Ebay Holograms

[Dinesh]

Sorry, perhaps I was venting a little. I have a profound opposition to "experts" and the "expert syndrome" . My way for learning about something, or finding out something, is to study the subject in great depth in all the aspects I can think of. While doing this, I ignore all the experts and all the "common knowledge" - the stuff that "everybody knows" - and come up with what-ifs.

For example, in the question of X-rays, I came up against this question a while back and came up with a hypothesis. I then tested my hypothesis by assuming that the dosage of X-rays (and probably, roughly the wavelength and bandwidth) of dental X-rays is about the same as the scanning X-rays for airline carry-on luggage. The next time I went to my dentist, I gave him a small piece of a dcg film and asked if he would zap it, as it was a dental X-ray, he's well used to my bizarre requests. I then developed and found I was right. If I'm wrong, I re-invent my model. Then I come up against an expert and I find that this expert is simply re-hashing the common knowledge - usually without understanding it! When I first started doing dcg image-planed holograms, everyone compared them to Augie. No one, even Augie, had ever done an image-planed dcg to the best of my knowledge. But, no one appreciated this! Then Colin came to our lab to learn dcg. Apparently, while he was in another part of the lab learning to laminate the hologram, he told Joy that the hologram we'd made (the mermaid) was every bit as good as an Augie piece. Joy told him not to mention this to me as it was a sore point with me. Colin, in a very diplomatic way, told me later that evening that Augie's holograms were almost as good as mine!

As for your test, I'm afraid I didn't even recognise any of the names. Steve Hart was over here and we had the same conversation. He asked me whether I'd pay 8 million dollars for a shark immersed in formaldehyde. "You're joking!" I said. The next day I got an email titled "Oh Ye of Little Faith" with a picture of a shark immersed in formaldehyde in an art gallery with a price tag of 8 million dollars! There are schools today that can't afford textbooks, the literacy rate in parts of the US and Europe are going below 19th century literacy standards and some one will pay this kind of money for a dead shark! Amazing!

[Tony]

Thanks Dinesh for your insight.
I work at a medical laser company and have worked with two types of what I call true Geniuses. Ones who are crazy methodical, everything modeled everything tested, even the thermometers are tested in boiling water to verify that the are reading correctly, use two power meters because the first may not be right, test 50 optics from the same batch to see how much they vary (there's a fun one for you). Hundreds of experiments and spare no expense on equipment and outside testing houses to try every possible way of achieving results.
Then there are the intuitive ones. These are the guys who work one a shoestring and are all feel. I cannot count how many times one of these guys will be presented with a problem (many times they have not even seen the set up) and say try "this". And even though your leaving his office shaking your head that it cannot possibly work, it does. They are wired very differently. When I'm asked what project do I want to work on, I often side to the intuitive ones. Why? Because for me it's like a magic trick, they put this stuff out of thier butts and pow they are right.
So what were the results of the X-ray test?

[Dinesh]

Exposed holograms did nothing. Unexposed dcg plates got harder. I expected that there would be an overall hardening, but I felt that it would not become so cross linked that it was not possible to shoot anymore. Sure enough, they did cross link a little, but I could still shoot them. The efficiency was lower than I would have got. Effectively, they aged the plate by what would have been a few weeks of dark reaction.

By the way, whoever it is that tests thermometers by putting them in boiling water, tell 'em they're not testing the thermometer. The boiling point is water is defined as the property of a particular substance that's achieved when the substance is placed in boiling water, the expansion of mercury to a particular volume, for example. All, they're doing is testing the definition!

[BobH]

By the way, whoever it is that tests thermometers by putting them in boiling water, tell 'em they're not testing the thermometer. The boiling point is water is defined as the property of a particular substance that's achieved when the substance is placed in boiling water, the expansion of mercury to a particular volume, for example. All, they're doing is testing the definition!

At the risk of being argumentative, aren't they really just verifying the calibration of the instrument? Testing the definition presumes a known and highly accurate thermometer.

[Jeffrey Weil]

Hello Dinesh and Bob,

For techi debating fun you guys can work that out but I have another problem with the test. It's highly unlikely it was run correctly in the first place. To do this you need perfectly clean water, reagent grade. The vessel has to be perfectly clean and it probably has surface finish requirements for the test to be accurate. Local defects in the vessel's walls and bottom have a large effect on boiling.

The accuracy and stability of the heat source over time would be doubtful. All the materials used would have to be up to spec. And who's to say what a "rolling boil" really is anyway? That's a pretty subjective concept. What altitude was the test run at? Only at sea level does water boil at 212F.

You would have to be NIST or something like that to do this test correctly and use it to actually calibrate thermometers. If not, I bet using this method the thermometer could be plus or minus a full degree or two. That's a swing of 4 degrees, not very accurate.

It might be fine for cooking and making candy but not for serious lab work, not the kind of work where your actually worried about the accuracy of your thermometer.

Jeff W

[Dinesh]

At the risk of being argumentative, aren't they really just verifying the calibration of the instrument? Testing the definition presumes a known and highly accurate thermometer.

