MSSpeech Forum : EMR and Vista speech https://www.msspeech-forum.com/ Copyright (c) 2006-2013 Web Wiz Forums - All Rights Reserved. Tue, 05 May 2026 00:36:35 +0000 Sun, 10 May 2009 13:40:05 +0000 http://blogs.law.harvard.edu/tech/rss Web Wiz Forums 12.02 360 https://www.msspeech-forum.com/RSS_post_feed.asp?TID=27 <![CDATA[MSSpeech Forum]]> https://www.msspeech-forum.com/forum_images/msspeech_forum.png https://www.msspeech-forum.com/ <![CDATA[EMR and Vista speech : &gt;&gt;However text still a no...]]> https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post411.html#411 Author: cmwilkerson
Subject: 27
Posted: 10/May/2009 at 1:40pm

>>However text still a no go, unless I want to use the Alternate Panel in my EMR.<<
 
No longer an issue with Windows 7 WSR. Thumbs Up
]]> Sun, 10 May 2009 13:40:05 +0000 https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post411.html#411 <![CDATA[EMR and Vista speech : I&#039;m using WSR on Windows...]]> https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post398.html#398 Author: cmwilkerson
Subject: 27
Posted: 09/May/2009 at 11:33am

I'm using WSR on Windows 7 with a Lenovo Tabletpc with internal array mic and 4 gigs of RAM. It works great but needs a medical lexicon. For example, the word, "radiculitis," is seen as "ridiculous." Using the TabletDictionary's import tool, I imported a 1000 medical words.

After the import, radiculitis is found, as are every medical word present in my import!!!
You can read about it here: http://www.medicaltabletpc.com/


Edited by cmwilkerson - 10/May/2009 at 1:42pm]]> Sat, 09 May 2009 11:33:39 +0000 https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post398.html#398 <![CDATA[EMR and Vista speech : robch_ms wrote:You can delete...]]> https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post198.html#198 Author: cmwilkerson
Subject: 27
Posted: 20/Nov/2008 at 9:33am

Originally posted by robch_ms robch_ms wrote:

You can delete your profile (and/or create a new one) in the Advanced Speech Control Panel. Say, "Show Speech Options", "Configuration", "Open Speech Control Panel", "Advanced Speech Options". Then, you can use "New" to create a new profile, or "Delete" to delete the selected profile.
 
If you want to remove words from you user lexicon, you can export the speech dictionary with the "Speech Dictionary.wsrMac" macro on http://code.msdn.microsoft.com/wsrmacros.
 
--rob
 
0utstanding! Thanks so much! Guess I lost sight of the forest through the trees. Smile


Edited by cmwilkerson - 20/Nov/2008 at 9:34am]]> Thu, 20 Nov 2008 09:33:29 +0000 https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post198.html#198 <![CDATA[EMR and Vista speech : You can delete your profile (and/or...]]> https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post197.html#197 Author: robch_ms
Subject: 27
Posted: 20/Nov/2008 at 9:26am

You can delete your profile (and/or create a new one) in the Advanced Speech Control Panel. Say, "Show Speech Options", "Configuration", "Open Speech Control Panel", "Advanced Speech Options". Then, you can use "New" to create a new profile, or "Delete" to delete the selected profile.
 
If you want to remove words from you user lexicon, you can export the speech dictionary with the "Speech Dictionary.wsrMac" macro on http://code.msdn.microsoft.com/wsrmacros.
 
--rob
]]> Thu, 20 Nov 2008 09:26:49 +0000 https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post197.html#197 <![CDATA[EMR and Vista speech : &#034;I have been told that performance...]]> https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post196.html#196 Author: cmwilkerson
Subject: 27
Posted: 20/Nov/2008 at 9:19am


"I have been told that performance will begin to suffer around about 1,000 words in the lexicon but the addition of words to the language model (context) is unlimited."
 
In my experience, there can be an inverse recognition to medical and lay recognition. In other words, if you fill your dictionary with medical words, the lay output can suffer. Perhaps this subjective experience is because of, "Overload." Or it may be brain fatigue on my part?
 
Is there any way to delete an entire profile or dictionary?
 
Knowing what I know now, I would be much more discretionary about what I import and scan.
]]> Thu, 20 Nov 2008 09:19:01 +0000 https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post196.html#196 <![CDATA[EMR and Vista speech : cmwilkerson wrote:Thanks...]]> https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post195.html#195 Author: BradT
Subject: 27
Posted: 20/Nov/2008 at 9:10am

Originally posted by cmwilkerson cmwilkerson wrote:

Thanks for the explanation Rob!
 
