Half a croissant, on a plate, with a sign in front of it saying '50c'
h a l f b a k e r y
You gonna finish that?

idea: add, search, annotate, link, view, overview, recent, by name, random

meta: news, help, about, links, report a problem

account: browse anonymously, or get an account and write.



Colour hearing

Translates light frequencies into sound frequencies.
  (+16, -1)(+16, -1)
(+16, -1)
  [vote for,

The human eye has 3 types of colour receptors, centred around red, green, and (despite what you may have heard) violet. Therefore a pure frequency of (say) orange light is indistinguishable from a certain blend of pure red and green light.

In contrast, the ear is sensitive to a continuum of frequencies, and can make sense of several frequencies mixed together. A mixture of (say) 200 Hz and 250 Hz sounds different from any intermediate frequency. In other words, you can hear chords in sound but you can't see chords in light.

Place a spectroscope over the end of a monochrome digital camera, so the camera sees a spectrum of whatever the spectroscope is pointing at. The camera is connected to a computer, which extracts a line of pixels corresponding to the spectrum. These values are sent to FFT software to convert from the frequency domain to the waveform domain. The output is sent to a soundcard.

Point it at red laser light and you hear a low frequency sine wave. Blue laser: high frequency sinewave. Real white light (eg daylight) (no surprise): white noise. But fake white light (eg on a computer monitor): a 3-note chord, corresponding to the blend of R, G, and B.

spidermother, Feb 05 2006

frequency translator frequency_20translator
Has a bat detector link [Dub, Feb 06 2006]

Audible light sources http://www.keelynet.com/audbin.htm
Looking at the modulation of light sources rather than an audible representation of their spectrum. [st3f, Feb 06 2006]

Urban Dictionary Entry for Arachnophile http://www.urbandic...p?term=Arachnophile
FAO lintkeeper2 [Dub, Feb 06 2006]

_22seeing_22_20bird_20colors not really related, but reminded me of this. [xaviergisz, Jul 21 2006]

Wikipedia: Neil Harbisson https://en.wikipedi...wiki/Neil_Harbisson
Mentioned in my anno. The prior artist [notexactly, May 28 2019]

Amasci: UFO Scope http://amasci.com/freenrg/ufoscope.html
Mentioned in my anno. The article by [wbeaty], a shorter version of which is linked above by [st3f] [notexactly, May 28 2019]

Wikipedia: Metamerism (color) https://en.wikipedi.../Metamerism_(color)
Mentioned in my anno. The wavelengths vs. colors one [notexactly, May 28 2019]

Wikipedia: Fluorescent tube § Phosphors and the spectrum of emitted light https://en.wikipedi...um_of_emitted_light
Mentioned in my anno. The obvious one [notexactly, May 28 2019]


       I'll give you a + for this, even though 450 nm sounds blue to me.
ldischler, Feb 05 2006

       "What's color like?"   

       "Well, it's like G flat."
notmarkflynn, Feb 05 2006

       Cool, a synesthesiasizer.   

       //These values are sent to FFT software to convert from the frequency domain // sp. "IFFT"
//Real white light (eg daylight) (no surprise): white noise. But fake white light // Bollocks.
AbsintheWithoutLeave, Feb 05 2006

       //sp. "IFFT"// Overly picky, and wrong: //A fast Fourier transform (FFT) is an efficient algorithm to compute the discrete Fourier transform (DFT) and its inverse.// Wikipedia definition, note last 3 words, go there and contest it if you will. Besides, your objection is like saying "you don't use a wordprocessor to check your spelling, you use a spell checker" Any set of FFT libraries that doesn't carry out the inverse transform isn't worth much.
//Bollocks.// Them's fightin' words. I just now, with the aid of a simple CD, compared the spectrum from my LCD monitor showing just white (all 3 colours at maximum values) with that from an incandescent lamp. In the case of the monitor, I can clearly see 3 separate colour regions centred around red, green and blue. In the case of the lamp, the spectrum is continuous. The difference would be starkly audible; after all, we distinguish between vowel sounds using far more subtle differences in spectral content. The whole point is that the ear is a better frequency analyser than the eye. Get your facts straight!
Besides, who are you calling defensive? ;-)
spidermother, Feb 06 2006

       If you're tone deaf, are you partially color blind?
Shz, Feb 06 2006

       Hmm... I suppose so. If you were deaf, on the other hand, you could run the sucker in reverse and create a display which changed colour according to ambient sounds. It would be of woefully low resolution compared to hearing, but better than nothing. You might just be able to distinguish a foghorn from a whistle.   

