entire picture must have enough resolution
to allow that level of analysis.