I think the problem may be one of definition. IIRC, "mountains" are caused by uplift and "hills" by erosion. What to call when both work together, like the Appalachians? Hmm, maybe "inbred?"
Anyway, here's how to put coal in a hill or mountain:
all of the "statments known to be false" that I've heard referenced here on Slashdot and in other discussions about similar topics, have all been verses taken anywhere from slightly to grossly out of context from what any unbiased reader simply reading along would come to understand as the meaning.
I am an atheist who believes a lot of the Bible. There really are factual errors in it. There really are facts in it too.
For instance, I believe that Genesis 10 "Table of Nations" is how the Hebrews viewed themselves as fitting into the world-at-large.
I do not believe it is completely factually accurate. Elam being Semitic, for instance, when the facts of archeology tell us they were related to nobody nearby, or maybe to the Dravidians. But not descendants of Shem.
How would the Hebrews know any better? Their guess was probably as good as could be expected.
If the Bible were inerrant, I would expect the Table of Nations to parallel Cavalli-Sforza's work exactly;-)
Supernovas already put the energy in to every atom heavier than iron (IIRC). We blow a few neutrons at it and we can get that supernova solar (FLoaBW) power back out.
Then we blow a few neutrons at the nuclear "waste" to make it into fuel or (more) stable isotopes.
It takes so little nuclear fuel to replace so much coal:
"one kilogram of uranium can theoretically produce... as much electricity as 1500 metric tons of coal." -Wackypedia entry "Uranium" (emphasis added and spelling enhanced)
One kilogram of uranium vs. one million five hundred thousand kilograms of coal.
1500000 : 1 pounds, tons, talents, however you measure your fuel. Every unit of uranium fuel replaces ~4500000 units of CO2 in the atmosphere.
Anyone who is pro-nuke is more of an environmentalist than anyone who is anti-nuke.
No general-purpose algorithm is going to expand the cloud of smoke preferentially in one direction, and then suddenly terminate the billowing edge of the smoke cloud against a clear sky, because general-purpose image enhancement algorithms do not model the behaviour of billowing smoke clouds. No general-purpose algorithm is going to cut out whole buildings and transplant them perfectly to other parts of the neighbourhood, because general-purpose algorithms do not recognize where buildings start and end against a backdrop of other buildings.
It's quite a bit more sophisticated than Gaussian blur or bicubic interpolation. DCT and FFT "back in the day" and I'm sure it's well past Laplace-land these days.
When you tag pixels as "bad" they are set to null and then you do the transform. Do the inverse transform and but go ahead and let the synthesis equations fill in the "bad" pixels. That's the process in a nutshell.
Marking pixels "bad" removes their contribution to the image. In the reconstruction the data has to come from elsewhere, it comes from the rest of the image through the inverse transform.
You're right, the algorithm certainly does not know "building."
But neither does my guitar digital delay box know "notes."
This algorithm "models" the behavior of the whole image, billows and all. The frequency domain representation *is* the model of the image. Which is why this algorithm looks so damn good when done right: the interpolated values are very true to what should be there.
I tested this by using bogus defect channels and compare (subtract) the reconstruction with the original. Usually the difference was well below the resolution of 16-bit ints and always indistinguishable to human eyes. Then when you ask too much of the algorith you get a disaster like the image we are discussing.
The widespread feathering is a common artifact of this kind of image processing. There are even knobs to deal with it on some GUIs (like this free-to-download implementation from Polaroid (IIRC).
We don't know how much user input went into making Hajj's defect channel. If he designated a block of Beirut as bad pixels, the algorithm would have to interpolate that whole block. IOW, the "selectivity" you attribute to the algorithm may well be the selectivity of the operator one Adnan Hajj.
And to address your other pole, the algorithm did randomly chop up other parts of the image: the aforementioned feathering. It went on a pogrom against large populations of gray (algorithmically these are close to null) all over the image.
You glossed over the upper left section of smoke, among other things. There was nothing there before hand, it was added
No, the picture was cropped. Following the time-honored rule "sky is boring" the top was brought down. Look in the correction image, it has much more room at the top.
This, maybe more than anything, is what makes the smoke seem bigger, when really a lot of it was cropped out.
