Jump to content

Wikipedia:Reference desk/Archives/Science/2015 February 27

From Wikipedia, the free encyclopedia
Science desk
< February 26 << Jan | February | Mar >> February 28 >
Welcome to the Wikipedia Science Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.


February 27

[edit]

NASA's beautiful new denoising algorithm

[edit]

Anyone seen [1] [2]? Titan has never looked so good. It traces back to [3] which is locked up. Do we have any good (and hopefully more approachable) explanation to what the algorithm is? Specifically, I'm wondering if it could work on pre-existing immunofluorescence images and the like. Wnt (talk) 02:34, 27 February 2015 (UTC)[reply]

A horribly formatted draft of the JGR article can be downloaded here (found using Google Scholar). It cites "A non-local algorithm for image denoising" by Buades et al., which can be downloaded here. The latter paper says "the denoised value at x is a mean of the values of all points whose gaussian neighborhood looks like the neighborhood of x"—in other words they are blending together similar-looking regions of the image to reduce noise. This seems similar to advanced upscaling algorithms like Genuine Fractals™ that use the image itself as a codebook. It strikes me as very dangerous to use this kind of algorithm as a prelude to image analysis, as these authors do, since there's a good chance that the details it "reveals" are not really there. -- BenRG (talk) 06:21, 27 February 2015 (UTC)[reply]
BenRG describes a bilateral filter, or an enhancement along those lines. Nimur (talk) 15:36, 27 February 2015 (UTC)[reply]
I don't know the terminology, but that appears to be different as it only averages nearby pixels. I should have linked non-local means, which is a term used by Buades et al. (You already linked it below.) -- BenRG (talk) 21:16, 27 February 2015 (UTC)[reply]
On close inspection, I see that you are correct. The key word that I overlooked is "all", as in a mean of the values of all points whose gaussian neighborhood looks like the neighborhood of x". A bilateral filter operates similarly but only considers points within a specified radius. Nimur (talk) 21:42, 27 February 2015 (UTC)[reply]
It's actually quite easy to design your own denoising algorithm with a performance that is far better than anything that is implemented in standard software. I use my own algorithms on my raw picture files which yields far better results than the in-camera noise reduction or anything that photoshop has to offer. Basically, what I do is I use a model for the a priori probability distribution of the set of grey values of the pixels in the picture (the prior probability of a picture in the absense of noise). The parameters describing that model will behave in a certain way under scaling, which allows me to estimate these. Then I have a good model for the noise which includes correlations and outliers. I can then solve for the most probable picture iteratively starting with an approximate guess. That guess is obtained by eliminating the outliers in some simple way (e.g. local median) and approximating the noise + picture model by a Gaussian model, allowing one to use Fourier transform methods to get to an approximate solution. Then this is the first approximation that can be improved iteratively.
On a fast pc, denoising properly takes a long time (hours, not minutes) and it requires quite a lot of preliminary calculations to do the model estimation before you can run the program. E.g. the point spread function should be specified, it matters if you specify the picture as perfectly sharp or if the point spread function is more spread out and the picture is not sharp at the level of the pixels. And, of course, when you are done you need to do the demosaicing and the transforms to e.g. SRGB colorspace before you have a picture, but the results are worth it. Count Iblis (talk) 14:00, 27 February 2015 (UTC)[reply]
Count Iblis describes a non-local means estimator, e.g. a spectral estimation filter or an iterated statistical estimator an enhancement along those themes. This is a statistical signal processing method.
It does not necessarily require hours to compute one of these, in its most basic form; but if you are iterating for the purpose of optimizing some kind of model fit, you can iterate as long as you wish. If you have an incredibly elaborate model, each iteration could take a very long time.
Nimur (talk) 15:33, 27 February 2015 (UTC)[reply]
What Count Iblis described doesn't sound much like what's described in the non-local means article. -- BenRG (talk) 21:16, 27 February 2015 (UTC)[reply]
A more sophisticated denoising algorithm can make images look subjectively better, but no algorithm can recover the actual detail that was lost to the noise. That information is not present in the noisy image and can only be guessed. More sophisticated algorithms make it less obvious which features of the output were guessed, which is good if the goal is a good-looking image, but bad if the goal is scientific analysis (which is what Wnt was asking about, I think). -- BenRG (talk) 21:16, 27 February 2015 (UTC)[reply]
Wnt, you should definitely find a way to read this article, which was published in IEEE Signal Processing about two years ago: A Tour Of Modern Image Processing. Nimur (talk) 15:43, 27 February 2015 (UTC)[reply]
  • I should thank all who responded, especially User:BenRG; I've been progressing slowly toward doing this, learning some C++ in the process. The Cassini paper seems straightforward, while the one referenced is a bit more confusing (what do I make of a calculation like Ga*|u(x+.)-u(y+.)|2(0) -- yes, that looks like a zero ...) but I should probably hit the mathdesk only after I've finished taking my best hack at the program beginning to end. Wnt (talk) 02:01, 3 March 2015 (UTC)[reply]
That is a kernel (e.g. a weighting function), convolved with a (derived) signal, then evaluated at zero. In this case, the kernel is convolved with the image values - rather, the norm (magnitude) of the difference between pixel values. Keep in mind that the output of a convolution is an entire sequence (e.g., a vector): so if you want a scalar parameter, you have to specify an element of that vector. For this algorithm, the 0th point is desired. (Review how convolution works if the motivation for that isn't immediately obvious).
The equation is complicated because it's building up an operation out of many many simpler operations. The notation - using the language of operators and functionals - is very familiar to mathematicians and signal processing theorists, but it's quite different from the notation that many programmers prefer, because it's nontrivial to turn that equation into pseudocode. Let us know if you get stuck. Nimur (talk) 00:04, 4 March 2015 (UTC)[reply]

