Detective Time (itch) (Mattie) Mac OS

Bartender

I've been playing with text to speech recently. After giving up on FreeTTS as more trouble than alternative approaches, I've been able to develop classes for the purpose for Windows (using a VBScript) and Linux, running in Sun VirtualBox (using a shell script and eSpeak) but I don't have access to, nor have ever so much as touched, a Mac. The Mac-specific code that follows is a cut-paste-edit of a full night's worth of Google searching. I took a look around, but I don't see any topics of this nature here.
This is not exactly a SSCCE, but it is as short as I could make it. If you have a Mac and a little time on your hands, feedback and bug fixes would be most welcome.import java.io.*; import java.util.*; public class MacSpeaker { protected File scriptFile; protected Process process; protected final String[] voices; protected int voiceIndex = 0; //protected int pitch = 0; // -10 to +10 protected int volume = 50; // 0 to 100 protected int rate = 100; // 80 to 400 public static void main(String[] args) { new MacSpeaker().test(); } private void test() { int numVoices = voices.length; if (numVoices 0) { return; } Random random = new Random(); for (int i = 0; i < 10 && i < numVoices; i++) { voiceIndex = random.nextInt(numVoices); volume = random.nextInt(100); System.out.println('Voice: ' + voices[voiceIndex] + ' Volume: ' + volume); speakAndWait('Hello World'); try { Thread.sleep(500); } catch (InterruptedException ex) { ex.printStackTrace(); } } } public MacSpeaker() { voices = populateVoiceArray(); } protected String[] populateVoiceArray() { Scanner input = null; try { Process p = Runtime.getRuntime().exec('say -v '?'); new ProcessReader(p, true); // consume the error stream input = new Scanner(p.getInputStream()); List<String> voiceList = new ArrayList<String>(); while (input.hasNextLine()) { voiceList.add(input.nextLine()); } return voiceList.toArray(new String[0]); } catch (IOException ex) { ex.printStackTrace(); } return new String[]{}; } protected void makeScriptFile() { String speakScript = ' + 'on run argvn' + ' set theCall to (item 1 of argv)n' + ' set theVoice to (item 2 of argv)n' + ' set theVolume to (item 3 of argv)n' + ' set theRate to (item 4 of argv)n' + ' set oldVolume to output volume of (get volume settings)n' + ' set volume output volume ((oldVolume * theVolume) / 100)n' + ' say -v theVoice -r theRate theCall with waiting until completionn' + ' set volume output volume (oldVolume)n' + 'end runn'; FileWriter fw = null; try { scriptFile = File.createTempFile('speech', '.scpt'); scriptFile.deleteOnExit(); fw = new java.io.FileWriter(scriptFile); fw.write(speakScript); } catch (IOException ex) { ex.printStackTrace(); } finally { if (fw != null) { try { fw.close(); } catch (IOException ex) { ex.printStackTrace(); } } } } public void speakAndWait(String call) { process = speak(call); if (process != null) { try { process.waitFor(); } catch (InterruptedException ex) { ex.printStackTrace(); } } } protected Process speak(String call) { process = null; try { process = Runtime.getRuntime().exec(new String[]{'osascript', scriptFile.getAbsolutePath(), '' + call + '', '' + voices[voiceIndex] + '', ' + volume, ' + rate }); new ProcessReader(process, false); // consume the output stream new ProcessReader(process, true); // consume the error stream } catch (IOException ex) { ex.printStackTrace(); } return process; } } class ProcessReader { public ProcessReader(Process process, final boolean errorStream) { final InputStream processStream = errorStream ? process.getErrorStream() : process.getInputStream(); new Thread() { @Override public void run() { try { BufferedReader br = new BufferedReader(new InputStreamReader(processStream)); String line; while ((line = br.readLine()) != null) { if (errorStream) { System.err.println(line); } else { System.out.println(line); } } } catch (IOException ex) { ex.printStackTrace(); } } }.start(); } }
Thank you for reading.

Bartender

Example use: (should work out-of-the-box on Windows XP and above, requires eSpeak to be installed for Linux).
http://www.fileden.com/files/2008/12/12/2221280/HousieGame.jar

author & internet detective

Bartender

author & internet detective

Say -v '?' Gave an error message at the command line.

Greenhorn

Darryl,
I don't know if this will help you or not, but in Terminal, if I type:
say -v bruce hello my name is bruce
the computer speaks in the Bruce voice: hello my name is bruce
If I type only 'say -v bruce' at the beginning then hit enter, I get the beach ball spinning for a few seconds, then whatever I type at the prompt is spoken in the bruce voice once I hit the enter key. Each phrase I have it speak is on one line in the Terminal.
Perhaps this is already known to you, but maybe not.
BD

Greenhorn

I ran your first code, and changed the ? to a name like Jeanne did. I stepped it to the while loop that read from the Scanner instance, which ended up being a null reference.
I pasted in your new code, and the terminal output read:
Voice: Ralph Volume: 61
and got a Null pointer in the try block at scripFile.getAbsolutePath() in the speak method.
Hope that helps. :-)
BD

Bartender

Thank you, Howard. Should I conclude that the man page for say is in error? Or am I looking at a page that's for a version later than that you're testing on?
edit Oh, I see that say -v '?' appears for ver. 10.7.4 and isn't there for 10.6.
What about the second option, the one uising NSSpeechSynthesizer? Does that fly?

Greenhorn

Far be it from me to declare a manual in error. I just typed some stuff into Terminal and reported to you what happened. :-)
I am running Lion. I did not consult the manual, just typed in a few commands to test how 'say' worked.
If you want me to test something specific, just ask. :-)
Cheers
BD