Wikipedia:Reference desk/Archives/Computing/2013 August 9

Welcome to the Wikipedia Computing 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.

August 9

Images that have pixels of nearly every color of the spectrum

I am doing some independant research, and i need to find out if there are any images out there that have nearly every RGB color expressible with r,g,b being from 0 to 255 as you see in modern computers. Obviously, a perfect match would be any image which contains 256^3 pixels, each with their own color, so that indeed, every pixel color is in there. However, i need this picture to be something that isn't just any pixels. It should be a real photograph if possible. I can write a java program to test whether or not there are pixels within 10 values of every possible color, but i would have to test it on some large amount of input images. So i was wondering if this is an already described property of an image, or if someone knows some great images to start off with.

Thanks!

216.173.145.47 (talk) 07:21, 9 August 2013 (UTC)

I know that what i am looking for has links to high color variance, where variance is a type of analysis done on images. I tried to google for such images and the only ideas i have thus far are perhaps images of differently colored rugs with lots of patterns or depictions on them, or images with many things in them. I guess i would like to add that the point is subtlety. The image shouldn't look like its terribly obvious that almost every color is there. Therefore, an image of a color wheel is not a good candidate. ;)

216.173.145.47 (talk) 07:37, 9 August 2013 (UTC)

mindvirus? 190.60.93.218 (talk) 14:57, 9 August 2013 (UTC)

How about this vegetable display? --.Yellow1996.^(ЬMИED¡) 16:44, 9 August 2013 (UTC)

No - that image is 694px × 364px which could AT MOST be 252,616 out of the 16 million or so possible 8-bit RGB values. However, since it's got a large area of white, I'm sure it's a lot less than that. (Besides: Do you see any blue in there? I don't!) SteveBaker (talk) 16:48, 9 August 2013 (UTC)

No, no blue at all - oh wow, how did I miss that?! Thanks for pointing that out, Steve! ;) --.Yellow1996.^(ЬMИED¡) 17:23, 9 August 2013 (UTC)

I think it's very unlikely for such an image to exist of a "natural" nature. The problem is that you'd need a 16 megapixel camera to capture it - which is do-able, but not easy. But cameras don't perfectly reproduce the scene - so it's really unlikely that you could deliberately set up a scene and photograph it and happen to get one of every shade with only 16 megapixels because there would be so much repetition. Also, some shades of magenta really don't exist in nature...so it would be hard to capture a scene with them all.

So what you need is a synthetic image - computer graphics of some kind - but even then, making a scene that doesn't look super-contrived would be exceedingly difficult.

Of course there is always stuff like THIS...which contains every possible color - but hardly counts as a natural image.

SteveBaker (talk) 16:53, 9 August 2013 (UTC)

See allRGB.—Wavelength (talk) 18:01, 9 August 2013 (UTC)

Hmmm - interesting...although they are making extensive use of "dithering" to hide the more vivid colors in darker areas. So when you look at the images closely, they are all some kind of a speckled mess. However, for the application our OP requires, they should suffice. SteveBaker (talk) 20:46, 9 August 2013 (UTC)

I wrote the algorithm to check for all colors, and so far, with 1,135,684 different colors, is the following image: Autumn Leaves

I did realize, using the same reasoning that Steve Baker used above, that smaller images just weren't going to cut it. This image does seem to hold 1/16th of the colors possible though, which is a great find. :) 216.173.145.47 (talk) 23:47, 9 August 2013 (UTC)

I just looked at allRGB too, which is a great resource! Some of these are a little too distorted for my taste, that or completely abstract. Some of them could pass though, like the more dull version of the mona lisa. :) 216.173.145.47 (talk) 23:52, 9 August 2013 (UTC)

I wonder if you check the pixel values in a colorspace other than RGB, if you will get results more similar to your intuitive expectation. Of course, changing color spaces does not change the number of pixels, but an image with a large number of distinct hues (in an HSV (hue/saturation/value) representation for each pixel) will perceptually appear to contain more distinct colors than an image with the same number of distinct pixel values, distributed across saturation and value. When you're looking at pixels that are represented as RGB values, this perceptual preference is a little harder to quantify. As an example, though, RGB=(100, 100, 200) is a dull green, and RGB=(100, 100, 255) is a brighter green; perceptually, they are nearly the same color, despite having very different numerical representations in RGB space. The same pixels represented as HSV will have an identical hue, and merely differ in saturation. It is very easy, comparatively, to construct numerical methods for HSV colors that correspond to perceptual ideals.

If you want to get very quantitative, you can construct histograms and color gamut coverage charts, and convert to a perceptual color space; and so on. Numerous different types of chroma/luma color-spaces are used by codec engineers, image processing algorithm designers, and others who need to develop software with a "human perception" application. I think your end-goal is to find images whose histograms for hue are nearly flat - in other words, images with the most number of distinct colors, with no single color dominating the image. Nimur (talk) 04:17, 10 August 2013 (UTC)

That autumn-leaves-collage image (HERE) is stored as a JPEG file - and one of the things that the file compression software does to make files smaller is to reduce the number of bits at which color values are stored. It's likely that finding the original image before it was JPEG compressed would produce MANY more unique colors. The largest version of the image that I could find with a Google Image search is the one linked to above - 5616x3744 - which is 21,026,304 pixels - more than the 16,777,216 you'd need to get every possible color. I was unable to find an uncompressed PNG or TIFF or whatever.

