5 ms·

> Almost anybody can say "That was a violin" or "That was a female voice" or "That was a guitar" with basic training. That's about timbre, not pitch.

jacquesm 5y ago
What is known as timbre is technically the degree to which certain harmonics are present or not. The most interesting case of this is the missing fundamental, which we apparently re-create in our heads to hear it even when it isn't really present! https://en.wikipedia.org/wiki/Missing_fundamental https://en.wikipedia.org/wiki/Missing_fundamental This really threw my for a loop while building my mp3->midi convertor.
- Grustaf 5y ago
  I know what timbre is, that's my point. The difference between a violin and a piano is not about frequency.
  - ogma 5y ago
    And their point is that timbre is determined by the intensity of the overtones, which are frequencies.
srcreigh 5y ago
That's my point. I believe red vs blue is more like violin vs voice than A440 vs Exyz.
- jacquesm 5y ago
  But it isn't. To make it a bit more specific: Red vs Blue is ~700 nm vs ~475 nm (about because they are ranges rather than specific frequencies where most people who are not vision impaired will agree something is either red or blue). Violin vs voice is more like 'triangular wave form with f, 2f, 3f, 4f, etc as the harmonics and voice would be 'mostly sinusoidal waveform with a bunch of vocal 'chords' acting as strings each of them with a sligthly different base pitch, with those same harmonics. But if you were to compare for instance to a reed instrument the harmonics would look completely different. Some singers by the way are capable of controlling their vocal chords in such a way that they can create rising and falling pitches at the same time. https://www.youtube.com/watch?v=vC9Qh709gas https://www.youtube.com/watch?v=vC9Qh709gas
  - srcreigh 5y ago
    Although I concede and agree with you on a basic level that base frequency is more like color than timbre, there are some other interesting factors. One. Almost all instruments have different timbre depending on the pitch. At least at large scales, your voice's deepest note does not have the same timbre as your midrange, or your highest note. Similarly with pianos. I wonder if this is also true on a micro level between A and B on a piano? Two. As I mentioned above, perfect pitch folks _don't recognize sine waves as well as piano notes_. Why? That's very curious. In any case, I was also going to mention that musical notes are interesting because they loop. A is 440Hz and 880Hz. I was expecting to find something like 2x blue frequency = yellow, which would highlight a difference between color and sound. However, interestingly, that is not the case. The entire visible spectrum of light is within one "octave" of frequency. Fascinating... :)
    - jacquesm 5y ago
      Yes, it is fascinating, highly recommended book about sound: "The Acoustical Foundations of Music" by John Backus.
    - akomtu 5y ago
      7 octaves above cover the UV spectrum and 7 octaves below cover the IR spectrum. I guess our eyes could see 14 octaves with 50 instead of 3 sensors per pixel. A more practical solution is a separate device, I mean organ, that works like our ear but for em waves: just one "pixel" but with lots of sensors and complex postprocessing to detect harmonics. This way we could hear em waves. Such a device can be built, actually.
      - jacquesm 5y ago
        Not octaves. An octave is a doubling of frequency. When you go seven octaves below 'red' you are much, much lower than IR and when you go 7 octaves above blue you are way higher than UV.
        akomtu 5y ago
- Grustaf 5y ago
  It's not though, red and blue have specific frequencies, just like a 440 Hz sine has.
  - jacquesm 5y ago
    No, red and blue have a 'range' where most people will agree on what's red and what's blue. They are not exact frequencies but frequency bands that have been culturally defined. You can most easily see this in green, there are 100's of 'greens' but we call all of them green.
    - Grustaf 5y ago
      Sure, but we are still talking about bands of _frequencies_, not sets of harmonics (although there are colours that are not pure). So it would correspond to identifying 440 Hz with a tolerance of N Hz. And we are not talking about colour naming, we are talking about colour perception. So the situation would be "here's a particular green colour, please find a patch from this heap that has the same colour." I mean, just because the word "green" is very broad doesn't mean we can't _see_ the difference.
      - jacquesm 5y ago
        Of course we are not talking about harmonics, the first overtone of 'red' would be a bit above ultra-violet and invisible. And as for green, yes that is the best color to do that test with because we have the biggest discriminatory capability for green. And most people would be able to distinguish with a large degree of accuracy an increasingly high frequency shade of green given similar intensity. But once you start varying the intensity and the hue at the same time I think people will get confused quite rapidly as to which shade has the higher frequency hue. Color is much more 'loose' than sound, that's why we 'tune' our instruments and why painters don't necessarily need to 'tune' their palettes so precisely to be able to make something that looks harmonious.
        Grustaf 5y ago
        Sure, it's complex, but the basic point still stands. Most people will probably be able to accurately identify 10-20 different colours, if we fix the luminosity etc and they are given cards with each colour. When it comes to sound though, even many musicians won't be able to find even a single absolute note, even if we fix the timbre, intensity etc. So colour and sound are definitely fundamentally different, which I don't find very surprising, there are few situations in the wild where it would help us to be able to distinguish absolute frequencies, timbre is more important.

jacquesm 5y ago

What is known as timbre is technically the degree to which certain harmonics are present or not. The most interesting case of this is the missing fundamental, which we apparently re-create in our heads to hear it even when it isn't really present! https://en.wikipedia.org/wiki/Missing_fundamental https://en.wikipedia.org/wiki/Missing_fundamental This really threw my for a loop while building my mp3->midi convertor.

