Speech Recognition - Spell out words

Hello,
I am writing an application that displays employees (pictures) and allows to mark them as enabled (in the office) or disabled (out of office). This will run on a big screen in our office, so that every employee can easily see if a person he is looking for
is in the office or not.
Now I am adding Kinect as user input, since very large touchscreens aren't that affordable. On top, Kinect is cool ;)
I already easily added the HandCursor to the program with great success. I had it running after 5 minutes and it works very fine, so Kudos for that!
Now I am looking into speech. I want to use speech as another way for users to sign in and out. Since all employees have shorthands (3 characters long) it would be nice to use phrases like that:
"Kinect ABC" (ABC = sample employee shorthand)
Then the tile of ABC gets focus and you can:
a) see if ABC is in the office
b) sign in/out
c) get further detail on ABC
So the question here is, can I use the single letters of the shorthand so that users spell out the shorthand? I think that this is the only consistent way to solve my problem, since first- and lastnames can get difficult to recognize.
Best regards,
zuckerthoben
chn

You don't say what language you are using, but, even although you think names will be difficult to recognize, I would give them I try and compare the results with shorthands.
I think you "could" use single letters, but UX will not be great. Users will propably need to pause too many between letters.

Similar Messages

How can we campare two wave spectrum??Related to speech recognition I have to find out whether a word is matching or not?

I have to match a store data to a given data for speech recognition so plz tell me how I can???

Hi Sattu,
Below is an example program on Voice Recognition. It is an awesome example that uses MSAgent (Or Dragon) to do voice recognition. This example should give you a good starting point.
Voice Recognition in LabVIEW
I hope it helps
Feroz
National Instruments

Has anyone figured out how to get speech recognition working with sticky keys enabled on mountain lion?

I'm trying to use speech recognition to input text on my iMac running the latest mountain lion, 10.8.3.
I have sticky keys enabled.
When I try to start speaking by pressing the function key twice nothing happens. I can only get it to work if I disable sticky keys.
The same problem occurs with all the other modifier keys as shortcut, they do not work with sticky keys.
When I try to select a different shortcut, I am unable to select a two key combination, but am limited to one.
If I select the F6 key, or any other single key, I am able to start speech recognition. However the second time that I press the key, it does not stop recognition and process my words. Instead, it restarts the recognition.
Has anyone figured out how to get speech recognition working with sticky keys enabled?
Or a way to get an individual key shortcut to start on the first press and process it on the second?
Or a way to get key combinations to work, as specified by the help:
Dictation On and Off
To use Dictation, click On.
When you’re ready to dictate text, place the insertion point where you want the dictated text to appear and press the Fn (function) key twice. When you see the lighted microphone icon and hear a beep, the microphone is ready for you to speak your text.
Shortcut
By default, you press the Fn (Function) key twice to start dictation. If you like, you can choose a different shortcut from the menu.
To create a shortcut that’s not in the list, choose Customize, and then press the keys you want to use. You can press two or more keys to create your shortcut.

I noticed with version 10.8.4 of OS X that I am now able to select F6 to activate, and the return key to complete the speech recognition. This is still different than the description of how these should function that's included in the help, but at least it's an improvement.

Required code on to find out mfcc in speech recognition

Hi frenzs
I'm doing my acdemic project in speech recognition. so i have to find out mfcc for a wav file. so please send me the code.
my id [email protected]
Thanks
Anjali

sabre150 wrote:
By publishing your email address ( [email protected] ) you are inviting a load of spam. I never publish an email address such as [email protected] since some search engines harvest them.
whatever he said ;-)