You point to a building at 2c and 2d in your file which is cloned to 3a and 3b. However, the one at 3a and 3b can be seen in the original, but was moved down to the lower section.
In the correction image, the good one, there is one building which in the botch is at 2c,2d. At the place which is 3a,3b there is a lot of smoke obscuring more regular boring buildings. I very much suspect the regularity behind the smoke provided "hinting" to make the "echo" happen as bad as it did.
I have tried to keep from calling the second-to-be-published image the "original" as it probably wasn't. It is a correction image. I would truly, truly love to have the stuff right out of Hajj's camera and computer. Would solve everything in a real hurry!
More importantly, it's not at quite the same relative postion within your gridlines.
A tilde can mean "approximately" and that's what I meant when I said "~16 pixels apart, cutting the image into 16 nearly-equal bands." Same with the words "nearly-equal." Maybe I should expand the symbols into words to make them more visible.
The whole lower half of the original appears to have been copied, sharpened, copied back in lower and to the left, and the smoke added in a vain attempt to cover it up, then cropped to hide the lower right corner which didn't have anything in it.
Give it a whirl if you think that's what happened.
And while you are committing psychology in public, perhaps you could explain why. Why take all that damn time to fake an image? A news photographer has ten pictures before and after that one. He's not hurting for content. He's hurting for time!
Dust and scratch removal is quite automated, and is even "on" by default in some products. It is part of the workflow. And so are bad lighting conditions which Hajj blames for having missed the awfulness of the bad image (laptop in the sun is my interpretation of bad lighting).
What motivation for taking all that time to copy a building here and there and then go put feathers in the smoke on regular 1/16th centers....
At least no UFOs were harmed in the construction of your conspiracy theory.;-)
The next step, if someone was paying me for this, would be to try to replicate the disaster using some readily-available dust & scratch removal software, like Sane for the GIMP.
If Hajj's lawyer or Reuters were laying appropriate bucks at my feet, I would explore the problem through SciPy and PIL.
Hajj's disastrous image is an example of the kinds of errors we will have to get used to recognizing.
In the olden days, we would correct scratches by putting a drop of light mineral oil on the negative and putting glass over that. The oil filled in the scratches similar to the way the DCTs fill in the scratches nowadays.
Reuters deserved some reputation damage, as Hajj's photos aren't all that great and quite obviously Reuters's photo editor was asleep at the switch.
But accusing them of publishing faked photos is in this case fakery itself: pretending to knowledge that nobody has.
(Claimer: I was a photojournalist for various school organs for about a decade. I've done DSP professionally several times, and love doing it in my free time as well. If you count my PWM synth for the Apple ][, I've been doing DSP since 1979.)
125 watts or thereabouts at rest, several times that when exerting.
The resting brain eats 30W. And thence you can see the idea that we only use 10% of our brain is bunk, as then the brain would be glowing at 300W an yer haid wood splode.
Or as I originally wrote it in Motorola-accented pseudocode about 1980 or so,
$2B | ~$2B = $FF
Code review:-) Why are you promoting the sum of two bytes to a 16-bit quantity? A sum might make at most a 1-bit carry. (A mult of course might need a 16-bit destination.)
That's not very surprising when you think about the roots of modern math: Islamist mathematicians during Europe's middle Ages... Europe wasn't always that mighty thing that it is now you know... In fact Europe got out of that period because of the insights and knowledge they "acquired" from Muslims during the "Crusades"...
The unfortunate image you refer to was, exactly like Adnan Hajj the photographer said, the result of dust removal and other image correction.
It was a bad image, but completely comparable to darkroom problems of bygone days. It should never have run, and a lot of the blame belongs to Reuters'eses (a little Gollum lingo there) photo editor.
To dispense with this urban bombing legend:
The artifacts are on 1/16th of the image width boundaries, a(n) = n/16 * width.
Even the duplicated building is moved in such increments.
I don't have to explain to this audience the significance, but for those reading over shoulders, computers move in powers-of-two, and the only things people do that way is fold stuff and most people are not very good at it. Origamists are experts.
The duplicated building is a DSP-created "echo" (unsure of aptness of metaphor). In the "corrected" image, we can see that region is very much middle-of-the-spectrum gray. There is even inadvertent "hinting" because of the regular grids of roads and building faces. And we do not know what the defect channel was, some software lets the user go nuts.
