WSR Add Word to Dictionary
Printed From: MSSpeech-Forum
Category: Windows™ Speech Recognition Forums
Forum Name: WSRToolkit
Forum Description: Questions and Answers for using the WSRToolkit features
URL: https://www.msspeech-forum.com/forum_posts.asp?TID=68
Printed Date: 15/Dec/2024 at 11:55am
Software Version: Web Wiz Forums 12.02 - http://www.webwizforums.com
Topic: WSR Add Word to Dictionary
Posted By: GG
Subject: WSR Add Word to Dictionary
Date Posted: 28/Feb/2009 at 9:38am
|
I am using the WSR toolkit and I am adding words to the dictionary. I am unable to find where the dictionary file(s) that I am creating are located. Any help would be greatly appreciated. I have also tried importing files (txt) into the WSR toolkit and I can not locate where this information is being stored. |
Replies:
Posted By: mmarkoe
Date Posted: 28/Feb/2009 at 11:47am
There are several ways to edit your added words.
This information becomes part of your speech profile. There is nothing to be looked at.
The Train From Text function improves the Acoustic Model of your speech profile. This means the unique way you pronounce words within the style of text you generally dictate is integrated into your personal speech profile.
The Add From File allows you to feed your typical documents so the Language Model or your personal speech profile is improved. This tool parses the documents for not just the words you use, but the context of surrounding words. This increases the probability of the words being recognized correctly.
Marty Markoe, eMicrophones, Inc. ------------- Marty Markoe, MVP Microsoft Valued Partner See us at: http://www.mymsspeech.com |
GG wrote:Posted By: GG
Date Posted: 28/Feb/2009 at 1:47pm
|
How do I know that the add to file was successful?
How do I use the dictionary on another computer?
How do I use the words I added froma file on another computer?
Would I use this to add a bunch of phrases that I use often?
Suppose I want to create my own individual vocabulary files (medical, legal, programming), is this the same adding the words to the dictionary or adding from file.
How do the medical vocabulary files that are sold on your site integrate with the files I have already created.
Is there a way to add the words to the dictionary without having to click the appropriate buttons for each word?
Can I share the files I create with anyone else to use with their WSR?
|
Posted By: mmarkoe
Date Posted: 28/Feb/2009 at 4:36pm
If you see the words you added (and added pronunciations to) are recognized properly
At this time you do not. This is one of the drawbacks to the free Windows® Speech Recognition built into Vista. The work group at Microsoft knows they need this feature. It is something they will probably add at some point are not saying when.
You could use the Add To Dictionary feature to add phrases.
No. You would need to Software Developers Kit for SAPI (speech application programming interface). It is not just a matter of getting words, it is also having linguists gather words and phrases and styles of speaking from all over the land so that it picks up different dialects and styles. You should be able to do this for approximately $100,000.
Directly into your speech profile. They should not conflict.
Sorry, I have no idea what you were trying to ask. Buttons, where, whose?
Only if you give them lists to feed in through the WSRToolkit.
Marty ------------- Marty Markoe, MVP Microsoft Valued Partner See us at: http://www.mymsspeech.com |
Posted By: GG
Date Posted: 28/Feb/2009 at 11:58pm
|
Looks like I am just full of questions, the more you answer the more questions I have. Now I am having a difficult time understanding this concept How do the medical vocabulary files that are sold on your site integrate with the files I have already created. Directly into your speech profile. They should not conflict. How are the vocabulary files added to my speech profile. Would I have to add the files through the add a file. If I decided to purchase a vocabulary file how does it get added to my speech file. This is confusing to me because if my profile is "my profile" how can something be added to it that is not actually "my voice"? Perhaps I am missing the point here. If I am understanding you correctly the WSRToolkit seems to be the basis for adding large quantities of information to WSR . By the way I love the way it works, but are there any plans in the future for adding anything else to it? What is the dictionary tool (for handwriting recognition) and how does it work as far as adding information to WSR. Or since I am using your toolkit should I just forget about using the dictionary tool. When I referred to the buttons, I meant the steps that are part of adding the words to the dictionary. |
Posted By: mmarkoe
Date Posted: 01/Mar/2009 at 11:53am
Here are 2 resources for you to explore:
1. The http://www.microsoft.com/speech/speech2007/default.mspx - Home Page for Microsoft SAPI
2. http://www.microsoft.com/communities/newsgroups/list/en-us/default.aspx?dg=microsoft.public.speech_tech&cat=en_us_dfd98a36-81e8-4006-a42b-e08a29afeb2b&lang=en&cr=us - Microsoft SAPI User Groups
In a nutshell to answer your questions:
1. The WSRToolkit is not meant to add vast amounts of information. It is not a language model builder. It is meant so that any professional can add anywhere from 20 to 100 documents to improve the likelihood that their typical type of speaking in terms of the context of the words they use will be improved. The train from text function as mentioned several times, parses the way you pronounce words and add that to the acoustic model which is separate from the language model which is separate from the added words. 2. We have a list of eight features we would like to add to the WSRToolkit. If you'll notice, the cost of the toolkit is $15.99. Do you know of any other software that does as much as the toolkit and costs as little? It was a labor of love and cost many thousands of dollars which is still a long way off from breaking even. Our list of eight items is based on things we have been asked for and we feel may be doable. I would not expect to see the update/upgrade for at least one year as the official release of the WSRToolkit has only been six months. Please do not ask us any more questions about this because I think I've made it clear that we do not preannounce as we do not wish to look like fools if it turns out something is extremely expensive, not likely to help the vast majority of people who use it, or just plain too complex to accomplish. There was another speech recognition specialist who has been claiming for 2 1/2 years that they will be releasing a command module, like they have for Dragon, for Windows Speech Recognition. As far as we know, that is vaporware and has always been vaporware. eMicrophones quietly and diligently worked on the WSRToolkit and only announced it when it was ready to be tested.