Latex gloves

[edit]

Do latex gloves go through a water test like condoms? And are latex gloves equally permeable?199.119.235.217 (talk) 04:06, 27 February 2015 (UTC)[reply]

Yes, the gloves that are used for sterile surgical procedures or for handling hazardous material go through a water test. I don't know about permeability. See http://www.astm.org/Standards/D3577.htm Looie496 (talk) 14:49, 27 February 2015 (UTC)[reply]

Physics: How can any black hole move?

[edit]

Give reference please! A. MohammadZade Iran --78.38.28.3 (talk) 06:44, 27 February 2015 (UTC)[reply]

Fixed the formatting. --70.49.169.244 (talk) 17:39, 27 February 2015 (UTC)[reply]
As per our article, "there is no observable difference between the gravitational field of such a black hole and that of any other spherical object of the same mass". Presumably they would therefore interact with spacetime the same as any other object except within the event horizon. In other words, they move the same as anything else. Remember, black holes aren't vacuums; if our sun were magically replaced with one of the same mass, it would get dark and cold on Earth, but that's pretty much the only difference, gravitationally speaking. Our orbit wouldn't change. Mingmingla (talk) 06:59, 27 February 2015 (UTC)[reply]

Equatorial bulge and viewing distant objects

[edit]

Since the Earth has an equatorial bulge where the Earth's curvature is at its greatest and the poles are the flattest points of the Earth's curvature, does that mean that we get to see things at greater distances at the poles than in the equator, especially when looking from North to South and vise versa? Let's say we are standing at the North Pole and there is a ship 5 miles away that we can barely see because of the distance. Only its masts are visible because of the Earth's curvature. Then, let's say we and the ship go near the equator and the same expanse and visibility are as clear as when we stood at the North Pole and the same ship is there at exactly at the same distance, will the ship be seen just as barely as it was in the North Pole, or will the whole ship be out of sight because it will be hidden under the Earth's curvature due to the equatorial bulge? Has such experiment ever been done near the poles and then near the equator? Willminator (talk) 06:58, 27 February 2015 (UTC)[reply]

