How to convert Speech to Text

I am doing project on speech signal analysis.how to convert speech to text any one help me

Answer 1

[Gabriele Bunkheila]

Hi Nemala,

I work at MathWorks and I noticed this old question of yours just now. In case it was useful to others, I thought I'd leave a link to our recent MATLAB Central sumbission speech2text:

https://www.mathworks.com/matlabcentral/fileexchange/65266-speech2text

That enables to leverage 3rd party speech-to-text web services (so far including from Google, Microsoft and IBM) directly from MATLAB using a simple MATLAB Function. The syntax looks like the following:

>> [samples, fs] = audioread('helloaudioPD.wav');
>> soundsc(samples, fs)
>> speechObject = speechClient('Google','languageCode','en-US'); 
>> outInfo = speech2text(speechObject, samples, fs);

>> outInfo.TRANSCRIPT =

ans = 
    'hello audio product Developers'

>> outInfo.CONFIDENCE =

ans = 
    0.9385

Please refer to the html folder in the submission for a step-by-step getting started guide.

Thanks,

Gabriele.

[gozde nur]

How can I create speechClient object ? I got stuck in this point. Thank you

[Gabriele Bunkheila]

Hello,

The property-value pairs for speechClient vary depending on the specific service used (i.e. Google, IBM, Microsoft). You should find it useful to take a look at the "Create speechClient" section of the "Examples" tab of the submission page for speech2text - that also include links to the documentation of the respective services.

I hope this helps.

Gabriele.

[gozde nur]

Ichecked thta section and run this line to the command window;

speechObject = speechClient('Google','languageCode','en-US')

but it gave me an error 'Size and validator functions not supported for abstract properties.

Property 'Name' is abstract in class 'BaseSpeechClient'.

'

but I cannot open BaseSpeechClient or any other .p files because of the protection. Do I use wrong speechClient line?

[Raja Palanimurugan]

Hello,

Your error stack looks like you hit a bug in speech2text code seen in MATLAB releases older than R2018b.

I have fixed the bug in the speech2text submission and verified that it works in MATLAB releases R2017a through R2018b. Please download the latest version of speech2text and try again.

Thanks,

Raja

[gozde nur]

Thank you so much it works so well

[Akshatha Prasad]

Hello

I have downloaded the latest version of speech2text in Matlab2018a.

When I am trying to run I am getting the following error:

Output argument "tableOut" (and maybe others) not assigned during call to

"BaseSpeechClient/speechToText".

Error in speech2text (line 45)

tableOut = connection.speechToText(y,fs,timeOut)

Error in stt (line 7)

outInfo = speech2text(speechObject,samples,fs,'HTTPTimeOut',25)

Immediate help is appreciated.

[Talal Salim]

Is there a way to use this speech2text program offline without connecting to the Internet?

Also, how can we add a new language package to this program?

[Nikos Korobos]

https://uk.mathworks.com/matlabcentral/answers/1678534-speech-to-text-sapi-offine-mode

Want to know too if you went through with it

[Gabriele Bunkheila]

Indeed, as of today speech2text uses a web API interface to access three different speech-to-text web services. You also need to create an account with either of the supported providers to use that feature.

If you need a local solution, you may want to try this recent submission based on the deep network wav2vec -- this will downloaded and run the actual network on your local machine as part of the setup process:

https://www.mathworks.com/matlabcentral/fileexchange/103525-wav2vec-2-0

Note that this is a simple end-to-end deep learning model and not a a fully-featured service. One of the limitations is that the model version available only works for English and it doesn't have any dialect-specific variations. That said, its is a state-of-the-art model as far as single networks go and its performance is remarkable for many ordinary test cases. Let us know how that works for you!

Answer 2

[Image Analyst]

Not an answer for you, but in case anyone is interested in the reverse direction (text to speech).....If you have Windows, try .Net. Try this MATLAB code snippet:

% Program to do text to speech.
% Get user's sentence
userPrompt = 'What do you want the computer to say?';
titleBar = 'Text to Speech';
defaultString = 'Hello World!  MATLAB is an awesome program!';
caUserInput = inputdlg(userPrompt, titleBar, 1, {defaultString});
if isempty(caUserInput)
  return;
end; % Bail out if they clicked Cancel.
caUserInput = char(caUserInput); % Convert from cell to string.

NET.addAssembly('System.Speech');
obj = System.Speech.Synthesis.SpeechSynthesizer;
obj.Volume = 100;
Speak(obj, caUserInput);

Speech to text is a lot more complicated of course.

[Image Analyst]

Did you overlook Gabriele's answer above? Does that not help?

[Image Analyst]

We can't do your project for you. You have to be more specific in what you need help with. At this point, it's just too vague for any of us to help. At this point in your project all we can do is to point you to software that can help you. If you have specific questions about small snippets of your code, we can help you, but please read this link first before you post your code, data, and specific question(s):

TUTORIAL: How to ask a question (on Answers) and get a fast answer

[Nikos Korobos]

Hi, is there any possibility to function STT without an internet connection?

Asking because it makes use of various SAPIs

[Gabriele Bunkheila]

Please see the response provided to this same question in the main thread

Answer 3

[zahoor mosad]

please send me the code to my gmail [e-mail]. It will be great help for my thesis. I'm never forget your kindness.

[Image Analyst]

Did you overlook Gabriele's answer above? Does that not help? Exactly what kind of code were you looking for? What is the topic of your thesis?