In some (many? most?) implementations, pixels flagged in the defect channel are ignored, and is replaced with data interpolated from other pixels. Simple interpolations are usually inadequate, and much better results are achieved with more sophisticated DSP, such as DCT, FFT, and even Laplace transform-based implementations.
Combining hinting, middle of spectrum, and a botched attempt at identifying defects produced the catastrophic image, comlete with an "echoed" building.
I strongly suspect the other goofiness in Hajj's pictures have similar technical problems.
Reuters has a policy against misleading photographs. And since many naive viewers were misled into thinking Adnan Hajj intentionally duplicated parts of the image, they at least temporarily quit using his services.
And even if the weirdnesses are all unintentional, their sheer number creates doubt as to Hajj's technical competency as a photographer. "You can have a heart of gold but if your pictures suck I can't use you." -a lambda character I just invented for the quote
Claimers: I've done a lot of DSP programming for a long time and got paid for a lot of it, including some DCT-based dust removal software, and I think I know the theory up to but not including Laplace transform yoga; I've done a lot of photography for a long time, all kinds except portrait, and never been paid for any of it.
Adnan Hajj's disastrous image reminded me of test runs of the dust removal software, when I would intentionally screw with the input in all (physical and programmatic) sorts of ways, use random defect channels, "Mondrian" defect channels and input, I even (ab)used screen caps of Babylon 5.
The wrecked image caught my eye as something familiar. I measured the defects, and suggest that those who impugn Hajj's motives are committing psychology in public.
Slashdot Mirror
I think the problem may be one of definition. IIRC, "mountains" are caused by uplift and "hills" by erosion. What to call when both work together, like the Appalachians? Hmm, maybe "inbred?"
Anyway, here's how to put coal in a hill or mountain:
1. Coal layer in sediment.
2. Uplift.
3. Erosion.
4. Hills with coal in them.
5. Profit!
For instance, I believe that Genesis 10 "Table of Nations" is how the Hebrews viewed themselves as fitting into the world-at-large.
I do not believe it is completely factually accurate. Elam being Semitic, for instance, when the facts of archeology tell us they were related to nobody nearby, or maybe to the Dravidians. But not descendants of Shem.
How would the Hebrews know any better? Their guess was probably as good as could be expected.
If the Bible were inerrant, I would expect the Table of Nations to parallel Cavalli-Sforza's work exactly
Supernovas already put the energy in to every atom heavier than iron (IIRC). We blow a few neutrons at it and we can get that supernova solar (FLoaBW) power back out.
Then we blow a few neutrons at the nuclear "waste" to make it into fuel or (more) stable isotopes.
It takes so little nuclear fuel to replace so much coal:
"one kilogram of uranium can theoretically produce
One kilogram of uranium vs. one million five hundred thousand kilograms of coal.
1500000 : 1 pounds, tons, talents, however you measure your fuel. Every unit of uranium fuel replaces ~4500000 units of CO2 in the atmosphere.
Anyone who is pro-nuke is more of an environmentalist than anyone who is anti-nuke.
A friend of mine from childhood has a geologist father who worked on siting power plants and other necessary infrastructure.
My friend grew up with death threats. They weren't just addressed to the father.
The Nimby Rousseau Banana Luddite saboteur terrorists are a nasty bunch.
It's best to take all possible precautions when dealing with such dangerous animals.
When you tag pixels as "bad" they are set to null and then you do the transform. Do the inverse transform and but go ahead and let the synthesis equations fill in the "bad" pixels. That's the process in a nutshell.
Marking pixels "bad" removes their contribution to the image. In the reconstruction the data has to come from elsewhere, it comes from the rest of the image through the inverse transform.
You're right, the algorithm certainly does not know "building."
But neither does my guitar digital delay box know "notes."
This algorithm "models" the behavior of the whole image, billows and all. The frequency domain representation *is* the model of the image. Which is why this algorithm looks so damn good when done right: the interpolated values are very true to what should be there.
I tested this by using bogus defect channels and compare (subtract) the reconstruction with the original. Usually the difference was well below the resolution of 16-bit ints and always indistinguishable to human eyes. Then when you ask too much of the algorith you get a disaster like the image we are discussing.
