🚀 go-pugleaf

Hi,there,

I am a Doctoral student, doing research on revealing live mouth images 
through speech signal analysis(also called automated lip-sync).
Currently, I need some sample data, of synchronised image and speech. 
This is a broadcast request for data of that kind.

Description:
The data should be 
a. continuous head and shoulder images, synchronised visual and audio; 
b. of read speech,  several minutes long; 
c. both the image and the speech signal should be accessible by computer,
frame by frame( e.g., 20ms/s). The data format is not important.

So, my first question is: can anyone tell me if such a database exists now 
and how can I get access to it?

My second question is about how to catch the data properly.

I've got a full motion digital video adapter DVA-4000/ISA produced by 
Videologic Limited. According to the producer's information, 
DVA-4000/ISA is a multimedia system adapter which can transfer both 
audio and video signals between IBM PC/XT/AT and audio/video sources 
such as video cassette recorder(VCR),and only Sony 9000 series(Umatic) 
VCRs are DVA-4000/ISA supported.
We are considering purchasing a Sony 9000 series VCR called VO9600P
(about A$5534 including a RS232 port) to construct the video and audio 
data acquiring system.
I would be grateful for any ideas and suggestions about the data 
collecting system. 
In particular, I would value information on some questions like: 
Can DVA-4000/ISA really transfer a speech signal from tape to computer
(with what command)?; 
Will the VO9600P do(or is there a better type)?

Thanks a lot.


Suhuai Luo
Speech Technology Research Group
Department of Electrical Engi.
The University of Sydney
NSW 2006
Australia

luosu@brutus.ee.su.oz.au
tel: +61 2 692 4509
fax: +61 2 692 3847