I'm having a surprisingly hard time finding a reference for this. Earth bulge#Distance to horizon (bulge has a different meaning here) and Horizon#Distance to the horizon both seem to deal only with a spherical Earth, as, I was amazed to discover, does Bowditch. Online I found this Distance to the Horizon university page which at least mentions that "Numerically, the radius of the Earth varies a little with latitude and direction; but a typical value is 6378 km (about 3963 miles)." and proceeds from there with a spherical Earth.
My impression from this is that the difference must not be significant, but let's try a calculation. Note that we are interested in the radius of curvature, not the actual distance to the center of the Earth. Earth radius#Radii of curvature just says to see Spheroid#Curvature, and that's more math than I wish to deal with at the moment. Assuming that the radius of curvature along a meridian at the equator is about equal to the semi-major axis at the pole and vice-versa, then the distances to the horizon should differ by about 1 part in 600 because the flattening of the Earth is about 1 part in 300 and there is square root in the formula in Horizon#Geometrical model. (sqrt(1-x) ≈ 1 - x/2 for small x.) But Horizon#Effect of atmospheric refraction says that with standard atmospheric conditions, refraction adds about 8% to the value calculated from the geometric model, and in unusual atmospheric conditions the distance to the horizon will vary by much more than the 0.17% difference we get from the varying geometric model from equator to pole.
So yes, it differs, but not enough to be noticeable. Sorry that I can't find any good references here. -- ToE 10:00, 27 February 2015 (UTC)[reply]
My assumption regarding the radii of curvature was a poor one as it significantly underestimates the difference in the radius of curvature along a meridian from equator to pole. Radius of curvature (mathematics)#Ellipses tells me that with a flattening of 1 part in 300, the two radii of curvature will vary by 1 part in 100. Thus the distance to the horizon from our geometric model will vary by 0.5%. This is still small compared to effects of unusual atmospheric conditions. -- ToE 10:20, 27 February 2015 (UTC)[reply]
Left open is the question of whether this has ever been observed experimentally. Perhaps, instead of taking separate measurements at equator and pole, it would be best to conduct such an experiment only at the equator, where the radius of curvature along the equator is 2 parts in 300 greater than the radius of curvature along a meridian. This yields a 0.33% difference in the calculated distance to the horizon. I suspect that during periods of settled weather the atmosphere would be sufficiently azimuthally isotropic for this difference to be detectable. -- ToE 10:39, 27 February 2015 (UTC) I'm 12.5°N. If only the seas were calmer and I had two equally nerdy friends with sailboats.[reply]
If Earth is approximated as an oblate spheroid (with no atmosphere), possibly the easiest way to calculate this is to first calculate the horizon for a sphere whose radius equals Earth's equatorial radius, and then scale where f is the flattening ratio. It follows that the horizon on an oblate spheroid is elliptical regardless of location, and that at the equator the straight-line distance between the northern and southern horizons is equal to 1−f times the straight-line distance between the eastern and western horizons, or 0.33% smaller on Earth, which agrees with ToE's calculation. At the pole you have to start with a height that's 1/(1−f) larger in order to have the correct height after rescaling z, and since the horizon distance goes roughly as the square root of the height, and rescaling z doesn't rescale the horizon in this case, the distance should be roughly 0.17% larger than the east-west distance at the equator, or 0.5% larger than the north-south distance at the equator. By a slightly different argument, it's 0.33% larger than the distance on a sphere with a radius equal to Earth's polar radius. -- BenRG (talk) 22:18, 27 February 2015 (UTC)[reply]
We might also want to consider that it's generally colder at the poles, which means that the air is denser - which (presumably) alters the refractive index. The effect may be a small one - but the curvature effect is also small...so it's hard to estimate. SteveBaker (talk) 06:27, 28 February 2015 (UTC)[reply]
Which is why I suggested an experiment at the equator, comparing north-south viewing distance with east-west viewing distance. Standing in my cockpit (no real "pit" -- just deck level behind the pilothouse) my eye height is 3.4 m above the water, which in typical atmospheric conditions gives a distance to the horizon (Young's method) of 7.1 km (3.8 nautical miles). Were I to position myself on the equator in calm weather, and two friends hung a light at the same 3.4 m above the water on their boats, with one heading away from me along the meridian and the other along the equator, the limit of visibility of the lights should be twice the distance to the horizon, or 14.2 km (7.7 nm). My and BenRG's calculation's above suggest that the actual distance along the equator should be 0.33% greater than that along the meridian, which is a difference of only 47 m. And in glassy calm conditions there is still usually a long period swell, and, even as low as 10 cm, that would reduce viewing distance by 210 m when both observer and observed were in a trough, with the maximum calculated distance attained only when both were on a crest. I could have my friends move slowly back and forth at the limit of visibility while trying to notice the moment their light becomes first visible, but, with a consistent train of swells, might the observer be phase-locked (wrt the wave train) with each observed? It is far from clear to me that the difference due to the equatorial bulge could be properly measured. Confounding this is that it is often during the calmest weather that unusual atmospheric conditions occur. See Fata Morgana (mirage). -- ToE 08:03, 1 March 2015 (UTC)[reply]

