Human hearing range is commonly known to be 20 Hz to 20 KHz
But what about speaking range? We know that speech Sounds are generated through our vocal cords. So What is the maximum value of frequency that human speech can have?
$\begingroup$
$\endgroup$
5
-
8$\begingroup$ "Wang Xiaolong holds the Guinness Book of Records title for the highest vocal note produced by a man (E8, 5243 Hz)." en.wikipedia.org/wiki/Whistle_register (But that doesn't include harmonics.) $\endgroup$– endolithCommented Jul 7, 2022 at 15:17
-
3$\begingroup$ Human voice/speech includes other mechanisms to produce sound than just vocal cords. You should define the question better if you want the whole range of human speech or only vocal cords. $\endgroup$– JustmeCommented Jul 9, 2022 at 14:29
-
1$\begingroup$ Has anyone measured a wailing baby? $\endgroup$– End Antisemitic HateCommented Jul 9, 2022 at 15:47
-
3$\begingroup$ @rock: Wailing babies peak at arounf 2khz, where human hearing is most sensitive. Evolution at it's best. $\endgroup$– MaxCommented Jul 9, 2022 at 20:20
-
$\begingroup$ @Max Ah thanks, is there any indication of what came first, babies wailing at 2kHz or human hearing being most sensitive at that frequency? $\endgroup$– End Antisemitic HateCommented Jul 9, 2022 at 21:07
Add a comment
|
2 Answers
$\begingroup$
$\endgroup$
3
Especially What is the maximum value of frequency that human speech can have?
This depends on how exactly you define it. Fricatives ("s","f","sh" ...) and plosives ("p","k","t", ...) are fairly broadband noises and have energy up to 20 KHz and beyond. However, the energy above 10 kHz is generally very small and cutting it of makes little practical difference. Bandwidth that is typically used for speech is
- 3.5kHz Telephone quality. Good enough for intelligibility but doesn't sound very good and you can't tell the difference between, for example "f" and "s" sounds.
- 8kHz High Quality speech. Sounds natural and easy to understand but there are still audible differences to the original. This is commonly used in Voice Assistants (Amazon Alexa, Apple Siri, etc.) and similar devices
- 20kHz : HIFI. Indistinguishable from the original. Used for Music and Movies/Videos.
-
3$\begingroup$ Fricatives can be broadband noise generated by turbulent air flow. An ultrasonic microphone can pick up some (small amount of) energy from these fricatives above 24 kHz. Depends on what noise floor you want to use to limit your upper frequency bandwidth. $\endgroup$– hotpaw2Commented Jul 7, 2022 at 23:12
-
$\begingroup$ @Hilmar, Sorry I mistakenly took the question in the context of speech communication and have deleted my answer $\endgroup$– malik12Commented Jul 13, 2022 at 6:00
-
$\begingroup$ "The /ʃ/ has most of its acoustic energy from around 2,000 Hz up to 10,000 Hz." -- Beyond 10kHz, I vaguely remember that we're probably around -60dB, compared to the normal speech volume. Thanks to auditory masking, the content above 10kHz is irrelevant to human ears in most scenarioes. $\endgroup$– KlawsCommented Feb 7 at 14:12
$\begingroup$
$\endgroup$
3
The fundamental speaking frequency of humans can reach up to around 1kHz, although higher values than, say, 500Hz usually appear only while singing. The harmonics and non-tonal parts of speech can have relevant energy beyond 10kHz. See the screenshot below for an example. This is a spectrum of ITU-T P.501 speech signals, male and female speakers.
Limiting the bandwidth to 4kHz was a trade-off between intellegibility and cost. It produces problems especially when trying to tell fricative sounds apart. (Like "f" and "s").
Edit: Limiting bandwidth always means less cost. Bigger bandwidth means more information to transmit. Plus: the old speakers and carbon microphones in the olden days where quite noisy in the upper frequency range.
-
$\begingroup$ Would you mind adding to your answer what you mean by "limiting the bandwidth to 4kHz was a trade-off between intelligibility and cost"? $\endgroup$ Commented Jul 9, 2022 at 15:45
-
1$\begingroup$ @RockPaperLz-MaskitorCasket If you bandwidth limit your signals, you can frequency shift different signals and merge them together, and then send that across a single copper wire. The limit here is the bandwidth of the copper. The more bandwidth you give to each phone conversation, the fewer conversations a piece of copper can support, and the more pieces of copper you have to run to support a city. $\endgroup$ Commented Jul 9, 2022 at 18:53
-
$\begingroup$ @CortAmmon Thanks Cort. That's what I thought the quote was likely about, but I was confused because I didn't see anything about telephones in the question or elsewhere in this answer! $\endgroup$ Commented Jul 9, 2022 at 21:02