Hello Forum. I am interested in finding some vocals of female whispering. If anyone happens to come across any in this database, if you would be so kind as to post them, i would be extremely grateful.

--- (Edited on 12/5/2008 6:08 pm [GMT-0600] by ) ---

>interested in finding some vocals of female whispering

I'd really like to know what kind of speech recognition application this might be for...???

Ken

--- (Edited on 12/8/2008 11:13 am [GMT-0500] by kmaclean) ---

Sphinx

--- (Edited on 12/8/2008 9:09 pm [GMT-0600] by akaletto) ---

>Sphinx

That's the speech recognition engine... what inquiring minds would like to know is why do you need this type of speech? 

Speech Recognition Engines trained with a certain type of speech (e.g. female whispering speech) work best at recognizing that same type of speech.  Why do you need this type of speech? What is your problem domain?

--- (Edited on 12/10/2008 11:56 am [GMT-0500] by kmaclean) ---

I understand your curiosity, however, this is for a voice recognition project I am working on in which I do not wish to speak publicly about, nor do I feel the need to conditionally explain myself to your satifaction. If you don't want to help me then don't.

--- (Edited on 12/10/2008 3:58 pm [GMT-0600] by akaletto) ---

Whispering is frequently used in an interpretation / translation context. Say you have a trial where the accused does not speak the official language. In the interests of smooth process you can't stop after every sentence to allow the court interpreter to translate. You give the interpreter a mike which transmits to an earpiece receiver for the accused. Then as the trial goes forward the interpreter sitting in the courtroom whispers (in this way the normal vocal processes are not interrupted) the translated version into the mike and the accused can follow the process in real time.

Generalizing, you can use whispering whenever there are two speech processes running concurrently and you need to benefit from the difference in volume levels. It is amazing how speech is preserved at very low volume which is later amplified. You might argue that there is a benefit to not blaring at high volume, you have more control over vocalization, resulting in better and clearer speech.

--- (Edited on 1/25/2009 12:59 pm [GMT-0600] by colbec) ---