Is there a scientific consensus on how long people who existed during the Stone Age lived?

[edit]

Is there a scientific consensus on how long people who existed during the Stone Age lived? If so, what is the consensus? Fanddlover5 (talk) 10:07, 27 February 2015 (UTC)[reply]

There is a huge consensus among the scientific community that, during the Stone Age, humans had a lifespan of 30 years. Icemerang (talk) 10:13, 27 February 2015 (UTC)[reply]
Is that maximum typical longetivity, or life expectancy at birth? (I.e. is it including or ignoring infant mortality?) Iapetus (talk) 12:34, 27 February 2015 (UTC)[reply]
I found The neolithic revolution and contemporary variations in life expectancy which (looking at the graph on page 2) suggests that life expectancy might have dipped below 25 during the Neolithic period, to increase again later. Alansplodge (talk) 13:14, 27 February 2015 (UTC)[reply]
Figures in the 25-35 year range are almost certainly life expectancy at birth, see also my links below. SemanticMantis (talk) 15:05, 27 February 2015 (UTC)[reply]
"Huge consensus" is an overstatement. Life expectancy is not so easy to calculate. You have to assume that the distribution of skeletal remains accurately reflects the distribution of the population, and there are a variety of reasons why that might not be correct. Looie496 (talk) 14:39, 27 February 2015 (UTC)[reply]
You could also say that, "In order to prove that humans evolved from beings that looked like apes, you have to assume that the distribution of skeletal remains accurately reflects the distribution of the population." But nobody makes that argument (well, creationists do, actually). Even before the development of genetics, a majority of people already accepted human evolution as fact, primarily due to fossil evidence (and when only using fossils as evidence to back up one's claims, the number of fossils one needs to prove that humans evolved is greater than the number of fossils one needs to prove that humans from the Stone Age lived until age 30). You'd be hard-pressed to find anyone who thinks that humans lived to older ages during the Stone Age. But to get to the point and to answer the OP's question (which is why we're all discussing this in the first place), there is already a scientific consensus on Stone Age lifespan, and this is acknowledged by even those who disagree with the consensus. As a few examples, Mark's Daily Apple admits that it has become an article of faith among virtually everyone that our ancestors lived short lives, and LIVNAKED admits that an underlying point which is widely shared among researchers and the public-at-large is that our ancestors did not live long enough to develop cancer, heart disease, and other chronic illnesses. The most influential person in paleodemography (as well as being the person to popularize it and set an example for future paleodemographers to follow) has been, since the 1950s, the late John Lawrence Angel, whose figures for Stone Age longevity, as stated in his book, The People of Lerna, are given as 27 years for females and 33 years for males, giving an answer of an average of 30 years for both sexes combined to the OP's question. Anthropologist Rachel Caspari has provided (further) evidence that most early humans rarely lived to older ages. In her own words, "the conclusion was inescapable". Icemerang (talk) 05:02, 28 February 2015 (UTC)[reply]
Here's an open access journal article on the topic, which also includes many references [4]. At birth, estimates usually fall in to the 25-35 year range. However, most of that low age is due to extremely high infant mortality. If you were a neolithic human who survived to ~18 years old, and survived childbirth if female, then the expectancy goes way up, and it wouldn't have been that odd to see a few 65 year old people around. Probably not as many as today. The main idea is that the oldest neolithic people lived about as long as the oldest people in the modern era, but that the high rates of infant mortality and death-by-birthing bring the average expectancy way down compared to modern life expectancy. However, Maximum_life_span hasn't changed nearly as much. SemanticMantis (talk) 15:04, 27 February 2015 (UTC) (p.s. I see that Alansplodge has already linked the the same article as I did)[reply]
Actually believe it or not, Medeis, lifespan was indeed 30 years. Icemerang (talk) 05:02, 28 February 2015 (UTC)[reply]
I am confused, you seem to be assuring me the truth of a claim I neither denied nor doubted. My concern was to provide a widely and cheaply available source for educated laymen, available at most libraries. μηδείς (talk) 20:29, 28 February 2015 (UTC)[reply]