3. We are sorry to hear that you feel there are too many buttons to press to add words to the dictionary. Our software just uses the code from Microsoft so you will have to take it up with them. Frankly, I have added dozens of words and phrases and never noticed it to be a chore. As a matter of fact, I do most of the adding using my voice. I will now check your other two postings. However, if they have similar questions, I will refer you back to this thread.
Sincerely, Marty Markoe, eMicrophones, Inc. ------------- Marty Markoe, MVP Microsoft Valued Partner See us at: http://www.mymsspeech.com |
Posted By: GG
Date Posted: 01/Mar/2009 at 10:33pm
|
Import a full folder: Well I can understand that this was not part of the original plan, but I hope that this is something that you will consider. I have purchased several products over the years that allow this feature...and hope that you will consider it as a possibility for the WSR Toolkit. I do not feel there are too many buttons, but I would like to be able to implement this process and walk away to work on other things is possible. I am a multi tasker and entering each of the words can be very time consuming if the list of words is long. Also if someone else is in the same room it is distracting to me to try to work on adding the list and answer their questions are well. Interuptions make it difficult at times. Perhaps I should practice adding more words using my voice. Okay I will not ask about your pending features, but I must say that I am curious to see what the new additions will be. When I ask you questions I am not criticizing your product, just trying to get a better understanding of how it works and how I can get it to work the way I need it to work for me. |
Posted By: mmarkoe
Date Posted: 02/Mar/2009 at 9:20am
We did not see any criticism. You even said several times you liked the WSRToolkit.
From a developers point of view, application development is time and money intensive. Therefore features that one or two people need might take time away from development of features that many people can use. I suggested you acquire:
This is an execellent place for you to learn how to make WSR work for you.
Marty ------------- Marty Markoe, MVP Microsoft Valued Partner See us at: http://www.mymsspeech.com |
Posted By: CapnBuck
Date Posted: 08/Jul/2009 at 7:09pm
|
When I use voice commands to try to add words to dictionary through WSRToolkit, WSR minimizes herself and turns off.....therefore, unusable.
She does other strange and erratic things since installing WSRToolkit (this is NOT critical of the program....I like it. Just wondering if others have similar happenings)
She now loves to press the end key.
A command from another Macro will appear in her text feedback area while she is (for example) speaking the local WX forecast and then she will respond to that command (if verbal response is called for, she will speak it). Giving herself commands and then carrying them out.
After completing nearly any command, she will display her alternates panel with many sound alike items from something in the previous action.
She might display the mousegrid for no reason.
She might (when supposedly asleep) open Windows Media Player and choose a song.
Several Macros made 'just for fun' - - "Hello, ------ " / "Hello, Bill", "thank you" / "You're welcome" I have had to delete because of her using them or parts of them without command.
Some of these irritants had begun before WSRToolkit (I am NOT blaming this program....rather asking for assistance. Windows Vista, Vista Speech Recognition Program, one MS default voice, one commercial voice from Cepstral, basic microphone and speakers. I have had great success from the beginning and am sure it will continue.