Speech Recognition stops working after a minute of nonuse

I'm a grad student trying to keep from exacerbating my carpal tunnel and just figured out how to use the voice commands on my macbook. So far, they work great, they hear my voice, it does what I tell it to do, and I'm not constantly hitting apple+tab and apple+c and apple+v and so on, which really helps my wrists!
BUT. If I go for more than a minute or two without using a voice command, it stops working. I can't pinpoint anything else that triggers it other than if I stop using it for a little bit while I'm reading a paper or something. The microphone is calibrated and working, the little black arrows and the blue and green bars flash on the widget like it hears me, but then it won't give me the sound that says the command was recognized and it doesn't perform the command any time I try to use it again after a few minutes of not being used. It's like the recognition part goes to sleep and won't wake up.
If I go to the system preferences and toggle Speakable Itemst off and then on again, it starts working again until another minute or two of non-use. If I use spoken commands constantly, it will stay on for a good long while, but I'm not using it constantly that's the end of speech recognition. I really can't stop to just tell it some unneeded command to keep it awake all the time! Eventually, the computer gets confused if I turn the "speakable items" on and off too many times from the System Preferences, and it gets confused about whether or not it's turned on or off, or the widget disappears but the System Preferences panel thinks the Speakable Items is on.
I read somewhere that other people with a newer OS have a similar problem if they don't use the default speaking voice, so I switched to Alex and restarted, and the same thing is still happening.
I'm running 10.5.8, and using safari, word 2011, preview, nothing too crazy. My speech recognition is in "Listen only while key is pressed" mode, and like I said, works great until I stop using it for a minute or so. I'm using the internal microphone on my MacBook (dob circa 2008). Any tips on how I can actually keep my Speech Recognition useful? Thanks in advance!

OK, I have one more question now, seems to be the same as this question which was never answered years ago:
Speakable Commands>Application Switching -- how to delete items?
When I click on the triangle from the speech command widget and open speech commands, there is a TON of stuff there that I'll never ever use, but these things do not appear in the Library/Speech/Speakable Items folder so that I might be able to delete them. For example, one item I'd like to turn off is just the word "Home" and the computer keeps thinking that I said "Home" instead of whatever else I meant to say. Another keeps opening iChat or Mail, when I say "open new window" or something like that. I tried clicking, control clicking, option clicking, nothing will highlight a command in the Speech Commands window so that I might delete it.
The "Home" command only shows up when I'm using Word, in the "front window" list. However, the Application Speakable Items/Microsoft Word folder is empty. Is there some other way to get into the speakable commands list and weed out unwanted stuff?

Ignore voice commands / Speech Recognition Server timeout

I'm fairly new to AppleScript but I've found it really entertaining and gloriously simple to pick up so far. I've been using it mostly to make my own speakable items for the Mac's voice control, and I've managed to write some basic things like voice commands to set/change volume and brightness. My goal at this point is to be able to do the majority of things I want to do on my computer by voice only. Sue me - I just rewatched Iron Man and I really wish I had a JARVIS system right now.
This means that the computer must always be listening for commands (I have set it to 'listen continuously with keyword', and set the keyword as 'optional before commands'). It defeats the purpose if I have to walk over and hit a key to start listening for commands. At the same time, I don't want to have to speak a keyword, because I will invariably forget to do so and it's also an added hassle (especially when I have several commands to say - I don't want to have to say the keyword before each command).
You might have an idea of where this is going. Obviously, if the computer is always listening, it will pick up on normal conversation and try to interpret it.
I could just deactivate voice commands when I'm not using it, but as I said earlier, I don't want to have to either hit a key or say a keyword. In other words, I want to be able to activate and deactivate voice commands by voice. Deactivating is easy. Activating is what presents an interesting problem, because if it's deactivated, how will it hear my reactivation command?
Thus, the first part of my question: is there a way - any way at all - to get voice commands to ignore commands until given some keyword, short of literally changing its settings to listen for a keyword before the command?
I decided to try to use speech recognition. This must, I think, be involved in some way, because as I said earlier, this presents the unique problem of having voice commands deactivated yet still able to recognise the reactivation command. I had noticed that while the computer is listening for a specific command, it ignores anything that isn't what it is listening for. I tried to use this to my advantage and wrote a script to listen for a reactivation command; until the reactivation command was given, the computer would ignore everything else. The problem is that the speech recognition server times out after two minutes. I have tried to increase this, but to no avail (see below for my attempts).
Thus, the second part of my question: assuming there is no other answer to my first question, is there a way to get speech recognition to listen indefinitely?
Here are the two things I tried to make the timeout last longer, specifically for five minutes (300 seconds) in both cases. Neither worked; both timed out after 120 seconds.
tell application "SpeechRecognitionServer"
set userinput to listen for {"resume voice commands"} giving up after 300
end tell
and
with timeout of 300 seconds
set userinput to listen for {"resume voice commands"}
end timeout

UPDATE: I have since come up with an alternate solution. I created two speakable items, one for 'deactivation' and one for 'activation'. The deactivation command uses UI scripting to set voice commands to 'keyword required' mode (i.e. the keyword is required before each command); the activation command simply sets it back to 'keyword optional' mode.
I figure that as long as I choose a keyword that's unique enough that it (and things that sound like it) won't come up too often in conversation, the computer will ignore anything I say once I activate the keyword requirement. Then, I only need to say the keyword once along with the reactivation command to bring it out of this 'ignore commands' state.
In a way, this is actually better, because it allows me to continue to use voice commands if I really need to, as opposed to having to say the reactivation command to get my computer to listen to me, saying my intended command, and then saying the deactivation command to make it ignore me again.
All the same, if anyone has answers to any of my two original questions, please don't hesitate to share your knowledge! It might come in handy later!

