[ Close ]
Username
Password


Forgotten your password?
 
Computer Projects
 Projects.spogel.com Students, Freelancing projects Acadamic and Freelancing Projects Computer Projects
Message Icon Topic: Converting voice message to a text signal Post Reply Post New Topic
Found this useful? Give a +1   Add to Facebook: Add to Facebook
Check out   Final Year Projects   &   Mini Projects

Free Project Downloads
CS / IT Project Downloads
.Net Project Downloads
JAVA Project Downloads
PHP Project Downloads
Networking Proj Downloads
Project Topics
Computer(CS/IT) Projects
Software Mini Project Topics
Android Project Topics
Networking Project Ideas
Dot Net Project Topics
Latest IEEE Projects List
IEEE JAVA Project Topics
Project Presentation Tips
How to choose Project Topic
Tips for Final Year Project
Steps for FYP, Mini Project
Preparing a Project Report
Problems faced in Projects
 
Get Free Software CS / IT Project Downloads:

Enter your email address:  

Author Message
jtcss
Newbie
Newbie


Joined: 16-Sep-2009
Online Status: Offline
Posts: 0
Quote jtcss Replybullet Topic: Converting voice message to a text signal
    Posted: 16-Sep-2009 at 10:39am
I am interested in making an application for the deaf wherein the voice message at the caller end will be transmitted in the form of data packets such that they appear in the form of text messages at the reciever's end.

As the reciever is a deaf person, he would use an instrument which has a LED that glows when he recieves a voice message.(instead of a ringtone). His reply to the same would be in the form of a text message which would be converted to voice at the other end. Please suggest a practical way to implement this (or something similar)as a final year project(BE)
IP IP Logged
SArioDabidG.
Newbie
Newbie


Joined: 18-Jan-2010
Online Status: Offline
Posts: 0
Quote SArioDabidG. Replybullet Posted: 18-Jan-2010 at 4:04am
Hi jtcss, I am a newbie also. I read a similar thread like this before. They said for speech recognition you must us dsp.   
IP IP Logged
kranthikumar
Senior Member
Senior Member


Joined: 21-Sep-2008
Online Status: Offline
Posts: 0
Quote kranthikumar Replybullet Posted: 23-Feb-2010 at 2:58am
A mobile communication terminal for performing a voice-text message conversion process comprising: a memory for storing condition information for determining when the voice-text message conversion process occurs and selecting a specific voice message among received and stored voice messages which is a target of the voice-text message conversion, said memory including a voice message area for storing voice messages and a text message area for storing text messages; a voice data detecting unit for extracting a voice signal from the specific voice message; a voice-text converting unit for converting the voice signal to a text message; a controlling unit for determining whether voice-text message conversion is required for the received and stored voice messages, and selecting the specific voice message, based on the condition information, and storing the text message converted by the voice-text converting unit in the text message area, wherein whether the voice-text message conversion is required is determined by comparing a total size of the voice messages with a threshold set in the condition information; selecting a voice message having a highest priority for voice-text message conversion from among the voice messages, based on selection reference information set in the condition information, if the total size of the voice messages is larger than the threshold; checking whether the voice message having the highest priority is locked for protection; and selecting the voice message having the highest priority for voice-text message conversion if the voice message is not locked, and selecting an other voice message having a second highest priority as a new highest priority if the voice message is locked.
IP IP Logged
SunishTheGreat
Groupie
Groupie


Joined: 21-Sep-2008
Online Status: Offline
Posts: 0
Quote SunishTheGreat Replybullet Posted: 23-Feb-2010 at 3:02am
Came across this related patent when googling...

The present invention relates to a method enabling the transmission of a message from a portable communication device, such as a mobile phone or a personal digital assistant, to a separate device, in order to generate a text message to be displayed either locally at said separate device, or at a distant end user terminal.

Today, mobile applications aiming at sending text messages (for instance via Short Message Services, Multimedia Message services or E-mails) are limited due to the fact that text input is difficult to perform at the portable device. This limitation could dramatically restrict the use of some new applications such as mobile weblog or blog, in which a great quantity of text may have to be input by the user of the mobile phone.

A known solution to counter message input issue is to use voice messages instead of text messages. Some mobile phones already implement this solution consisting in recording at the mobile phone a text message which is sent to an end user terminal via MMS (Multimedia Message Service). However, this solution requires that end user terminal (for instance another mobile phone) also supports this service. In addition, the end user might prefer to receive a text message to be displayed instead of listening to an audio message.

For blog applications, it is also already known to call a defined number corresponding to a server, and to leave a voice message which can be listened by other users through this server. Here again, there is no possibility to display a text message.