However, I used the GIMP image editing program's "Colorcube Analysis" tool to count the number of colors and there are 1,144,842 present in the original image (Why does this differ from the OP's answer of 1,135,684??). But then I used GIMP's "Scale Image" tool to increase the size of the image to 10,000x6,667 pixels using "Sinc" interpolation - and then re-counted the colors and ended up with 4,434,205 unique colors. (Cubic interpolation resulted in 4,389,216 and linear interpolation just 4,300,112 colors). The reason this works is that the interpolation process generates new pixels with colors that are intermediate between their neighbors...so if two neighboring pixels differed by more than one bit, the new pixels would be different from either of them.

So that's a useful trick to get more subtle in-between shades that were not present in the original image! Repeating this process is difficult because of the amount of memory required for the ungodly large images that result...but I would imagine that a modest increase over 4 million might result from quadrupling the size with interpolation. Since "Sinc" interpolation does a good job of improving the appearance of the image - it could be argued that this is still a "natural" image.

SteveBaker (talk) 14:51, 11 August 2013 (UTC)

Weird PC speaker problem

I have them for about 6 years (some model of Alteclansing). for some reason, every 10 seconds (approximately), there is a weird noise coming from both speakers equally. it lasts beteewn 3 to 7 seconds each time. reinstalling drivers didn't help, the cables seems to be connected fine, and i have win7, any ideas? before i pay a stiff for new speakers? Thank you, blessings. 09:06, 9 August 2013 (UTC) — Preceding unsigned comment added by Ben-Natan (talk • contribs)

This could be caused from speaker as well as Parasitic. Have you something that can produce waves such as mobile phone. Have you something that produce sparks every 10 seconds. 2A02:8422:1191:6E00:56E6:FCFF:FEDB:2BBA (talk) 12:40, 9 August 2013 (UTC)

If you have a cell phone near your computer try moving it further away. There is a pretty characteristic intermittent pulsing buzz/hum kind of sound (i know, terrible description) that comes over poorly shielded speakers when a cell phone communicates with the cell network nearby. Every 10 seconds sounds too frequent for that too me, but I haven't had speakers that pick up the sound since smartphones and constant data use have become more common. You could also try borrowing a different pair of speakers to see if it is just your set. Plugging in headphones to the headphone jack on the PC would confirm that the sound isn't coming from software. 209.131.76.183 (talk) 12:33, 9 August 2013 (UTC)

I agree with 2A02 and 209; It's more than likely some type of interference. Definitely try a different set of speakers and also earphones/headphones to rule out a software problem. --.Yellow1996.^(ЬMИED¡) 16:40, 9 August 2013 (UTC)

Wi-Fi wireless modems can also cause interference similar to a mobile phone. --220 of ^Borg 21:04, 9 August 2013 (UTC)

Hey u guys, I don't have any Cell\Wireless phones \ Wireless networks on the desk, i do have only Iphone 4 on the other side of the room, but when i take it off, the problem persists. could something be damaged with the Speakers themselves? it seems that if it does, it's an affect of time (because they never fell or something like that). Ben-Natan (talk) 21:28, 9 August 2013 (UTC)

It's certainly possible. After all, they're about 6 years old. To completely rule out interference then you'd have to try an alternate set of speakers; but due to the age of your current ones I think they've probably just gotten broken from being old, and probably need replacing. --.Yellow1996.^(ЬMИED¡) 00:54, 10 August 2013 (UTC)

We are floundering about a bit in the dark here Ben-Natan. Perhaps a few more details are needed, the type of speaker system i.e. Stereo (you said "both speakers"), stereo with sub woofer, etc, model of speakers, and a description of the "weird" sound. Is it a clicking, whistling or humming for example?

• Are you anywhere near an airport or other place that may have radar transmitters as I have heard 'weird' periodic sounds from radios when a radar antenna sweeps nearby. It should be possible for it to come though on PC speakers too.
  
• Is this a new problem, or has the "weird noise" always been evident?
  