Grustaf 5y ago
I know what timbre is, that's my point. The difference between a violin and a piano is not about frequency.
- ogma 5y ago
  And their point is that timbre is determined by the intensity of the overtones, which are frequencies.

srcreigh 5y ago

That's my point. I believe red vs blue is more like violin vs voice than A440 vs Exyz.

jacquesm 5y ago
But it isn't. To make it a bit more specific: Red vs Blue is ~700 nm vs ~475 nm (about because they are ranges rather than specific frequencies where most people who are not vision impaired will agree something is either red or blue). Violin vs voice is more like 'triangular wave form with f, 2f, 3f, 4f, etc as the harmonics and voice would be 'mostly sinusoidal waveform with a bunch of vocal 'chords' acting as strings each of them with a sligthly different base pitch, with those same harmonics. But if you were to compare for instance to a reed instrument the harmonics would look completely different. Some singers by the way are capable of controlling their vocal chords in such a way that they can create rising and falling pitches at the same time. https://www.youtube.com/watch?v=vC9Qh709gas https://www.youtube.com/watch?v=vC9Qh709gas
- srcreigh 5y ago
  Although I concede and agree with you on a basic level that base frequency is more like color than timbre, there are some other interesting factors. One. Almost all instruments have different timbre depending on the pitch. At least at large scales, your voice's deepest note does not have the same timbre as your midrange, or your highest note. Similarly with pianos. I wonder if this is also true on a micro level between A and B on a piano? Two. As I mentioned above, perfect pitch folks _don't recognize sine waves as well as piano notes_. Why? That's very curious. In any case, I was also going to mention that musical notes are interesting because they loop. A is 440Hz and 880Hz. I was expecting to find something like 2x blue frequency = yellow, which would highlight a difference between color and sound. However, interestingly, that is not the case. The entire visible spectrum of light is within one "octave" of frequency. Fascinating... :)
  - jacquesm 5y ago
    Yes, it is fascinating, highly recommended book about sound: "The Acoustical Foundations of Music" by John Backus.
  - akomtu 5y ago
    7 octaves above cover the UV spectrum and 7 octaves below cover the IR spectrum. I guess our eyes could see 14 octaves with 50 instead of 3 sensors per pixel. A more practical solution is a separate device, I mean organ, that works like our ear but for em waves: just one "pixel" but with lots of sensors and complex postprocessing to detect harmonics. This way we could hear em waves. Such a device can be built, actually.
    - jacquesm 5y ago
      Not octaves. An octave is a doubling of frequency. When you go seven octaves below 'red' you are much, much lower than IR and when you go 7 octaves above blue you are way higher than UV.
      - akomtu 5y ago
Grustaf 5y ago
It's not though, red and blue have specific frequencies, just like a 440 Hz sine has.
- jacquesm 5y ago
  No, red and blue have a 'range' where most people will agree on what's red and what's blue. They are not exact frequencies but frequency bands that have been culturally defined. You can most easily see this in green, there are 100's of 'greens' but we call all of them green.
  - Grustaf 5y ago
    Sure, but we are still talking about bands of _frequencies_, not sets of harmonics (although there are colours that are not pure). So it would correspond to identifying 440 Hz with a tolerance of N Hz. And we are not talking about colour naming, we are talking about colour perception. So the situation would be "here's a particular green colour, please find a patch from this heap that has the same colour." I mean, just because the word "green" is very broad doesn't mean we can't _see_ the difference.
    - jacquesm 5y ago
      Of course we are not talking about harmonics, the first overtone of 'red' would be a bit above ultra-violet and invisible. And as for green, yes that is the best color to do that test with because we have the biggest discriminatory capability for green. And most people would be able to distinguish with a large degree of accuracy an increasingly high frequency shade of green given similar intensity. But once you start varying the intensity and the hue at the same time I think people will get confused quite rapidly as to which shade has the higher frequency hue. Color is much more 'loose' than sound, that's why we 'tune' our instruments and why painters don't necessarily need to 'tune' their palettes so precisely to be able to make something that looks harmonious.
      - Grustaf 5y ago
        Sure, it's complex, but the basic point still stands. Most people will probably be able to accurately identify 10-20 different colours, if we fix the luminosity etc and they are given cards with each colour. When it comes to sound though, even many musicians won't be able to find even a single absolute note, even if we fix the timbre, intensity etc. So colour and sound are definitely fundamentally different, which I don't find very surprising, there are few situations in the wild where it would help us to be able to distinguish absolute frequencies, timbre is more important.