Speech Recognition stops paying attention.

hey all--
I was having trouble with my speech recognition (as well as a few other things) and so I finally deleted and reinstalled my system out of frustration. There are a lot of commands in the voice commands I don't want/use so I took some of them out or changed their wordings. I am also adding ones of my own.
My main trouble is that after a little bit, or 4-5 spoken commands, it stops listening to me. The green indicators indicate that it is hearing me just fine, but suddenly I can say "computer" all I want and it won't do a darn thing. If I open speech prefs, turn speakable off and back on, that'll fix it for another 4-5 commands. I am not in a noisy environment. It is very irksome!!
~G

OK, I have one more question now, seems to be the same as this question which was never answered years ago:
Speakable Commands>Application Switching -- how to delete items?
When I click on the triangle from the speech command widget and open speech commands, there is a TON of stuff there that I'll never ever use, but these things do not appear in the Library/Speech/Speakable Items folder so that I might be able to delete them. For example, one item I'd like to turn off is just the word "Home" and the computer keeps thinking that I said "Home" instead of whatever else I meant to say. Another keeps opening iChat or Mail, when I say "open new window" or something like that. I tried clicking, control clicking, option clicking, nothing will highlight a command in the Speech Commands window so that I might delete it.
The "Home" command only shows up when I'm using Word, in the "front window" list. However, the Application Speakable Items/Microsoft Word folder is empty. Is there some other way to get into the speakable commands list and weed out unwanted stuff?

Java speech recognition software ?

Hi Software Experts.
I have 2 short questions.
1. does any one here have any knowledge of any java speech recognition software ?
2. is it true that all speech recognition software has to be trained ?
Are there some simple ones on the market that will test for 1 word and therefore not need to be trained ? Or trained a lot less ?

IBM makes Via voice, a commercial package. Almost all apps need some training, to be truely effective, as the user's speech patterns vary. I am not aware of any Sun classes for this, there may be some 3rd party classes out there.... - Bart

How can I spell out names in pages dictation

I need to spell out various peoples names in pages by voice dictation, not by typing on keyboard. Is there a verbal cue to help with this?

There is no annunciate on/off voice command that informs OS X dictation to stop translating spoken letters to guessed words.
In my experience with pages dictation, and in particular, enhanced dictation (local), there just isn't sufficient interpretation for it to identify accurately, individually annunciated letters that comprise a name. I am a patient person, and it frustrated me. Dictation does learn your voice patterns, and after some point in time, it might become more accurate.
If you don't have it already, here are some Apple dictation recognition commands.

Speech recognition for a programmer with RSI (not how to do in Java)

Guys, someone that I know has developed RSI and has almost had to drop out of programming.
Has any other programmer here been faced with this and been able to work around it? In particular, have you found a decent speech recognition system that allowed you to continue to work?
Note that I am not interested in learning how to implement speech recognition in a Java program, which is what [most of the posts here|http://forum.java.sun.com/thread.jspa?threadID=5310499&tstart=0] are concerned with.
I am aware of Dragon Naturally Speaking, but there are some terrible reviews of it which have made me nervous. But if any of you give a glowing review, I will reconsider.
Since this is a java forum, and one of the best academic research projects in this field is [Sphinx-4|http://cmusphinx.sourceforge.net/sphinx4/] which is based on Java, does anyone know of an end user application comparable to Dragon Naturally Speaking which uses Sphinx-4?

pm_kirkham wrote:
No, we just get an intern and pair-program.Thats the first good reason to pair program that I have heard! (I like most of the rest of XP, but PP normally drives me insane...)

