How do I create and save an audio file from the output of text to speech synthesizer

How do I save an audio file generated by the text to speech synthesizer, for instance using the code given below. My aim is to take the synthesizer audio output and continue processing it. % 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. %*https://www.mathworks.com/matlabcentral/answers/159113-text-to-speech-synthesis-matlab-code#answer_155653*% NET.addAssembly('System.Speech'); obj = System.Speech.Synthesis.SpeechSynthesizer; obj.Volume = 100; Speak(obj, caUserInput);

Answer 1

[Jan]

See: https://www.mathworks.com/matlabcentral/fileexchange/18091-text-to-speech

And:

NET.addAssembly('System.Speech');
synth = System.Speech.Synthesis.SpeechSynthesizer;
synth.Volume = 100;
synth.SetOutputToWaveFile(fullfile(tempdir, 'yourFile.wav'));  % [EDITED, typo]
Speak(synth, 'Hello world');
synth.SetOutputToNull;   % [EDITED] Close the file

[Gwood]

Thanks, the code helped. If I run the code a second time, this error message pops up: The process cannot access the file 'myFile.wav' because it's being used by another process. How do I make the code keep updating the file upon re-running the code without having to change the file path

[Jan]

See [EDITED] in my code. You need the SetOutputToNull command.

[Gwood]

Thank you! Worked fine

[Muhammad Zeeshan Ahmed Khan]

How can i just read this wave file in workspace ?? just without using audioread

[Jan]

audioread is the perfect tool for reading wav files. I cannot imagine a reason to import the file without this built-in function.

[Muhammad Zeeshan Ahmed Khan]

but audioread doesn,t work for me it ggives error for this tyoe of wav file

[Muhammad Zeeshan Ahmed Khan]

@Jan thankyiu for that good support I just want to plot a spectrogram of the audio file . can you please help me out.

[Muhammad Zeeshan Ahmed Khan]

Error using audioread>readaudio (line 157)
The file type is not supported

Error in audioread (line 136)
[y, Fs] = readaudio (filename, range, datatype);

Error in synth (line 23)
[y,Fs] = audioread(fullfile(tempdir, 'myfile1.wav'));

[Walter Roberson]

@Muhammad Zeeshan Ahmed Khan

Your system might require that you install a "CODEC" for that particular kind of WAV file.

On Windows, audioread() can read any .wav file that is supported by Windows Media Centre. The operating system is asked to read the file.

The third-party program VLC https://www.videolan.org/ might be able to read the audio file and save it in a different encoding. Certainly it would be able to tell you which codec it uses.

There are a few encodings, "Indeo", that are not available on 64 bit operating systems, as the authors of the encodings want those old encodings to die off.

[Muhammad Zeeshan Ahmed Khan]

@Walter Roberson I got u .. Hope it works .. but I was expecting a spectrogram , I don't need an audio file . Is there any way to do so ....

[Walter Roberson]

@Muhammad Zeeshan Ahmed Khan

NO.

.wav files are not just a header and then a plain list of samples .

.wav files have a header, and then they have chunks of encoded audio data. Each encoded chunk contains a 4-character code that indicates which encoding technique has been used for the audio data. One might have code 'H243' and another might code 'IND5' for example. A few of the codes are standardized by Microsoft, but vendors are permitted to create their own codes. The 4-character codes are references to algorithms for encoding audio data. Some of the algorithms are public; some of them are known only to one vendor. The simplest algorithm supported just lists the samples out in full. Another algorithm uses the samples unchanged but uses ZIP compression. Other algorithms might use "perceptual encoding". Some algorithms might be secret and might involve encrypted audio data (for example to enforce Digital Rights Management.)

So in order to read audio data from a .wav file, you need the file itself, and you need a program that generally knows the structure of wav files --- and you need a "plug-in" that knows how to decode audio data that has been encoded according to the particular 4-character algorithm.

It would certainly be possible to write a MATLAB program that found the chunk of encoded audio data, and made the encoded data available to the MATLAB user -- but unless you know the algorithm used by the company that did the encoding, you cannot take that and go further to access the audio content.

Imagine for example a similar text system (instead of audio) in which one chunk of data had 4-character code 'SIML' and content "The lock combination is 311724". And then imagine another chunk of data had 4-character code 'SHRT' and content 'Lck cmb CAAGBD' -- same information, algorithm not difficult to figure out. And then imagine another chunk of data had 4-character code 'Mxyz' and content '8yJQ*c' : MATLAB might be able to tell you 'Mxyz' and '8yJQ*c' but it would have no idea how to convert that to meaningful text.