Wikipedia:Reference desk/Archives/Miscellaneous/2023 February 1

From Wikipedia, the free encyclopedia
Miscellaneous desk
< January 31 << Jan | February | Mar >> Current desk >
Welcome to the Wikipedia Miscellaneous Reference Desk Archives
The page you are currently viewing is a transcluded 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 1[edit]

Handwriting[edit]

With regards to handwriting, is there a correct ratio between the height of upper and lowercase letters. I vaguely remember being taught that there was but can't find anything on the internet. --Ykraps (talk) 06:25, 1 February 2023 (UTC)[reply]

When I was taught handwriting in school, we used ruled handwriting guide paper, as seen here. Most images of such paper I find on the Web allot one extra x-height to ascenders and descenders. I think most majuscules also had to fit between the baseline (line 2 counting from the top) and the top line, making cap height equal to ascender height, the height of b, d, h, and so on. (As I was taught to write them, the cursive capitals G, J and Y also had descenders, and the Q had a tail that did not rest on a preprinted line, like seen here.)  --Lambiam 08:50, 1 February 2023 (UTC)[reply]
Nothing official (in most places), but schools can select and teach a particular handwriting style that demands a particular ratio. I remember that ruled handwriting guide paper. At my school, the ascenders and decenders added about 3/2 of an x-height and the capital letters were as high as the ascenders. PiusImpavidus (talk) 10:50, 1 February 2023 (UTC)[reply]
Because my father was in the British Army and we moved frequently, I had changed schools six times by the age of ten. Each school taught a different style, so my handwriting suffered. After then I deliberately ignored further 'advice' and evolved my own style (such as Є,є for E,e). {The poster formerly known as 87.81.230.195} 90.221.194.253 (talk) 12:46, 1 February 2023 (UTC)[reply]

Thank you all for replying. Perhaps I should've explained that there is nothing wrong with my handwriting, I am making handwritten invitations and will be using temporary ruled lines for extra neatness so I thought I might as well get them in the 'right' place. --Ykraps (talk) 18:20, 1 February 2023 (UTC)[reply]

In that case, pick a specific standard script and follow what its rules are. --Jayron32 15:36, 2 February 2023 (UTC)[reply]
  • The proviso is that the kind of script you're doing isn't really "handwriting" in the usual sense. It's closer to calligraphy. No matter what template teachers use to teach students how to form letters, pretty soon their individual personalities come into play and their writing becomes as unique as their fingerprints. And it changes over time. -- Jack of Oz [pleasantries] 19:24, 3 February 2023 (UTC)[reply]

Frameless video[edit]

If I understand correctly, modern video cameras shoot in a sequential still frame manner that instantly combines all still frames into a video. Instead of capturing individual frames, is it technically possible to shoot in a continuous stream manner, similar to how a dictation machine records sound? 212.180.235.46 (talk) 21:48, 1 February 2023 (UTC)[reply]

That would require an infinite number of individual frames. ←Baseball Bugs What's up, Doc? carrots→ 03:38, 2 February 2023 (UTC)[reply]
If a dictation machine is digital, it also records data in the form of chunks of audio data, in a manner similar to a CD. However, if I understand correctly, if any media recording device operates in the analogue realm then one could understand the stream of data to be continuous as you describe, and this would be true of videotape (say VHS) or audio tape. --Ouro (blah blah) 07:57, 2 February 2023 (UTC)[reply]
Videotape has individual frames, just like the films or TV programs they record from. ←Baseball Bugs What's up, Doc? carrots→ 08:15, 2 February 2023 (UTC)[reply]
Oh does it now? Well then, strike all of that.--Ouro (blah blah) 08:49, 2 February 2023 (UTC)[reply]
It is theoretically possible to record separate analogue continuous streams for each of the R, G and B values of each pixel, using e.g. a 3-CMOS camera. For a (now) modest image resolution of 3840 x 2160 pixels, this would require 24,883,200 separate tracks, each with its own recording head, to record these streams in parallel. This may be practically impossible. It may already be impossible to get the continuous signals out of the camera, since this also requires 24,883,200 separate signal-carrying wires connected to the image sensor.  --Lambiam 10:03, 2 February 2023 (UTC)[reply]
The thing you want to record is 3 dimensional: width, height, time. If the recording medium is 2 dimensional (like analogue film), you have to chop it up somehow. The normal method is to take a 2D image, wait 1/24 of a second and take the next 2D image, chopping it up in time. But that's not the only way; it's just the most practical and something people already knew from photo cameras. For example, you could also put a special lens at the focal plane of your film camera that splits the image into parallel horizontal lines, use some mirrors to pack those lines into one long line and project that on the film. Move the film continuously; no shutter required. The resulting recording will be chopped up in the vertical direction, but not in horizontal or time. When you film something with narrowly spaced horizontal lines, you'll get some (spatial) aliasing, but backward spinning wagon wheels (temporal aliasing) will be gone.
Or you could use a narrow vertical slit in your camera that moves sideways. When it has moved over part of your field of view, the next slit begins the next scan. The result is that your recording is chopped up diagonally: partly horizontal, partly in time. Each frame as you see it on the resulting film is a diagonal band, with the horizontal position giving the x coordinate and the vertical giving the y-t coordinate. You may already know the resulting deformation/aliasing; see rolling shutter (usually a horizontal slit).
If the recording medium is 1 dimensional (like an analogue television signal), you have to chop is up twice. Again, it doesn't have to be exactly in horizontal, vertical or time; you can take a mixture. Analogue television uses a single pixel that scans the image in a sawtooth pattern. No shutter required, just a pair of rotating mirrors. Different parts of the image get recorded at different times, but with a CRT television, they also get shown at different times, so that's perfect. PiusImpavidus (talk) 11:43, 2 February 2023 (UTC)[reply]
Speaking about chopping up, see Nipkow disk.  --Lambiam 15:26, 3 February 2023 (UTC)[reply]
We don't seem to have articles for Single-pixel camera or Single-pixel imaging. But here's an article in Nature: Principles and prospects for single-pixel imaging. This turns it into a two-dimensional problem, where there doesn't have to be a hard boundary between frames, and then yes, you'd get a continuous stream of data from which you could use any arbitrary amount to assemble a still image.  Card Zero  (talk) 19:45, 3 February 2023 (UTC)[reply]
See also Space-filling curve. Copies of the Hilbert curve can be glued together to get a closed curve with an arbitrary aspect ratio.  --Lambiam 11:23, 4 February 2023 (UTC)[reply]
Here is a video that does a really good job explaining how it is done. --Jayron32 14:20, 6 February 2023 (UTC)[reply]
A nifty feature of Hilbert scanning is that you can double the resolution of the sender or the receiver independently of the other, without losing compatibility. [1]Tamfang (talk) 22:02, 6 February 2023 (UTC)[reply]