       BTW, what got me thinking along these lines was someone asking why you can't see chords in light. Looking at it from a design perspective, if you saw with the equivalent frequency resolution and dimensionality of hearing, you would be swamped with information, as you would have many more dimensions to deal with - vision is already roughly seven dimensional as things stand! There! That's an assertion to get stuck into.   

       I suppose the concentration of colour perception per se in the centre of the field of vision is another way of reducing the information load. If I were to design a visual system I might want to include an extra, higly focussed, eye that did a sophisticated spectral analysis of whatever small region it was pointed at. It might just be possible to build such a sense in, using a network of electrodes in the brain linked to an analysing spectroscope. It would probably need to be installed in infancy to be incorporated as a fully fledged extra sense. Then you would need names for all those hundreds of extra colours. Damn. Ethical issues. Hence the piggybacking onto the only decent spectral analyser we come with.
spidermother, Feb 06 2006

       Sorry [spidy], but I'd be most surprised if Abs has it wrong... Trust me on this, he's probably forgotten more about Fourier Transforms than most people will ever know.   

       If you're already looking at a spectrum, you use an inverse transform to turn it back to a signal. Thus his I(F)FT.   

       Another way of doing it is "just to" shift and transform the frequencies...Like a super-duper bat box.[+]
Dub, Feb 06 2006

       Interesting applications for the visually impaired.
wagster, Feb 06 2006

       If you regularly swept the field of view, then the audio signal would change. I wonder if the brain could be trained to distinguish objects with these repetitive sounds?
Ling, Feb 06 2006

       I wouldn't be surprised - Abs used to debug some of his DSP code with a radio. (Dribble-slurp - Sorry master, don't beat me for suggesting your code had bugs!)
Dub, Feb 06 2006

       Cute idea. I can't see an appliccation for it, though and I wish I could.
st3f, Feb 06 2006

       [Dub], //Thus his I(F)FT// <flogging a dead horse> DFT/IDFT, FT/IFT, etc., are pairs in technical usage, but "FFT software" can surely be used for IFT calculations. I'll try to use less ambiguous terms in future. </f.a.d.h.>   

       //"just to" shift and transform the frequencies...Like a super-duper bat box.// Wait a moment, are you implying that we already have the light information in the waveform domain, and we transpose the frequencies down to the audible region? Can you _do_ that? What's a bat box? I just find artificial nesting structures for winged mammals. Do you mean you get bat-like super hearing?   

       [wagster] //Interesting applications for the visually impaired// Yes. Especially given that the auditary world is so much more present for long-term blind people than for sighted. I have a friend who was born with no sight at all. She can hear the shape of her surroundings - doorways etc. have a distinctive sound. If she received additional sound cues according to the brightness and colour of whatever she was facing (I'm imagining the device built into glasses or otherwise worn) it would be useful, or at least interesting.   

       [Ling] I think (by analogy with listening to music) that if you swept along a single dimension (a straight line) the brain could keep up, but attempting to build up a 2-D scene with parallel sweeps would be like trying to make sense of many peices heard simultaneously - or worse, matching the finescale structures of pieces heard in rapid succession. Quite an ask. And keeping up with a changing scene in 2D? Forget it, I'd say. Too much information.
spidermother, Feb 06 2006

       (f.d.h Reminds me of an old joke about sadomasicism, necrophilia and bestiality.)   

       A bat box/bat detector is used to shift the ultrasonic sounds bats use, down to a range the human ear can hear (search the bakery/web [link->link] ). So I imagined something that went above ultrasound e.g. IR->UV -> 50Hz-20KHz (probably non-linear!)
Dub, Feb 06 2006

       Ah! [Dub], now we're on the same wavelength. You see, the technology to do analog to digital sampling and signal processing at radio frequencies is fairly new, and I thought you were assuming the ability to do this at visible light frequencies, and that was way too scary.
The range of sound frequencies could be scaled to fit within a comfortable hearing range for whatever range of light wavelengths you were interested in, by applying a suitable transform to the raw spectral data. Then you would have a kind of spectral zoom function.
spidermother, Feb 06 2006

