Reproduction of timbre

Discussion in 'Sound Science' started by old tech, Apr 23, 2019.

1. This is probably a naïve question, but one for audio science. If reproduction of sound consists of amplitude and frequency (and frequency is the same as time), then what accounts for differences in timbres for notes that are the exact same amplitude and frequency?

2. There would be no differences in timbre for notes that are the same amplitude and frequency. In practice though, notes are never the same frequency, notes (and indeed all naturally occurring sounds) are combinations of different frequencies, a fundamental frequency plus a number of harmonics (mathematically related frequencies) and it's the number and balance (relative amplitude) of all those frequencies that account for differences in timbre. The exception is if we generate an (artificial) note which is just a single sine wave, in which case there are no other frequencies (harmonics) but then of course there would be no differences in timbre either.

G

PhonoPhi, pstickne, buonassi and 3 others like this.
3. Thanks Greg

Makes sense. What confused me is the 3rd para of the wikipedia article which talks about timbre being a different sound even when the pitch and loudness is the same - assuming pitch has the same meaning as frequency and loudness the same as amplitude. So I take it then that second and higher order harmonics are what is important (but even so, wouldn't a note with the same pitch and loudness have the same harmonics)?

https://en.wikipedia.org/wiki/Timbre

Last edited: Apr 23, 2019
4. Yes, it's that assumption that's causing the problem. "Pitch" doesn't have the same meaning as frequency, the "pitch" of a note usually refers to the frequency of the fundamental and effectively ignores all the harmonic frequencies. And, loudness is a perception that's not directly related to amplitude or rather, amplitude is only one of several factors which contribute to our perception of loudness. In other words, we can have two very different amplitudes which can be perceived as the same loudness and we can have two very different combinations of frequencies which can be perceived as notes of the same pitch (but different timbre).

G

Last edited: Apr 23, 2019
pstickne, buonassi and old tech like this.
5. I've heard audiophiles talk about "timbre matched" speakers for multichannel, and it never made any sense to me. A speaker should produce sound with a degree of fidelity. Speakers shouldn't have harmonics of their own. If one speaker can in theory produce a particular response curve, another one that can do the same should sound the same. I think in this case people are mixing up the term timbre with frequency response.

TronII likes this.

6. I've used timbre matched before when discussing multichannel, but agree that the more accurate term is frequency response.

7. Strictly speaking, it could make some sense. A change in timbre (both actual and perceived) can occur by simply changing the relative balance between the various harmonics which comprise a note/sound, you don't necessarily have to add or remove harmonics to change the timbre (although in practice this is usually the case). So a speaker does not have to add harmonics of it's own to change the timbre, it just needs to have a response which emphasises (or de-emphasises) frequencies which correspond to harmonics in the recording. This is always going to happen to some degree because no speaker has a perfectly flat response and the acoustics of the room the speakers are placed in will usually cause far larger peaks and troughs than the speakers themselves. Therefore, in certain circumstances, using the term "timbre" with regard to the perceived output of a speaker might not be incorrect, although I entirely agree with you; we really should be using the term "frequency response" to more precisely describe what is occurring and avoid confusion.

G

8. My practical ordinary world experience is that as long as a speaker can produce a full range of frequencies clean and at a decent volume level, you can just EQ them to match each other. If a speaker has an uncorrectable "timbre", there's probably something wrong with it.

9. Wiki has been corrected.

10. Agreed and if speakers are lending their own "timbre" to the sound,wouldn't the addition of more "timbre matched" speakers just be the addition of more of the same distortion?

Last edited: May 2, 2019
11. Thing is, they have so much variance in the production of their drivers, the have a hard time selecting even two that sound "close enough".

12. I've done a brief search on the reproduction of timbre. Am I right to say that accurate reproduction timbre (or failure to) is 100% the result of its frequency response.

Does that mean that a bass boost (or any adjustment from flat) will inevitably result in shifted timbre of an instrument such as a foot drum?

AlwaysForward likes this.
13. That's why God created equalizers!

Yes, timbre is affected by frequency response. All sound is affected by frequency response. Sound is basically frequency and amplitude in time. It would take a very large timing error to start affecting the timbre of musical instruments.

Last edited: Jul 22, 2019
SilentNote likes this.
14. “Large timing error” so that means when certain instruments or vocal sounds so off that I notice it, that means they’re completely off the mark...

15. Large timing error could cause high frequency cancellation making the response sound wonky. You could try to fix it with EQ but it wouldn’t fix the problem. Not likely anything like that would happen though.