The aim of the invention is to remedy the above drawbacks by proposing a solution enabling an easy input of message at the portable communication device's side, while offering the possibility to display a text message at the destination terminal's side.

To this aim, an object of the present invention is to provide a met hod for transmitting a message from a portable communication device to a separate terminal, comprising a step of recording a voice message at the portable communication device, and a step of generating, at the portable communication device, a multimedia file comprising at least said voice message, characterized in that it further comprises the following steps:

transmission from said portable communication device to said separate terminal of a signal including said multimedia file and a specific flag for indication that voice message is to be displayed as a text message;
reception at said separate terminal of said signal, extraction of said voice message and detection of said specific flag;
upon detection of said specific flag, speech-to-text conversion at said separate terminal in order to transform said voice message into a text message to be displayed.
In one possible embodiment, the text message is for local use, i.e. to be displayed at separate terminal.

Alternatively, the method according to the invention may comprise a further step for transmitting said text message from said separate terminal to an end user terminal such as a mobile phone or a personal computer.

Transmission of signal from portable communication device to separate device can be made by any kind of transmission links, such as wired or wireless links.

Similarly, transmission of text message from separate device to end user terminal can be made through wired and/or wireless communication links.

In one possible embodiment, specific flag indicating that voice message is to be displayed as text message is included within multimedia file comprising the recorded voice message.

Alternatively, said specific flag is sent in the signal as an attachment to said multimedia file.

A second object of the invention is a portable communication device for implementing the method, characterized in that it comprises:

means for recording a voice message, and for generating a multimedia file comprising at least said voice message;
software means for generating a signal including said multimedia file and a specific flag for indication that voice message is to be displayed as a text message;
transmission means for transmitting said signal to a separate terminal.
A third object of the invention is a terminal for use as separate terminal in the method, characterized in that it comprises:

means for receiving signal transmitted by said portable communication device, for extracting said voice message from the received signal and detecting said specific flag;
upon detection of said specific flag, software means for converting said voice message into a text message to be displayed.
IP IP Logged
SunishTheGreat
Groupie
Groupie


Joined: 21-Sep-2008
Online Status: Offline
Posts: 0
Quote SunishTheGreat Replybullet Posted: 23-Feb-2010 at 3:08am
One more patent...may be we can try completing before ....

US Patent 6366651 - Communication device having capability to convert between voice and text message

Abstract:

The communication device provides the capability to automatically convert between voice and text messages. This communication device enables the calling party to input a message in voice mode, then activate the communication device to automatically convert the message into a text message format for transmission to the called party. The communication device can display the content of the text message prior to transmission or can retrieve the message from its memory and play the message back to the calling party in an audio mode. The communication device automatically initiates the outgoing call and, since the message is transmitted in a text mode to the called party, the destination can be a computer system or output to a printer. This feature therefore enables hands free operation for E-Mail. The communication device also provides a data retrieval feature where the user can retrieve E-Mail or voice mail messages using the communication device and have the retrieved messages output in either text or audio format. The retrieved messages, since they are stored in memory in the communication device, can be forwarded to another destination once the user has reviewed their content.
IP IP Logged
kranthikumar
Senior Member
Senior Member


Joined: 21-Sep-2008
Online Status: Offline
Posts: 0
Quote kranthikumar Replybullet Posted: 23-Feb-2010 at 3:18am
Came across this info from a technical person.....with a piece of code logic....

"Speech to text conversion is a big animal. This needs a large HMM dictionary for "all" words to match.

And this still is on top of the speech recognition process extracting by ceptra of the moving speech, VAD and the silence detection.

BTW, my last project study also was on DSP. The first plan was to do a speaker-independent speech recognition in a very noisy environment. I shelved it for I don't have enough time to do the beast.

I just proposed an implementation of a adaptive noise cancellation algorithm and was accepted by my adviser. BTW, I used python (with pylab extension) as the PC simulation program and not in MATLAB. For reference for those interested and may improve on it, here's the source code of the algorithm"

# Noise Removal Spectral Subtraction Algorithm Simulation Software
# Copyright (c) 2009, Regulus Berdin
# All rights reserved.
#
# Permission is hereby granted, free of charge, to any person
# obtaining a copy of this software and associated documentation
# files (the "Software"), to deal in the Software without
# restriction, including without limitation the rights to use,
# copy, modify, merge, publish, distribute, sublicense, and/or
# sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following
# conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
# EXPRESSED OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
# OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
# NONINFRINGEMENT. IN NO EVENT SHALL REGULUS BERDIN BE LIABLE FOR
# ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF
# CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
# CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
# SOFTWARE.