       ...and thus "Sing the Blues"
Dub, Feb 06 2006

       Change your username, and I'll read your idea.
blissmiss, Feb 06 2006

       I'm not too up on my electronics or how much processing power FFT's take, but this sounds fairly bakeable to me, not to mention interesting.
wagster, Feb 06 2006

       Colour Hearing?
Sounds more like an LSD trip than high tech...
neutrinos_shadow, Feb 06 2006

       //Change your username// [blissmiss] Please explain. Is my username offensive somehow? It's not meant to be. It's after a song by one of my favourite local musicians, Baterz (deceased):   

       I came home one drunken Friday evening to find
My father crying on the porch
So I said, "Hey what's up, Dad, your computer eaten all your software again?"
And he replied "No, I'm afraid it's much worse than that, son, your mother has turned into a spider"

       [wagster] I use ZynnAddSubFX, an (I)FFT based soft synth. Based on what I throw at it, I think this would be feasable on a modern PC. I've also been looking at some high-speed, high bitdepth monochrome digital cameras that would be just the ticket. I agree - quite bakable. I might try cobbling together a few soundbites.
spidermother, Feb 06 2006

       Nothing wrong with your username, just [bliss] dislikes spiders. I, however, enjoy them.

       I haven't figured your idea out all the way. I used to have embryonic thoughts about colourful fish that lived out of visual depth & perhaps their colours gave off sound waves blah blah blah but I never did quite get my head around it. Same thing going on here, I'm afraid.
lintkeeper2, Feb 06 2006

       [lintkeeper2] //I enjoy them// - "I like (or dislike) them" sounds reasonable. "I enjoy them" sounds downright kinky! [link]
Dub, Feb 06 2006

       Two things I've always wondered: if red is a very long wavelength and violet is much shorter, then what wavelength is burgundy, and why do red and violet look so similar? Is there something with those wavelengths that borders on the 'fading to infra / ultra' response from our vision or something?
RayfordSteele, Feb 10 2006

       + just because it's interesting to contemplate.   

       It seems that the visible spectrum of a scene reaching your spectrometer would be quite heavily influenced by the means of illumination.   

       In sunlight, most scenes would produce fairly broad bands of overlapping color spectrum, and correspondingly broad bands of noise. In fluorescent light, you would get narrower response, depending on the interaction of the illuminating spectrum with whatever is reflected back to the spectrometer. (Anybody recall how many spectral lines in a fluorescent tube?)   

       You are quite right that most modern CRTs use relativly narrow band phosphors, so Illuminating a scene with "white" light from a CRT (actually RGB) would produce some sort of audible triad.   

       RGB (and possibly other light color/frequency lasers) used as the illuminating source would give nice narrow peaks in the spectrum, and could be arranged to form varying chords composed of whatever audio frequencies are chosen to track the source light colors.   

       As I said, it's pretty interesting, but what practical applications are there? Maybe quarter-baked.
csea, Feb 10 2006

       There's a well-known prior art example for this: Neil Harbisson's antenna attached to his skull. [link]   

       He's colorblind (compared to most humans), so it turns colors into sounds for him to hear (among other things). But it's chromatic rather than spectral, so it gives him similar color perception to most humans, rather than spectroscopic perception like this idea.   

       // Audible light sources // link   

       I read the longer version of that article on [wbeaty]'s own site [link] many years ago. It was how I found Instructables, actually, by looking for an alternative amplifier circuit because, I think, I found his too complex (LOL). That was back when that site had fewer than 800 projects posted. It was more user-friendly (and a more friendly company) in those days, I think, or maybe that's just rose-colored memory glasses.   

       // Two things I've always wondered: if red is a very long wavelength and violet is much shorter, then what wavelength is burgundy, and why do red and violet look so similar? Is there something with those wavelengths that borders on the 'fading to infra / ultra' response from our vision or something? //   

       Some perceptible colors cannot be produced by single wavelengths of light. And, except when you're looking at the output of a laser or a monochromator, all colors you see are almost certainly the result of mixing of multiple wavelengths (sometimes smooth distributions and sometimes spiky). See [link].   

       // Anybody recall how many spectral lines in a fluorescent tube? //   

       Depends. See [link].
notexactly, May 28 2019


back: main index

business  computer  culture  fashion  food  halfbakery  home  other  product  public  science  sport  vehicle