not easy... but it is possible to identify some spectrographic
features to be linked to spoken letters or couples of them. In some
cases it is easy... in other cases you don't see a specific section
to be linked to a letter, but you see just a transition...
you can try to label different sections... and then you can try to
"fill the holes" to make a meaningful concatenation of letters...
then you need to separate meaningful words... and concatenate them
into a meaningful phrase...
vowels can be easily identified because of their clean harmonic structure
u is the vowel with energy mostly concentrated on low frequency
i is the one with energy mostly at high frequency
sss can be easily identified because of the high frequency noise you
see on the spectrogram...
of course, an expert phonetician could help more,
Gianni
At 14.43 20/06/2006, you wrote:
>To all recordists,
>
>Recent discussion concerning spectrograms prompts the following: I've
>been making spectrograms of speech, does anyone in the group have the
>ability to transcribe into English what was said directly from a
>spectrogram of the phrase or sentence?
>
>Mike
>
>
>
>No virus found in this incoming message.
>Checked by AVG Anti-Virus.
>Version: 7.1.394 / Virus Database: 268.9.1/369 - Release Date: 19/06/2006
--------------------------------------------------------------
Gianni Pavan
Email
Centro Interdisciplinare di Bioacustica e Ricerche Ambientali
Universita' degli Studi di Pavia
Via Taramelli 24, 27100 PAVIA, ITALIA
Tel +39-0382-987874
Fax +39-02-700-32921
Web http://www.unipv.it/cibra
----------
No virus found in this outgoing message.
Checked by AVG Anti-Virus.
Version: 7.1.394 / Virus Database: 268.9.1/369 - Release Date: 19/06/2006
|