from pylab import *
import wavefile

def only_positive(v):
    return (v - abs(v))/2

class vad_process:
    def __init__(self, limit = 30, overhang_size=5):
        self.overhang = 0
        self.speech_flag = 0
        self.sd = 0
        self.dist = 0
       
        self.limit = limit
        self.overhang_size = overhang_size
       
    def update(self, signal, noise):
        self.sd = log(signal) - log(noise)
        self.sd = greater(self.sd,0)*self.sd/log(2)
        self.dist = sum(self.sd)
    
        if self.dist > self.limit:
            self.overhang = self.overhang_size
            self.speech_flag = 1
        else:
            if self.overhang > 0:
               self.overhang -= 1
              
        if self.overhang == 0:
            self.speech_flag = 0
       
srate = 8000
GainOut = 3.0

W = 256
STEP = 102
InitNoise = 0.2 #ms
NoiseAvgFactor = 0.01
alpha = 8
beta = 0.003

InitNoiseAvgFactor = fix((InitNoise * srate) / STEP)
SilentFrames = copy(InitNoiseAvgFactor)
       
fwi = wavefile.wavefile("noisy3.wav")
x0 = fwi.read();
x = x0.reshape(x0.size)/32768.0

noise = zeros([1,W/2])[0]
cx = zeros([1,x.size])[0]
xvad = zeros([1,x.size])[0]

L = x.size - (x.size % W) - W - 1

vad = vad_process(40, 2)

for i in range(0, L, STEP):

    frame = x[i:i+W] * hanning(W)
    y = fft(frame)/W
    
    signal = (y.real[:W/2] ** 2) + (y.imag[:W/2] ** 2)

    if SilentFrames>0:
        SilentFrames -= 1
        noise = noise + (signal / InitNoiseAvgFactor)
    else:
        vad.update(signal, noise)

    if vad.speech_flag == 0:
        noise = noise + (signal - noise) * NoiseAvgFactor
        clean_signal = beta * noise
    else:
        xvad[i:i+W] = ones([1,W])[0]
        clean_signal = signal - alpha * noise
        clean_signal = greater(clean_signal, 0) * clean_signal
        clean_signal = maximum(clean_signal, beta * noise)
    
    ratio = sqrt(clean_signal/signal) * W

    sp = y.real[:W/2] * ratio + y.imag[:W/2] * ratio * 1.j
    sp_rev = conj(sp)[:0:-1]
    spectrum = hstack([sp, sp_rev, 0])
    
    cx[i:i+W] += ifft(spectrum).real * GainOut
    xvad[i:i+W] = ones([1,W])[0] * vad.speech_flag

for i in range(0,cx.size):
    if cx>1: cx=1
    if cx<-1: cx=-1

py = array(range(0,cx.size)) / (srate * 1.0)

if 1:
    figure(2, figsize=(10,4.5))
    title('Noisy Speech')
    xlabel('time (s)')
    ylabel('level')
    plot( py, x, "k-" )
    axis((0.0, cx.size / srate,-1.0,1.0))
    savefig('fig1.png')

    figure(3, figsize=(10,4.5))
    title('Enhanced/Clean Speech')
    xlabel('time (s)')
    ylabel('level')
    plot( py, cx, "k-" )
    axis((0.0, cx.size / srate,-1.0,1.0))
    savefig('fig2.png')

if 1:
    figure(4)
    title('Noisy Speech Spectogram')
    ylabel('frequency (Hz)')
    xlabel('time (s)')
    specgram(x, Fs=8000)
    savefig('fig3.png')
    
    figure(5)
    title('Enhanced/Clean Speech Spectogram')
    ylabel('frequency (Hz)')
    xlabel('time (s)')
    specgram(cx, Fs=8000)
    savefig('fig4.png')
    
    show()
    
print "Finished."


IP IP Logged
Vishal
Newbie
Newbie
Avatar

Joined: 10-Jul-2010
Location: India
Online Status: Offline
Posts: 0
Quote Vishal Replybullet Posted: 10-Jul-2010 at 10:07pm
I am a final year computer engineering student.

I am interested in this project.
Could you send me the details of the project?

I Think that i may modify your algorithm.
IP IP Logged
Post Reply Post New Topic

Free Programming Courses in JAVA, .Net and PHP (Register Now!)

Java Free Online Course

.Net Free Online Course

PHP Free Online Course

Printable version Printable version

Forum Jump
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot delete your posts in this forum
You cannot edit your posts in this forum
You cannot create polls in this forum
You cannot vote in polls in this forum



This page was generated in 0.016 seconds.
© Spogel.com   |   Privacy policy