Well, the standard (and older) definition of the temperature in the Celsius scale is that water boils at 100 C and freezes at 0 C (actually, zero is the temperature of a mixture of ice and water). Thus to calibrate the thermometer, you'd have to put it in boiling water, mark the position of the mercury as 100, then put it in a mixture of ice and water and mark that position as zero. You'd then divide the interval into a hundred equally spaced units. To find just one point does not give you a scale, it just gives you one point. So, by putting the thermometer into boiling water and noting that the mercury has reached a pointed marked "100", you've confirmed that the physical property of the measuring material - the expansion of mercury - in the instrument does go to the correct position. Under the definition of 100C you've either accepted the expansion of mercury and confirmed that boiling water has a temperature of 100, or you've accepted that boiling water is 100 and confirmed the behaviour of the mercury. In the latter case, you'd be right and he's just verifying the calibration.

To actually test the thermometer, you'd have to have a standard known and highly accurate thermometer, as you say. You'd also have to have an atmosphere of controlled pressure. You'd then heat the water, take the temperature at various intervals, measure with your standard, and note the position of the mercury as a ratio between the zero point and the "100 point ". You'd then plot the measured temperature given by the "good" thermometer against the ratio of the high and low points of your thermometer and (hopefully) get a straight line.

Jeff,
You've left out that the thermometer has to be suspended in the water. If it sits at the bottom of the vessel in which the water is boiling, it's measuring an equilibrium temperature of the sides of the vessel and the water. The water must also j-u-u-st be boiling, otherwise the released air bubbles will form a protective coating around the bulb of the thermometer. Also, the pressure must be "standard", ie 15 pounds/sq inch or 1 atmosphere.

By the way, the freezing and boiling points of water are no longer the standards of tempertaure. The new scale (well, new since the 40's) is the absolute scale. Under this scale, the freezing point of water is defined as 273.15K at it's triple point - that value of temperature and pressure at which an equilibrium exists between all three phases of water. The "boiling point" of water is then (273.15 + (373.15/273.17)). Under these scales, water freezes at 0.01 C and boils at 99.9839 °C.

[Jeffrey Weil]

Hello Dinesh,

Lucky for us we don't need to be that accurate with temp. I heard a great quote when I was getting a degree in advertising photography. "A constant error eliminates itself". Here's an example. If were using a thermometer that's 20 degrees off, but always the same 20, we would find the best temp through trial and error and as long as we always use that thermometer and that reading, no matter how wrong it was, we would always get a great hologram.

That's a big part of the reason I don't like when people suggest times and temps, or really any other specs like that, and other's use those standards without doing experiments of their own. If my thermometer is off by 20 degrees and I use your numbers I get crap, if I start with yours and through experiments find the best temp for my lab, with my equipment and my techniques it'll always work.

Jeff W

[Tony]

Well, the standard (and older) definition of the temperature in the Celsius scale is that water boils at 100 C and freezes at 0 C (actually, zero is the temperature of a mixture of ice and water). Thus to calibrate the thermometer, you'd have to put it in boiling water, mark the position of the mercury as 100, then put it in a mixture of ice and water and mark that position as zero. You'd then divide the interval into a hundred equally spaced units.

I see what type of R&D engineers you guys would be

[Dinesh]

I heard a great quote when I was getting a degree in advertising photography. "A constant error eliminates itself".

Actually, Jeff, that quote has another dimension. "Error" in the physical sciences refers to the small variations between what is read off an instrument and the "true" value. So, if you read 20 degrees off a thermometer, you'd compare the rise of the mercury to the number 20 and note that as the temperature. But, the scales are usually 1 degree apart, so the "true" temperature could be anything from 20.5 and 19.5 (halfway between scales). If you repeated the reading and got 20 again, the real temperature could still be between 19.5 and 20.5, so you still have the +/- 0.5 degree error between the actual temperature and your reading of it. If you took repeated readings and got 20 each time, the real temperature could be 19.5 sometimes and 20.5 at other. In the long run, statistically, the fluctuation of 1 degree between 19.5 and 20.5 would mean that the real temperature could be anything from 19.5 and 20.5 - sometimes above 20 and sometimes below - and you'd still read the thermometer as 20. The constant error has eliminated itself. Of course in quantum mechanics there is no "true" reading.

Actually, measurement theory has always fascinated me. Bringing this back to holography, we know that one of the mismatches between the object beam and the reconstructed is affected by the size of the source. Is it possible, I wonder, to work this backwards? Can you measure the mismatch of an image and work out the size of the source? I know there are well-known relations between the image blur and the source size, the Van Cittert-Zernicke Theorem will tell you that, but how do you measure the blurring of an image and give a figure for it?

[Jeffrey Weil]

Hello Dinesh,

I think you could measure the blur by making a hologram of a resolution test chart. Play it back with the most proper source you can find and measure the target to set a standard of obtainable resolution.

From there you can do the measurements of other sources, possibly with machine vision.

What do you think? Would that work?

O, I have another one. Make a hologram of a single focused point in real space. Put a somewhat small sensor at that point. The less blur, the more gain in the sensor. More blur would expand the size of the focused spot so less would be on the small sensor.

You could make different holograms with the spot at different projected distances to get a nice range of testing materials.

I think that one might be better.

My last would be more machine vision. Is it possible for machine vision to set a size for the circles of confusion in an out of focus image? That could be a good quantization of the blur. Maybe an edge detection technique.

I think the focused spot idea might be the best one I had here. What are your thoughts?

Jeff W