Ötzi (3300 BCE (end stoneage start copperage)) was estimated to be 45 years old. I doubt the use of copper tools caused a jump in life expectancy so humans could likely reach higher lifespans atleast at the end of the stone age. --Kharon (talk) 02:07, 28 February 2015 (UTC)[reply]
Otzi was one of the rare individuals to reach an age beyond 30. Higher lifespans were still rare at the time of Otzi's death. Icemerang (talk) 05:02, 28 February 2015 (UTC)[reply]
How is living to 45 necessarily rare amongst a group with a life expectancy at birth of 30 (or even 25) when a significant contribution to the low life expectancy is an extremely high infant mortality? From Life expectancy: "In countries with high infant mortality rates, LEB is highly sensitive to the rate of death in the first few years of life. Because of this sensitivity to infant mortality, LEB can be subjected to gross misinterpretation, leading one to believe that a population with a low LEB will necessarily have a small proportion of older people.". -- ToE 07:03, 1 March 2015 (UTC)[reply]
For one thing, it was tough making it to 40. I don't know if they had a life expectancy at birth of 30 or 25. Their life span was 30. A little different from life expectancy. Icemerang (talk) 08:11, 1 March 2015 (UTC)[reply]
Ah! OK, I see that you have consistently been speaking of life span, from your initial response to this question to your reply to Medeis which she confused as a confirmation worded as a contradiction. The other respondents, however, feel that these numbers represent life expectancy at birth, and the paper (Galor & Moav) linked to by Alansplodge and SemanticMantis specifically states life expectancy at birth for its figures, which it gives as about 32 - 33 years for the Copper Age. Above you gave a couple of links to what I believe are paleolithic lifestyle blogs, but both of them appear to contradict what you are asserting. You quoted from both, but what you quoted are what the blog entries claim is the popular but incorrect perception which does not take into account the skewing effect of infant mortality, and they both go on to argue that, once surviving to early adulthood, these people would have had, on average, much longer lives than most ill informed people believe. You also mention Angel's work, but without it in front of me I can't tell whether or not your 27 & 33 year figure is for life expectancy at birth. Do you have any scholarly reference which specifically indicates that these low figures are maximum expected life spans and not LEB?
More generally, can anyone here find a reference which gives life expectancy at age of maturity during these time periods? -- ToE 14:02, 1 March 2015 (UTC)[reply]
Yes, the blog entries do state otherwise, though the point I was trying to prove was that they both admit that there is a consensus among nearly everyone that life in the Stone Age was short. Since these are apparent paleolithic lifestyle blogs, their arguments for longer lifespans may be a little biased. Also, the very same article that Mark's Daily Apple uses as evidence for longer lifespans in the Stone Age states "Contemporary hunter-gatherers have been affected by global socioeconomic forces and are not living replicas of our Stone Age ancestors.". The article that Alansplodge and SemanticMantis linked to is a working paper, so I wouldn't put too much faith in it. Also, Angel's book, The People of Lerna, which gives the 27 and 33 year (giving an average of 30 years) figures for Stone Age lifespan, states on page 72, "I must stress that all these statistics are simple statements based on frequencies of deaths; in no sense do they represent life expectancy". Articles from scientific journals also confirm that life in the Stone Age was short. As an article in Nature put it, the majority "died between twenty and forty years", though even making it to forty was tough. This is as much of a fact as evolution is. There have been other articles dealing with this topic, as well. And I don't know life expectancy at age of maturity, but if it helps, people reached adulthood at age 14, which was also the age they had children. Icemerang (talk) 04:45, 2 March 2015 (UTC)[reply]

Mobile Phones

[edit]

Once a day I used my smartphone - using WiFi - while charging it , after about ten minutes I noticed that its temperature became too high , why did that happen ? and what bad results maybe happened to my phone ? 149.200.218.248 (talk) 16:35, 27 February 2015 (UTC)[reply]

They do that. "Too high" is a matter of opinion. Can it be bad for your phone? Check the article I linkes for an explanation. Mingmingla (talk) 16:42, 27 February 2015 (UTC)[reply]

Is the speed of the electron being a (determinant of) main in nature?

