Wikipedia:Reference desk/Archives/Computing/2015 July 24

Hoo boy. I just moved into a co-op, and I've been elected the IT guy and I've been learning everything I can to improve the house's internet speed. It hasn't been easy, and I'm basically going to walk you through what I've learned:

• There are four routers in this house, we'll call them A, B, C and D. Every router has its own network with its own password.
• I had purchased a set of four wifi boosters in an attempt to improve their speed. But I was getting very confused at the fact that the boosters were only successfully connecting to Router A. Every other router could not connect to its assigned booster, either wirelessly or with an Ethernet cable, making it impossible for me to boost their signal.
• I just learned today that only Router A has actual service from our cable company provider (TWC). It's also the only router TWC provided. The other three routers were purchased separately by the house and (I can only assume) are piggybacking off of router A. We pay a single bill to TWC for all our internet.

My question: Should I leave those other routers in? Are they actually limiting traffic from router A and freeing up bandwidth or helping in any way, or are they only making the problem worse since router A is now having to handle multiple networks? I was planning to upgrade the routers, should I upgrade all of them? --Aabicus (talk) 01:32, 24 July 2015 (UTC)[reply]

I'm guessing router A couldn't handle all the traffic alone, so they adding more routers to "help". I suspect upgrading router A so it can handle all the traffic alone would have been the better option. My suggestion is to upgrade A to a level that can handle all the traffic alone, then test performance with the rest connected, then disconnected. StuRat (talk) 04:06, 24 July 2015 (UTC)[reply]

If it's a large house, that would explain their setup. Typically, large sites have to be served by multiple access points to provide adequate range.

Typically, you design large wireless networks with multiple access points (APs, or what you refer to as routers). For example, an 802.11g wireless network, under ideal conditions outside, has a range of 300 feet. Any obstructions (e.g., walls) will lower that range significantly. So, you first perform a site survey by measuring the strength of the signal at various distances from the access point. You can just use a laptop for the survey. Once the signal becomes too weak, you place another access point at that location. But you put the AP on a different channel. For example, 802.11g has three clean channels -- 1, 6, and 11. If the first AP was on channel 1, you place the next one on either channel 6 or 11. That way, they don't interfere with each other. So, my guess is that's what someone already did. They put four APs at your site because it's too big to be served by a single AP. They put each one on a different wireless "network" so it could have a different channel.

However, you don't have to create different networks to get a different channel. You can give them all the same name (i.e., SSID) but just change the channel. Keep the password the same for all the APs. That's how I would set it up.

Having said that, the channel numbers and ranges I gave above are only valid for 802.11g networks. Different 802.11 standards different ranges and channels, so I'd need to know the model number of the APs to give you more specific information. It'd also help to know the IP subnet you're using along with the wireless settings the APs are using (namely, the channel, 802.11 standard, and frequency band -- 2.4 or 5 GHz). Also, are all the APs on the same subnet? You can find this information by logging into the Web page for the AP. I'm also guessing the APs are connected to each other using Ethernet cables? Typically, the AP will have a single router port and multiple switch ports. They're usually colored differently. The router port is used to connect different subnets and the switch ports connect to the same subnet. Are they just cabled through the switch ports or the router ports? I would just connect them using the switch ports. Connect Router A to the cable modem using its router port but connect each AP to Router A using the switch ports. Then, make sure each AP is on the same subnet. Then, disable DHCP on all the APs except for one. Then, reboot every end-user device and AP at the site so they get new, consistent IPs. That will simplify things and make everything operate faster.—Best Dog Ever (talk) 09:20, 24 July 2015 (UTC)[reply]

If I remember rightly from last Saturday, both the nearer-to-camera tree and farther-from-camera group of trees (in the center, not the stuff farther away on the other side of the river) were the same color. Can we guess how they ended up in different colors in this image? For this photo, I used the 55-200 mm zoom lens for my Nikon D3200, zoomed in all the way, and it wasn't taken through a window or other medium. With other images on the same settings, e.g. File:Robert Buckles Barn from the road.jpg, it's not produced the same result, so I doubt that it's a systemic problem with the camera. Nyttend (talk) 04:14, 24 July 2015 (UTC)[reply]

I'm not sure that this is really a matter for the Reference Desk, but to my eye it just looks as if the air is a bit hazy. If you look at groups of trees at progressively larger distances from the camera, the farther they are the duller the colors are. --65.94.50.73 (talk) 04:31, 24 July 2015 (UTC)[reply]

Also bluer, due to Raleigh scattering. The camera has accurately recorded the difference in colour (I'm a Brit) that was actually there. However, when you were looking, your brain was automatically processing the "raw data" from your eyes to correspond better with your "mental model" of the world. It does this all the time in several differents ways, which is why, for example, you don't usually perceive the more distant of two similarly sized objects as "smaller", even though its image is smaller on your retina. Our article Visual perception hopefully gives some leads relating to this subject. {The poster formerly known as 87.81.230.195} 212.95.237.92 (talk) 13:22, 24 July 2015 (UTC)[reply]

Also, when you say the trees were the same color, did you go closer to them to determine this ? If so, the apparent colors of the trees would then be the same, without as much hazy air in between, as in the photo.

As for the Raleigh scattering, that is exemplified in the America the Beautiful line "For purple mountain majesties". (The mountains aren't really purple, but appear that way from enough of a distance.)