The widespread feathering is a common artifact of this kind of image processing. There are even knobs to deal with it on some GUIs (like this free-to-download implementation from Polaroid (IIRC).
We don't know how much user input went into making Hajj's defect channel. If he designated a block of Beirut as bad pixels, the algorithm would have to interpolate that whole block. IOW, the "selectivity" you attribute to the algorithm may well be the selectivity of the operator one Adnan Hajj.
And to address your other pole, the algorithm did randomly chop up other parts of the image: the aforementioned feathering. It went on a pogrom against large populations of gray (algorithmically these are close to null) all over the image.
This, maybe more than anything, is what makes the smoke seem bigger, when really a lot of it was cropped out.In the correction image, the good one, there is one building which in the botch is at 2c,2d. At the place which is 3a,3b there is a lot of smoke obscuring more regular boring buildings. I very much suspect the regularity behind the smoke provided "hinting" to make the "echo" happen as bad as it did.
I have tried to keep from calling the second-to-be-published image the "original" as it probably wasn't. It is a correction image. I would truly, truly love to have the stuff right out of Hajj's camera and computer. Would solve everything in a real hurry!A tilde can mean "approximately" and that's what I meant when I said "~16 pixels apart, cutting the image into 16 nearly-equal bands." Same with the words "nearly-equal." Maybe I should expand the symbols into words to make them more visible.Give it a whirl if you think that's what happened.
And while you are committing psychology in public, perhaps you could explain why. Why take all that damn time to fake an image? A news photographer has ten pictures before and after that one. He's not hurting for content. He's hurting for time!
Dust and scratch removal is quite automated, and is even "on" by default in some products. It is part of the workflow. And so are bad lighting conditions which Hajj blames for having missed the awfulness of the bad image (laptop in the sun is my interpretation of bad lighting).
What motivation for taking all that time to copy a building here and there and then go put feathers in the smoke on regular 1/16th centers....
At least no UFOs were harmed in the construction of your conspiracy theory.
Yes, Adnan Hajj's unfortunate images were "doctored" as in "given too much medicine," the medicine being dust & scratch removal.
But it was not faked, nor was image content "cloned" with that tool.
This Image Is Not Faked
The next step, if someone was paying me for this, would be to try to replicate the disaster using some readily-available dust & scratch removal software, like Sane for the GIMP.
If Hajj's lawyer or Reuters were laying appropriate bucks at my feet, I would explore the problem through SciPy and PIL.
Hajj's disastrous image is an example of the kinds of errors we will have to get used to recognizing.
In the olden days, we would correct scratches by putting a drop of light mineral oil on the negative and putting glass over that. The oil filled in the scratches similar to the way the DCTs fill in the scratches nowadays.
Reuters deserved some reputation damage, as Hajj's photos aren't all that great and quite obviously Reuters's photo editor was asleep at the switch.
But accusing them of publishing faked photos is in this case fakery itself: pretending to knowledge that nobody has.
(Claimer: I was a photojournalist for various school organs for about a decade. I've done DSP professionally several times, and love doing it in my free time as well. If you count my PWM synth for the Apple ][, I've been doing DSP since 1979.)
This is no troll its a useful message.
I think the purpose of local governments is to limit the damage they can do.
Okay, say the next company comes along with a very good system.
It shows Republicans winning by ever bigger majorities.
Placing doubt on the doubt.
But it's not about "the environment" or "saving the world" or whatever.
It is about technophobia and bad religion.
Within ten years, the United States could stop emitting CO2 from power plants simply by replacing them with nuclear power plants.
Nukes are the answer.
So, why are these Planetary Messiahs and Priests of Gaia against nuclear power?
Because they hate people. Not in a Sarte "hell is other people" way, but in a deep, sick, self-loathing way.
And being too cowardly to commit suicide they seek political means to kill people.
Being anti-nuke is being anti-human.
We should be able to be proactive about it and kill those luddites before they kill us.
I figured it out. That is, how such an uninformed post was modded Insightful.
The I scrooled back to the top and re-read the Bush bash.
Explains the moderation.
All Bush bashing is insightful.
It is the definition of insightful.