[edit]

If the speed of the electric current in all environments been constant, because all environments were had the electrical current had been a permanent magnetism, so that what is been the speed of light?--85.141.239.201 (talk) 19:52, 27 February 2015 (UTC)[reply]

Nothing permanent had been. InedibleHulk (talk) 20:15, 27 February 2015 (UTC)[reply]
Recent research published in Science [5] indicates that certain spatial structures of photons travel through a vacuum at speeds notably lower than c. Some popular press coverage here [6]. That's probably not what you're asking about, but then again it might be... SemanticMantis (talk) 21:42, 27 February 2015 (UTC)[reply]
This is another paper that plays word games with different definitions of velocity, like the perennial "superluminal" anomalous dispersion papers. Science seems to choose papers not by merit but by how much pop-press coverage they'll get. -- BenRG (talk) 22:55, 27 February 2015 (UTC)[reply]
I get that, but I'm also unable to rigorously critique the methods and definitions. The coverage in Science News said something along the lines of "the structured photons got the the target several micrometers ahead of the unstructured photons, per meter traveled." Do you have any reason to believe that statement to be false/inaccurate? At least this is claiming light can move slower in a vacuum, not faster. Are there any peer-reviewed responses/critiques yet? I know this is off-topic, and I could have posted a new question, but our recent poster going on about light speed and the recent article got me interested :) In the mean time, I guess I'll read the Science article more carefully... SemanticMantis (talk) 23:59, 27 February 2015 (UTC)[reply]
I see the word "determinant" in permanent. Not sure if that helps, but it rhymes. Maybe just an unavoidable pattern? Might be prudent to factor in the Thue–Siegel–Roth theorem, might not. Depends what you mean by main, perhaps. InedibleHulk (talk) 22:02, 27 February 2015 (UTC)[reply]
The speed of electrons is not constant. See Speed of electricity. Even in a vacuum, the velocity depends on the accelerating voltage. Mr.Z-man 23:13, 27 February 2015 (UTC)[reply]
I’m take it thinking so, that since the electric current is always been presented in the environments constantly had’s a magnetism, so that the speed of the electric current in these environments is always been a constant (speed of electron which is always been a constant in all environments), thus this the speed of light is nothing as the movements of the electric current in the optical conductor (optical medium).--83.237.204.207 (talk) 07:22, 28 February 2015 (UTC)[reply]
Assumes, that the electron(s) had a different states of gravity, by this been explained the weakness of the light (the weakness of inertia of the light) in nature.--83.237.219.86 (talk) 09:01, 28 February 2015 (UTC)[reply]
Note: The speed of the electron(s) is always been a constant in all environments, because Ampere Force and Volt’s of Ampere Force did not been depended on the speed (accelerations) of the electron(s) in nature.--83.237.219.86 (talk) 09:14, 28 February 2015 (UTC)[reply]
And so also for overcome, if you referring to a state of vacuum, you should be know, that vacuums are been different, they been: sparse, inert and sparsely inert, and such much over.--85.141.235.74 (talk) 11:48, 28 February 2015 (UTC)[reply]
What have you been reading? Dbfirs 15:30, 28 February 2015 (UTC)[reply]
Thanks for your question. By physics and chemistry had implies that aggregate states of the vacuum can be different, so the vacuum had different aggregate states.--83.237.204.85 (talk) 16:25, 28 February 2015 (UTC)[reply]
Since the physical environments are been different, so and vacuums are been different, because the vacuum is been the same physical environment.--85.141.236.21 (talk) 22:19, 28 February 2015 (UTC)[reply]
Mr Alex Sazonov, why don't you log in into your main account here, user:Alex Sazonov? 79.237.93.71 (talk) 01:23, 1 March 2015 (UTC)[reply]