So the trick to optimizing WSR with Medical is:
1. Import a pertinent dictionary-Lexicon
2. Scan prior medical documents into WSR so the words can be seen in their overall frequency of occurrence within the context of other words. (words often related to other words).


Maybe Rob will validate this, but I believe the best way with WSR is to:

1. Scan prior medical documents into WSR so the words can be seen in their overall frequency of occurrence within the context of other words. (words often related to other words).

Then, after certain words/phrases fail repeatedly, add them to the lexicon.

I have been told that performance will begin to suffer around about 1,000 words in the lexicon but the addition of words to the language model (context) is unlimited.


]]> Thu, 20 Nov 2008 09:10:40 +0000 https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post195.html#195 <![CDATA[EMR and Vista speech : Thanks for the explanation Rob!...]]> https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post194.html#194 Author: cmwilkerson
Subject: 27
Posted: 20/Nov/2008 at 8:56am

Thanks for the explanation Rob!
 
So the trick to optimizing WSR with Medical is:
1. Import a pertinent dictionary-Lexicon
2. Scan prior medical documents into WSR so the words can be seen in their overall frequency of occurrence within the context of other words. (words often related to other words).


Edited by cmwilkerson - 20/Nov/2008 at 8:59am]]> Thu, 20 Nov 2008 08:56:25 +0000 https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post194.html#194 <![CDATA[EMR and Vista speech : Sorry, Marty. I tried to make...]]> https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post192.html#192 Author: robch_ms
Subject: 27
Posted: 18/Nov/2008 at 10:47am

Sorry, Marty. I tried to make that clear in my previous post.
 
For Speech, there are 2 things:
1.) Lexicon - This the list of words that can be used. Each word has an optional pronuncation. If a pronuncation isn't specified we'll use LTS (letter to sound) rules to "guess" a pronuncation.
2.) Language Model - This is the relative frequency of occurrence of words in the context of other words.
 
The DictionaryTool that Dr. Wilkerson mentioned is simply adding items to the lexicon (without specified pronuncations). He's calling that a vocabulary. That's a reasonable thing to call it, although, it's not what we call it in Vista for speech. We call it a dictionary, sort of like a dictionary that you'd pick up at a local bookstore. It has words, their spellings, and their pronuncations in it.
 
--rob
]]> Tue, 18 Nov 2008 10:47:41 +0000 https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post192.html#192 <![CDATA[EMR and Vista speech : robch_ms wrote:Sorry... Sometimes...]]> https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post191.html#191 Author: mmarkoe
Subject: 27
Posted: 18/Nov/2008 at 10:11am

Originally posted by robch_ms robch_ms wrote:

Sorry... Sometimes the wrong information gets out of MS... Actually, the two systems are very very linked.
Thank you for letting us know of your new take on the initial question.
 
I would like some further clarification. It seems that Dr. Wilkerson is calling this a vocabulary. It seems to me that this is just a matter of adding words? This is not a true language model which includes the context associations as well as the pronunciations?
 
Marty
]]> Tue, 18 Nov 2008 10:11:52 +0000 https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post191.html#191 <![CDATA[EMR and Vista speech : Sorry... Sometimes the wrong information...]]> https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post190.html#190 Author: robch_ms
Subject: 27
Posted: 18/Nov/2008 at 9:49am

Sorry... Sometimes the wrong information gets out of MS... Actually, the two systems are very very linked.
 
I can confirm that the two sub-systems (both Handwriting and Windows Speech Recognition) do indeed use the same underlying store for supplemental vocabularies.

It's kind of funny, in fact... After finishing up the re-write of the Speech platform back in the late 90s in the speech team, I left that team, and landed in the Tablet PC team, where I led about half the platform development team for handwriting platform features (in addition to the very early implementations of Speech Recognition on Tablet PC).

The Handwriting team needed a system to enable the same types of features for a dictionary that we had developed for the Speech Recognizer but instead for the Handwriting Recognizer. So ... Instead of duplicating the speech dictionary code, or building something new, we just use the exact same store (for lexical entries). Not only was it easier to develop (because it was already done), it's a great end user sharing feature as well.

In addition, both correction experiences are tied together. For example, if you were to dictate something to the speech recognizer, and then select it for correction with the handwriting recognizer, you'll see the speech alternatives in the handwriting UI. Same is true in reverse. If you handwrite something and insert it into the document, you can select it with the pen and say "Correct that", and you'll see the handwriting alternatives in the speech correction UI.

]]> Tue, 18 Nov 2008 09:49:17 +0000 https://www.msspeech-forum.com/emr-and-vista-speech_topic27_post190.html#190