Thanks for any help,
CapnBuck ------------- CapnBuck |
Posted By: mmarkoe
Date Posted: 08/Jul/2009 at 10:46pm
WSR and WSRToolkit are not animate objects and are not imbued with human emotions like love, hate or spite. It sounds as if the program is getting confused as you have added so many commands and voices. It is just plain misinterpreting as commands what you are saying or what readback voices are saying.
The things to do are:
Marty ------------- Marty Markoe, MVP Microsoft Valued Partner See us at: http://www.mymsspeech.com |
Posted By: CapnBuck
Date Posted: 08/Jul/2009 at 11:57pm
|
"It sounds as if the program is getting confused as you have added so many commands and voices."
I had 2 commands that I mentioned as 'fun' (deleted). And as for the "so many" (all useful Macros copied DIRECTLY from the Macros libraries), how in the world would you know? Vista comes with TWO voices....I have TWO voices.
I don't know what your problem is, but try to just stick to my initial complaint about 'it' minimizing itself and not being usable in adding words to the dictionary? Do you have ideas or suggestions about THAT?
------------- CapnBuck |
Posted By: mmarkoe
Date Posted: 09/Jul/2009 at 7:11am
That sounds as if you're getting a little touchy. Your posting was not clear and had to be read several times to get the gist of it.
First of all, I believe my suggestion about the possible confusion for trigger words and commands is valid for your basic problem. Something you are saying is likely being interpreted as triggering the minimize command. The fact that you copied many macros does not preclude the possibility of similar sounding trigger words as the macros were not written by one person but many different people.
I see no mention of your considering suggestions number one and three. If you find what is triggering the minimize command and it relates to another macro, then the hyphen suggestion and enunciating very clearly is quite valid to solve your problem.
Another method to confirm whether there's a conflict with trigger words from macros, is to not load the macros program. If the window does not minimize then there is a likely conflict with a similar sounding trigger word.
Finally, if you are using a headset microphone, make sure the microphone element is off to the corner of your mouth and not in front of it. If it is in front of your mouth it may be picking up inhales and exhales as extraneous words/commands. Also, before starting to dictate, run the Set Up My Microphone wizard to set the correct volume levels for your voice and the background noise.
Marty ------------- Marty Markoe, MVP Microsoft Valued Partner See us at: http://www.mymsspeech.com |
Posted By: CapnBuck
Date Posted: 09/Jul/2009 at 3:48pm
|
I am not asking questions about my feelings....only the questions and problems at hand. I do enunciate each word clearly. I did hyphenate the phrase Add-Words. The speech recognition engine still minimizes and turns off 'while configuring add to dictionary'. When I click cancel on the Toolkit, the speech recognition engine wakes up with the text area message '....finished configuration...'. The speech recognition engine works as advertised with all other parts of the Toolkit. I removed ALL Macros from the macros folder and performed the above actions with NO macros on my system. I am not able to edit the commands in the Toolkit....the command Add Words is part of the Toolkit. I can close WSRToolkit and command 'open the speech dictionary' and 'it' will. I can then perform all functions.....but NOT in the WSRToolkit. Thank you for any consideration you might give this problem. ------------- CapnBuck |
Posted By: mmarkoe
Date Posted: 09/Jul/2009 at 4:23pm
In your first message you wrote:
"When I use voice commands to try to add words to dictionary through WSRToolkit, WSR minimizes herself and turns off.....therefore, unusable.
She does other strange and erratic things since installing WSRToolkit (this is NOT critical of the program....I like it. Just wondering if others have similar happenings)
She now loves to press the end key."
We were not talking about your feelings. We were talking about you imbuing WSR with a sense of humanity when you say things like, "WSR minimizes herself......" and "She now loves to press the end key." Our point was WSR is not a thinking/feeling entity, just software programmed by people.
Good to know.
I see now your main question was actually the first sentence in your first message:
"When I use voice commands to try to add words to dictionary through WSRToolkit, WSR minimizes herself and turns off.....therefore, unusable."
I just tried adding a word through the WSRToolkit Add-Words function. I was able to move left and right from one to the next by using the commands, "Move right" and "Move left." I could move down to the text input field by just saying the command, "Tab."
In the text field for Add-Words, I dictated Michael Mendick and it originally came up as, "Michael mendik." I corrected "mendik." At no point did the toolkit minimize itself or turn itself off. If I am not understanding you, please correct what I need to try to duplicate your problem. If in fact I do understand your problem, you can try on installing the toolkit from, Control Panel, Programs and Features. Then try reinstalling. If the problem persists, perhaps you have another computer you can duplicate the problem. Let us know how this goes?
Marty ------------- Marty Markoe, MVP Microsoft Valued Partner See us at: http://www.mymsspeech.com |