One way the camera actually could change the color of just certain trees is if the far trees were out of focus, and thus the leaf color was blended with the color of the background. However, it doesn't look to me like that happened here. StuRat (talk) 15:40, 24 July 2015 (UTC)[reply]

Couldn't go closer (thus the poorer-quality distant photo), because it's on the grounds of a power plant, and this is the closest the road goes to the site. I suppose the Rayleigh scattering is the right answer; it never occurred to me that the brain would "fix" it, but your statement makes complete sense. And of course the camera doesn't know to "fix" such a thing. Here I figured it was something with the digital camera; that's why the question came here, not WP:RDS. Thanks for the help! Nyttend (talk) 16:46, 24 July 2015 (UTC)[reply]

It's not impossible to fix this digitally. The camera would need to be able to determine depth at various points in the frame, measure the color change with depth, and then compensate for it. Of course, this wouldn't be perfect, as the absolute color might actually change for items farther away. Also, I think many people like the natural "purple mountains" effect. StuRat (talk) 16:59, 24 July 2015 (UTC)[reply]

• That's purple mountains' majesty. —Tamfang (talk) 08:43, 25 July 2015 (UTC)[reply]

• Not according to America the Beautiful#Lyrics and America the Beautiful#Idioms. StuRat (talk) 01:23, 26 July 2015 (UTC)[reply]

• Well, damn. I'll still maintain that the version I learned makes more sense. —Tamfang (talk) 18:00, 26 July 2015 (UTC)[reply]

I agree that this is most likely either Raleigh scattering or Mie scattering - primarily because you mention a the distance disparity between the two trees. But there are other ways this could theoretically come about. Suppose one tree had leaves that reflected only green light - and the other had some small amount of reflectivity in red light as well as in green. If you viewed the scene in white light - at midday, perhaps - the difference in color between the two trees might be too subtle to notice. But at dawn and at dusk, when the sunlight is very reddish/orange - the tree that reflects no red light would appear significantly darker than the one that reflects a tiny bit of red so that small difference in color would be greatly magnified. There are other possibilities too - the average orientation of the leaves to the direction of the sunlight might be different in one than in the other and that would produce dramatic differences depending on the angle of the sunlight when the photo was taken. If one tree had very shiney leaves and the other was relatively dull, then the difference might be hard to see on a cloudy day, but very evident when it's sunny. There are many possibilities. SteveBaker (talk) 03:01, 26 July 2015 (UTC)[reply]

Looking at the Y channel of the pic in the CMYK color space the foreground tree in question is very bright, nearly all the background black. Is that Raleigh scattering?—eric 04:08, 26 July 2015 (UTC)[reply]

More or less. The Rayleigh scattering (note spelling, everyone) makes distant objects bluer, and yellow is complementary to blue. I don't know much about CMYK, but when you're on the blue side of white, it will probably use little or no yellow ink (unless it's a very dark blue). -- BenRG (talk) 20:40, 26 July 2015 (UTC)[reply]

The camera is using color balance. During the night the visible spectrum is cut by the electric light. Light bulbs have missing blue light. Compact fluorescent lamps and other fluorescent lamps not not give enough red light. The digital camera does the color balance by gamma correction and looking up in the pictures histogram for the lightest dots (equals pixels). There are several parameters more used for good gamma correction. Some cameras allow to turn of the color balancing or saving raw images (uncompressed bitmaps from the sensor) to the flash memory. Gama correction is always a lost of quality. By amplifying the missing blue of lightbulb lighted pictures, the blue channel of the pictures becomes noisy. For that reason the camera takes multiple pictures to cover the sensors noise. This disallows to display fast moving objects clear and sharp when even focussed correctly. Settings decide between noise colors and sharp captured quick moving objects of the picture. A simple trick to help the color balance algorithm is to put a white paper or similar beside the objects an keep the paper in the view of the camera which is being cut out or removed by photo editing, later. --Hans Haase (有问题吗) 08:51, 27 July 2015 (UTC)[reply]

All very interesting, but how does it relate to the Q ? StuRat (talk) 20:00, 27 July 2015 (UTC)[reply]

Are there reliable statistics about computer literacy? For example, how many people can navigate the web, send and read emails, write a simple text with a word processor, make a table with a spreadsheet, install a program, configure a mouse, and so on? The corresponding article here in wk is kind of thin, btw. I'd like to find stats for Europe, US and Japan mainly. --YX-1000A (talk) 08:29, 24 July 2015 (UTC)[reply]

It is very difficult to tell if people can use what they have. It is easier to tell what they have and assume that if they are paying for something, they know how to use it. (As an analogy, I run reports on medications. I don't know what people take. I know what their insurance paid for and I know what the doctor prescribed. I assume that most people take some of what the insurance pays for and, if it is over-the-counter, they take some of what is prescribed. When working with millions of patients, I ignore the specifics about any one patient. Statistics over anecdotes.) Internet access is rather easy to track because someone has to pay for it. You can see U.S. statistics on Internet use here. UK internet statistics are here. I didn't find Japanese statistics, but I found Japanese websites that appear to be Internet statistics. If I could read Japanese, I could be certain. 209.149.113.45 (talk) 13:10, 24 July 2015 (UTC)[reply]

Schools and colleges might be a good way to check computer literacy for kids and young adults, as that is one factor on which they are likely to report these days. Employment agencies might be another source of info for adults. StuRat (talk) 17:01, 24 July 2015 (UTC)[reply]

Here is a "Survey of Literacy, Numeracy, and ICT Levels in England" (from 2011), published by the Department for Business Innovation and Skills (a ministry of the UK government). It's 425 pages, and computer literacy is only one of its topics, but there might be something useful in there. AndrewWTaylor (talk) 19:29, 24 July 2015 (UTC)[reply]