Hi,
I undersand that UWB speech is speech signal sampled at 32KHz frequency.
I really could not find enough reasons for why we need the speech at this high sampling rate which needs a very high bandwidth compared to narowband and wideband too.
Human voice range covers 300 Hz to 3500 Hz only and the most energy is contained in 1000Hz. So I am worried why do we need the full 32KHz sampling ??
Can you please tell me or discuss all the benefits of the UWB speech....
I am more interested in UWB speech wrt its use/requirement in real life problems and real life systems or better to say where can it make money the UWB research.
bye
Fricatives like /s/ or /f/ go over 3500Hz, IIRC. I think I remember reading that to some speakers, (I forgot if it was Dutch or German or something else,) compression made /s/ sound weird.
I know nothing about the application you ask about but look at vocals with a spectral display in many music recordings. The fundamentals may not go so high, but there is often considerable harmonic energy even above 15kHz.
Yeah but those harmonics are added in the studio with crystallizers and harmonic exciters to make the voice sound fuller and brighter, they wouldn't be used in any UWB speech situation
Yeah but those harmonics are added in the studio with crystallizers and harmonic exciters to make the voice sound fuller and brighter, they wouldn't be used in any UWB speech situation
[a href="index.php?act=findpost&pid=359183"][{POST_SNAPBACK}][/a]
so if I were to record myself saying "sally slapped sammy silly" I won't get anything near 15khz? (i will have to try this later )
I guess UWB may be useful when signal contains not only speech but also some music or background noise which in some situations may be welcome. Like in these "please wait for the assistant to pick up the phone" [insert some boring music here], or in radio broadcasts where they read stories with special effects in the background to make them more illustrative, etc.