I wish there was some help around the corner with #1 in your post.  This has been my #1 problem at least for the last 2 versions.  Like you, I have tried to train DNS but I've come to the conclusion that Nuance put the Spell That function in there just to make us believe something is happening - like a placebo.

The experts here will tell you to train these words in combination with other words - in phrases.  Quite honestly, I haven't found this to work.  They will also say to make sure you are using the best mic (and audio card).  I think part of my problem is the somewhat weak audio card built into my Toshiba M200 laptop.  I've tried to get around that by using USB microphones such as the Phillips SpeechMike.

'and', 'in', and 'an' do some different the way I enunciate them yet DNS just does not want to cooperate.  

I am surprized that Lunis has not jumped in yet with his special solution: add phrases to your vocabulary. This is the only way to help the Dragon with the small words.

Just look at it this way: whenever your Dragon makes a mistake, he needs your help. And the help must be to extend the acoustical signals representing the words.

And phrases do have more acoustical information than small words.

Training in context is ok, but adding phrases is much better.

With the best wishes from Augsburg/Bavaria

Willi

In this case (small words - and, in, he, she, a, the, etc.), adding phrases is not likely to help.
This is because there are thousands of phrases involving small words.

Likewise there is no point in training single words.

Adding a phrase will help for that particular phrase, but most likely will decrease recognition for that particular small word. For example, if you add "He said" recognition might increase for "He said", but will decrease for "She said" or "He had" or .....

Once again, the most important factor is good dictation technique. Dictate in smooth fluent phrases and take time to enunciate.

I would consider this a less than optimal setup for serious workflow production.
Although DNS will work on a MacBook with Bootcamp, it is really designed for a PC.

I would also question the Andrea external soundcard, most reports of the Andrea Soundcards are less than optimal for use with DNS. This is probably your problem with small words as well as SR speed. The microphone experts will help you here.

Finally, I love my 12" Tosihba Laptop but it is not optimal for work production. Maybe a quick dictation done at home but not for 2-3 hours of report preparation. For work production, I dictate using two High Resolution Sony 21" Monitors.

Bottom line is how much is your time worth? If you are serious about using DNS for work production, get the right hardware. The Intel Core 2 Duos/Quads are the biggest bang for the buck for SR in a long time. 

Sounds good here.

The trick to templates is Automation. Right now, you are using very basic templates. It also sounds as if they are sluggish on your MacBook. You should be able to move from field to field almost instantly.

You have two options here. The first is to pay a VAR to help write some custom templates for you. This probably won't be cheap, but will be cost effective in the long run.

The second option is to write your own. Again a bit of a learning curve but you probably can get a lot of help on the forums. Bottom line here is that you have to enjoy writing the code and making it work. Before everyone else jumps in, the best starting place is:

by Larry V. Allen

So to answer your original question, you probably are not doing anything wrong but there is a lot that you can do to increase your productivity using DNS.

so what's wrong with adding thousands of phrases? The acoustical content of "he said" versus "she said" or "he had" is always more informative than the small words alone. For our customized vocabs we use a tool listing the 2- or 3-word-combinations in the frequency order they occur in the customer's texts and add the most frequent ones to the vocab. Works great for many users, in addition to a good dictation style of course.

Willi

You bring up a good point. This is ideal for people like you and myself who do not save dictation with documents. However, there are couple of caveats to this approach.

1. If you set the "Disk space reserved for playback" to zero instead of the default 40 MB, then you are relying on the DNS speech buffer to store your playback. The problem with this is that if the speech buffer reaches its limit, DNS starts removing speech data from the buffer beginning with your earliest recorded utterances. What this means is that if you are dictating a very long document, playback within the document will be limited to your last approximately 100 utterances. Therefore, if you dictate 20 pages, only the last 100 utterances will be available in that document for playback. The result will be that at a certain point speech data will not be available until you save the document with a linked dra file and reopen it. Not a problem for most. However, if you want to playback your dictation on page 1, it won't be available under this setting until or unless you save the document with your dictation. In other words, if you want to go back to page 1 and make a correction, the playback will not be available, and will only be available up to the point of your last 100 utterances if you haven't saved your document with the dictation dra file.

2. I've never experimented with this, but there may be other issues relative to the fact that it may effectively disable the "Always preserve wave file data" in the Advanced settings. Nevertheless, this is definitely one way of preventing the prompt that WayneE was concerned about in that post on the ScanSoft public forum.

Therefore, you would have to experiment with this further to determine what the ultimate consequent effect might be.

Chuck Runquist
Former DNS SDK & Senior Technical Solutions PM for DNS

If you hear the sound of hoofbeats, think horses not zebras.
Law of Parsimony (Occam's razor)

That setting doesn't necessarily improve speed or performance.  The space reserved for playback is treated as a "cache."  As such, the only impact that it has at all on performance (speed) is if the user has "Save recorded dictation with document" set to either Ask me or Always.  Otherwise, it generally has no effect on speed or performance.  Reserving that disk space results in anything written to that space being done on-the-fly as you dictate on an utterance by utterance basis when the speech buffer limit has been exceeded.  Since you're only writing every now and then, the fact that that disk space is treated as a "cache" has little impact.

The only benefit at all in terms of speed or performance is that it eliminates the caution, or warning, about exceeding the space available for speech data and the delay in restoring speech data when switching documents.  However, the latter is only impacted if the user is saving their dictation to a dra file.

In other words, all that setting does is to save the overflow in the speech buffer.  However, since it doesn't save large amounts of speech data when writing to it, the writes to that disk cache are insignificant.  For example, I never save my dictation speech data with my documents but I have that "Disk space reserved for playback" set to 500 MB just in case I need it.  Doing so never ever slows down the performance on my system.  The developers designed it that way so that it would not slow down performance while preserving your dictation audio.

Understand that I'm not saying this is not a valid setting for some users under some conditions.  What I'm saying is that your assumption about increasing speed or performance is incorrect.  I load and run my users from a thumb drive (USB flash drive).  If what you were saying were correct, I should be taking a significantly greater performance hit because reading and writing to flash drives is slower than reading or writing to a hard drive.  My dictation is instantaneous in DNS 9.5 whether I'm working on Windows XP or Windows Vista and never varies regardless of the settings.  Whether I have that setting to 0 or 500 MB the performance of DNS on my systems never varies.  If it did, then I would set the disk space reserved for playback to 0.

Chuck Runquist
Former Dragon NaturallySpeaking SDK & Senior Technical Solutions PM for DNS

It is not so much my confidence that scientists will get it right as it is my confidence that nonscientists will get it wrong.  Isaac Asimov

To which I add my own version, "I am not so confident that technically oriented people will get it right as it is my confidence that non-technically oriented people will get it wrong."