Speech recognition problem with null pointer exception

im trying to run the simple hello program that came with IBM speech for java
i've put the speech.jar file in the right folder and it compiles all fine but when i run it i get this error:
locale is en_US
java.lang.NullPointerException
at Hello.main(Hello.java:171)
but i dont know enough about java speech to figure out what is null and what shouldn't be null...
thx
nate
P.S.
here is the program code
import java.io.*;
import java.util.Locale;
import java.util.ResourceBundle;
import java.util.StringTokenizer;
import javax.speech.*;
import javax.speech.recognition.*;
import javax.speech.synthesis.*;
public class Hello {
static RuleGrammar ruleGrammar;
static DictationGrammar dictationGrammar;
static Recognizer recognizer;
static Synthesizer synthesizer;
static ResourceBundle resources;
static ResultListener ruleListener = new ResultAdapter() {
public void resultAccepted(ResultEvent e) {
try {
FinalRuleResult result = (FinalRuleResult) e.getSource();
String tags[] = result.getTags();
if (tags[0].equals("name")) {
String s = resources.getString("hello");
for (int i=1; i<tags.length; i++)
s += " " + tags;
speak(s);
} else if (tags[0].equals("begin")) {
speak(resources.getString("listening"));
ruleGrammar.setEnabled(false);
ruleGrammar.setEnabled("<stop>", true);
dictationGrammar.setEnabled(true);
recognizer.commitChanges();
} else if (tags[0].equals("stop")) {
dictationGrammar.setEnabled(false);
ruleGrammar.setEnabled(true);
recognizer.commitChanges();
} else if (tags[0].equals("bye")) {
speak(resources.getString("bye"));
if (synthesizer!=null)
synthesizer.waitEngineState(Synthesizer.QUEUE_EMPTY);
Thread.sleep(1000);
System.exit(0);
} catch (Exception ex) {
ex.printStackTrace();
int i = 0;
String eh[] = null;
public void resultRejected(ResultEvent e) {
if (eh==null) {
String s = resources.getString("eh");
StringTokenizer t = new StringTokenizer(s);
int n = t.countTokens();
eh = new String[n];
for (int i=0; i<n; i++)
eh = t.nextToken();
if (((Result)(e.getSource())).numTokens() > 2)
speak(eh[(i++)%eh.length]);
static ResultListener dictationListener = new ResultAdapter() {
int n = 0; // number of tokens seen so far
public void resultUpdated(ResultEvent e) {
Result result = (Result) e.getSource();
for (int i=n; i<result.numTokens(); i++)
System.out.println(result.getBestToken(i).getSpokenText());
n = result.numTokens();
public void resultAccepted(ResultEvent e) {
Result result = (Result) e.getSource();
String s = "";
for (int i=0; i<n; i++)
s += result.getBestToken(i).getSpokenText() + " ";
speak(s);
n = 0;
static RecognizerAudioListener audioListener =new RecognizerAudioAdapter(){
public void audioLevel(RecognizerAudioEvent e) {
System.out.println("volume " + e.getAudioLevel());
static EngineListener engineListener = new EngineAdapter() {
public void engineError(EngineErrorEvent e) {
System.out.println
("Engine error: " + e.getEngineError().getMessage());
static void speak(String s) {
if (synthesizer!=null) {
try {
synthesizer.speak(s, null);
} catch (Exception e) {
e.printStackTrace();
} else
System.out.println(s);
public static void main(String args[]) {
try {
if (args.length>0) Locale.setDefault(new Locale(args[0], ""));
if (args.length>1) Locale.setDefault(new Locale(args[0], args[1]));
System.out.println("locale is " + Locale.getDefault());
resources = ResourceBundle.getBundle("res");
recognizer = Central.createRecognizer(null);
recognizer.allocate();
recognizer.getAudioManager().addAudioListener(audioListener);
recognizer.addEngineListener(engineListener);
dictationGrammar = recognizer.getDictationGrammar(null);
dictationGrammar.addResultListener(dictationListener);
String grammarName = resources.getString("grammar");
Reader reader = new FileReader(grammarName);
ruleGrammar = recognizer.loadJSGF(reader);
ruleGrammar.addResultListener(ruleListener);
ruleGrammar.setEnabled(true);
recognizer.commitChanges();
recognizer.requestFocus();
recognizer.resume();
synthesizer = Central.createSynthesizer(null);
if (synthesizer!=null) {
synthesizer.allocate();
synthesizer.addEngineListener(engineListener);
speak(resources.getString("greeting"));
} catch (Exception e) {
e.printStackTrace();
System.exit(-1);

yes, the problem seems to be with allocating the engine, but I don't know exactly why recognizer.allocate() returns null...I have already installed IBM ViaVoice and Speech for Java, and have set the supposedly correct paths...

How to set up Speech Recognition?