• Have you been able to try ear/head-phones or another set of speakers as Yellow1996 suggested? --220 of ^Borg 13:15, 10 August 2013 (UTC)

Hey, In short, it's nothing from the above (though, i didn't try to replace the audio system cause i don't have an alternative on hand. anyways, the problem seems to pass by itself. it just seems not to happen anymore.. Ben-Natan (talk) 04:07, 15 August 2013 (UTC)

hey beotches

I have TWO facebook accounts to one email,, I MEAN, how is that even POSSIBLE , SEEING THAT THAT FORBID THAT KIND OF thing ???? — Preceding unsigned comment added by 77.35.6.72 (talk) 11:09, 9 August 2013 (UTC)

Hi! Wikipedia avoids unnecessary capitalization, not even for for emphasis. See: MOS:CAPS. Thanks! Miss Bono ^[zootalk] 12:42, 9 August 2013 (UTC)

Actually, there is no "fixed" anti-all-caps rule. It might be OK to caps a single word for emphasis if a different emphasis (i.e. bold/italic) is hard to produce. There could be devices which don't produce the " ' " single quote easily. Half a sentence is a bit excessive, though. Users who contribute to WP on a regular basis should put some efforts towards a decent device to do so ... right now, I cannot find where I read this. - ¡Ouch! (^{hurt me} / _{more pain}) 16:01, 10 August 2013 (UTC)

How do you know you have 2 accounts? 190.60.93.218 (talk) 15:45, 9 August 2013 (UTC)

Unless you did this, then there's no way. --.Yellow1996.^(ЬMИED¡) 16:36, 9 August 2013 (UTC)

Committing an edit with javascript:void(0)

On the page http://en.wikipedia.org/wiki/Wikipedia:TWA/1/Start?tour=twa1&step=7 , the a href leads to "javascript:void(0)". Yet the effect is that it makes an edit to Wikipedia under your name. How does it work? Wnt (talk) 21:27, 9 August 2013 (UTC)

The page relies on very complex javascript; when you view source for that link, you only see a small portion of the total dynamic content. Several dozen external scripts are also loaded, and they all affect the page's DOM and the link mechanics. Specifically, jQuery is used; this requires the browser support for a particular flavor of javascript, permitting the page to bind mouse events to DOM elements, implicitly triggering arbitrary javascript. The presence of javascript:void(0) is a bit of a red herring; in actual fact, it tells your browser to do nothing when the link is clicked; allowing the mouse-event to be handled by a different javascript function that was registered elsewhere. In other words, this feature is buried in the arcana of the new MediaWiki Visual Editor implementation.

Clicking on the link Wnt described will modify your user page (by submitting a "tutorial" test-edit to a sub-page). This very unexpected result would have been explained on page zero of the "tutorial" - "Wikipedia:The Wikipedia Adventure," so ordinary users who are taking the tutorial shouldn't be surprised. My personal opinion is that the new visual editor relies on bug-prone, platform-dependent, and totally unintuitive JavaScript. For an end-user, actions occur without express user intent. For a developer, the source is scattered spaghetti code with numerous unwanted side-effects. Check my recent contributions to see how often I accidentally edited pages because of these horrible scripts - as I was viewing the page-sourcecode outside of an ordinary browsing environment!

Despite the developers' best intentions, data collected by the Foundation shows that the new visual editor features drive away new users, makes them more unproductive, and makes them more likely to contribute vandalism-like edits that need to be reverted. Ironically, the new editor was advocated as an easier-to-use user-interface. Apparently, the MediaWiki Foundation now subscribes to the increasingly popular theory that non-native GUI widgets, high-contrast colors, large buttons, and WYSIWYG fonts, are necessary and sufficient to make an interface "intuitive" and "user-friendly," even when the very same user-interface performs actions without user input. Needless to say, I disagree with that philosophy. Nimur (talk) 05:20, 10 August 2013 (UTC)

I appreciate the answers but I feel like the bottom line is... no, you didn't find it. :) Anyway, I just downloaded a program called Kdiff3 I should have long ago, and found that adding the GET parameters to a redlinked page with scripts off introduces mw.loader.load([...,"ext.guidedTour",...]), changes the link rel="canonical" value to include the parameters, and alters the logout link. I assume that simply loading the extension is the important one, explained (a little) at [1]. But where the mw.loader code is defined, so I can see the ext.guidedTour source being used... I was thinking it had to be in the first script tag [2], which references it, but if it's defined there it is by some automation of jQuery that I'm missing! I ought to come back and look at this further, but if anyone happened to trace deeper in to see how this works in the meanwhile... Wnt (talk) 21:28, 10 August 2013 (UTC)

(I should note that the PHP source is at [3], but my feeling is that it is "purer" and safer to figure out how the browser is thinking ... looking at the server side adds yet more distractions) Wnt (talk) 22:21, 10 August 2013 (UTC)

To follow up, it appears that the "load.php" script as accessed by Firefox debugger contains much of the logic to do this. Between the minimized code and the very indirect way of doing things it is not an easy thing to trace, and I still haven't found where the text comes from, but I think it is doing a find for everything in class "guider_buttons" (which is set in a surrounding div), and using that to set the onclick property. The Firefox inspector function is really showing its limitations here, because apparently this doesn't show up when you inspect the button. The load.php seems to do everything from setting up the structure of the guider window to handling normal non-guided posts, so there are a lot of ways things might be getting called back and forth there. I still haven't figured out where in hell the text comes from though - it's not in the source of any of the scripts, not in the HTML, I even checked the CSS in a moment of paranoia. Wnt (talk) 19:33, 13 August 2013 (UTC)