Being insightful is easy.
125 watts or thereabouts at rest, several times that when exerting.
The resting brain eats 30W. And thence you can see the idea that we only use 10% of our brain is bunk, as then the brain would be glowing at 300W an yer haid wood splode.
Or as I originally wrote it in Motorola-accented pseudocode about 1980 or so,
$2B | ~$2B = $FF
Code review:-) Why are you promoting the sum of two bytes to a 16-bit quantity? A sum might make at most a 1-bit carry. (A mult of course might need a 16-bit destination.)
Or they want to work on their own time and schedule and terms.
Prague
Proggy music to wash your brain in DRM-free MPEG-4 AAC playable in iTunes, mplayer, WinAmp, and so on.
I'm thinking of a "listenware" license, if you download it you are obligated to listen to it once all the way through.
Especially Make It Stop is that the perfect name or what?
The open source Audacity was used for various audacious tasks in the production of this music.
I will try to put new music up a few times a year so check back once a season or so.
I will check back here later in the day for comments, suggestions, reviews, and insults.
If you like it tell your friends.
Enjoy!
-Toddhisattva
Prague
ps: no odd meter? WTF? The odd-meter monster is stuck in editing.
Yeah it may crash but it has games!
(Microsoft will not be defeated by any competition. They will be defeated by wrongful death lawsuits.)
Build a Yamato and put these on it instead of the half-meter rifles the last one had.
Islamic not Islamist
Anyway, the roots of modern math were planted by some amazing Hindus and Persians and carried to the West by Mohammedans.
Here's one way:
12112111211112111112....
It's a pattern and it will never repeat.
It will just get more and more boring!
But the coastal Arabs had such an enviable setup the Romans called it "Happy Arabia."
CO2 is not pollution, it's incredible aerial fertilizer!
(I am only half kidding)
The unfortunate image you refer to was, exactly like Adnan Hajj the photographer said, the result of dust removal and other image correction.
It was a bad image, but completely comparable to darkroom problems of bygone days. It should never have run, and a lot of the blame belongs to Reuters'eses (a little Gollum lingo there) photo editor.
To dispense with this urban bombing legend:
The artifacts are on 1/16th of the image width boundaries, a(n) = n/16 * width.
Even the duplicated building is moved in such increments.
I don't have to explain to this audience the significance, but for those reading over shoulders, computers move in powers-of-two, and the only things people do that way is fold stuff and most people are not very good at it. Origamists are experts.
The duplicated building is a DSP-created "echo" (unsure of aptness of metaphor). In the "corrected" image, we can see that region is very much middle-of-the-spectrum gray. There is even inadvertent "hinting" because of the regular grids of roads and building faces. And we do not know what the defect channel was, some software lets the user go nuts.
In some (many? most?) implementations, pixels flagged in the defect channel are ignored, and is replaced with data interpolated from other pixels. Simple interpolations are usually inadequate, and much better results are achieved with more sophisticated DSP, such as DCT, FFT, and even Laplace transform-based implementations.
Combining hinting, middle of spectrum, and a botched attempt at identifying defects produced the catastrophic image, comlete with an "echoed" building.
I strongly suspect the other goofiness in Hajj's pictures have similar technical problems.
Reuters has a policy against misleading photographs. And since many naive viewers were misled into thinking Adnan Hajj intentionally duplicated parts of the image, they at least temporarily quit using his services.
And even if the weirdnesses are all unintentional, their sheer number creates doubt as to Hajj's technical competency as a photographer. "You can have a heart of gold but if your pictures suck I can't use you." -a lambda character I just invented for the quote
Claimers: I've done a lot of DSP programming for a long time and got paid for a lot of it, including some DCT-based dust removal software, and I think I know the theory up to but not including Laplace transform yoga; I've done a lot of photography for a long time, all kinds except portrait, and never been paid for any of it.
Adnan Hajj's disastrous image reminded me of test runs of the dust removal software, when I would intentionally screw with the input in all (physical and programmatic) sorts of ways, use random defect channels, "Mondrian" defect channels and input, I even (ab)used screen caps of Babylon 5.
The wrecked image caught my eye as something familiar. I measured the defects, and suggest that those who impugn Hajj's motives are committing psychology in public.