Referring to following link for detailed discussion,
I would like to know if I create a batchfile in following directory, then after getting through those training, does anyone have any suggestions on how to set where the speech locates
this batch file?
Does anyone have any suggestions?
Thanks, to everyone very much for any suggestions :>
C:\script\run.bat
http://answers.microsoft.com/en-us/windows/forum/windows_7-desktop/how-to-set-up-speech-recognition/d3a5693e-c17b-4f5b-84fd-9c1c262ae9ba
Thanks in advance for any suggestions

Hi,
You can try to add new words to open bat file in Speech dictionary:
Say "start listening," or tap or click the microphone button to start the listening mode.
Say "open Speech Dictionary" and do any of the following:
-To add a word to the dictionary, say "Add a new word," and then follow the instructions.
For details commands you can use, please refer to this article:
How to use Speech Recognition
http://windows.microsoft.com/en-IN/windows-8/using-speech-recognition
Kate Li
TechNet Community Support

What is Speech Recognition and how do you use it?

Hello, people. I was on my System Preferences and just out of curiosity, I clicked on Speech. Then there was a tab on the right called Speech Recognition. What is Speech Recognition and how do you use it?

When you turn it on the default should be listen only when ESC key is pressed. Hold the ESC key and say, in your normal voice, "show me what to say".
For fun.. say "tell me a joke"
Mac OS X 10.6 Help: About speech recognition
About speech recognitionhttp://docs.info.apple.com/article.html?path=Mac/10.6/en/8355.html
Mac OS X 10.6 Help: Speech Recognition Commands preferences
Speech Recognition Commands preferenceshttp://docs.info.apple.com/article.html?path=Mac/10.6/en/8407.html
Mac OS X 10.6 Help: Speech Recognition Settings preferences
Speech Recognition Settings preferenceshttp://docs.info.apple.com/article.html?path=Mac/10.6/en/8408.html
Mac OS X 10.6 Help: Calibrating speech recognition to your environment
Calibrating speech recognition to your environmenthttp://docs.info.apple.com/article.html?path=Mac/10.6/en/8406.html
Mac OS X 10.6 Help: If the Speech Recognition pane is empty
If the Speech Recognition pane is emptyhttp://docs.info.apple.com/article.html?path=Mac/10.6/en/8907.html

Speech recognition not working properly!

I just started trying out the speech recognition tool in Snow Leopard and I seem to be having problems with it. When I calibrate it (using the internal mic) it works fine and understands those phrases perfectly but when I actually say the exact same thing outside of the calibration process the computer doesn't seem to understand. I know it's picking up my voice because the little widget thing animates as is does when it hears sound and the "Speakable Items" option is set to on . It actually worked once yesterday when I tried it for the first time, I told it to switch to an app and it did but that was the end of it.
What should I do? I haven't updated the software to 10.6.1, would this perhaps solve the issue? I don't know if it worked before I upgraded to Snow Leopard because this is the first time I've tried it out

Had the same problem. Could only resolve by going back to Snow Leopard.

Speech Recognition Not Working. Why?

I have a 6-month old Macbook (OSX 10.5.5), and I want to try out the built in speech recognition software. For some reason, it won't accept any input, no matter the configuration I put it in. When I click calibrate, it's slow to open, and no matter what I do with the slider, it won't recognize my commands. I've tried it in a new account, but without success. It appears to be system-wide. Also, there are no items in the speakable items folder, and a feedback box does not appear on the desktop. Any ideas?

Im assuming you've used the Sound pane to check that the built-in mic is working, and that you've remembered to click on the "Speakable Items" radio button.
There are a couple of fixes that normally work when changing user helps, but it is possible that two user accounts have the same problems!
1)
Have you got any items in Address Book that are marked as "Company" but have no company name in them? This is an old problem that was supposed to have been fixed with Leopard, but I've heard people complaining about it occasionally since. If there are, try going through them and putting in the company names, then log out and back in again.
2)
Try deleting the preference files
com.apple.speech.recognition.recognizer.prefs.plist
com.apple.speech.recognition.AppleSpeechRecognition.prefs.plist
com.apple.speech.recognition.feedback.prefs.plist
all in ~/Library/Preferences/
Failing those, and the missing Speakable Items suggests that's not the fault, try downloading and reapplying the latest Combo updater.
If that doesn't do the trick, you may have to reinstall the OS.
Sorry!
Archie

Speech Recognition - Spell out words

Similar Messages

Maybe you are looking for