EMRFD Message Archive 260
Message Date From Subject 260 2006-12-24 12:20:57 ka3j12 Voice Power Spectrum Density Hi - I've been searching for a graph showing power versus frequency for
a typical male and female voice. No luck. I recall hearing somewhere
that for the male voice, although the fundamental frequencies are
pretty low, most of the power is in the harmonics and/or in the
nonvoiced higher frequency (sibilant) consonants. Is that correct?
I'm interested in this because simple crystal filters don't offer as
much suppression at lower frequencies and I was wondering or hoping
that it didn't make quite as much difference due to the power spectrum
of the voice. Thanks for your help.
Happy holidays everyone!
269 2006-12-24 21:36:07 Rick Re: Voice Power Spectrum Density Hi Ron,
Interesting question, and no simple answer. The power spectral density of a male radio
announcer is different than a man in normal conversation. The power spectrum desired for
maximum intelligibility is not the same as for natural sounding speech. There are many
resources on the web. Here is an instructive one:
Also, the long-term average spectrum isn't the whole story. Transmitter power, efficiency,
and spectral purity off-channel factor into the discussion along with intelligibility and natural
sound. Experiments are in order--but as a benchmark, the late 1950's Collins S-line SSB
transmitters sound good on the air with male voices, and that system has been described in
detail in many publications and handbooks. I believe it's just a modestly flat, 6th order filter
with a relatively narrow bandpass (nominal 2.1 kHz) and placement of the carrier 20 dB down
the filter skirt.
271 2006-12-25 05:32:37 w4zcb77 Re: Voice Power Spectrum Density 272 2006-12-25 15:02:00 Lasse Re: Voice Power Spectrum Density Ron,
I might have one graph buried in a 25 year old stack of paper... did some work way back at the salt mine. You have to QRX a few days until it opens again :) As for intelligibility vs. power most power is in the low end but the high part is absolutely vital for understanding. That is why you often find that speech processors do amplify high frequency, and even try to reduce low end of the spectra. Listening on a pile up, those who cut through is often very high pitched and sounds awful, but, you are able to pick up the call sign. But it is not something to be used for round table chats on 80 m !
On Sun, 24 Dec 2006 20:20:31 -0000
"ka3j12" wrote:> Hi - I've been searching for a graph showing power
>versus frequency for
> a typical male and female voice. No luck. I recall
> that for the male voice, although the fundamental
> pretty low, most of the power is in the harmonics and/or
> nonvoiced higher frequency (sibilant) consonants. Is
> I'm interested in this because simple crystal filters
>don't offer as
> much suppression at lower frequencies and I was
>wondering or hoping
> that it didn't make quite as much difference due to the
> of the voice. Thanks for your help.
> Happy holidays everyone!
> Ron (KA3J)
> Yahoo! Groups Links
> <*> To visit your group on the web, go to:
> <*> Your email settings:
> Individual Email
273 2006-12-26 16:37:55 ka3j12 Re: Voice Power Spectrum Density Thanks for the help folks! The information confirms that one can
sacrifice the low end of the spectrum and still maintain decent
intelligibility at the cost of naturalness.