I ran into issues similar to this, but I forget the details. When I get a chance, I'll fire up my microphone and test.
Select-and-Say is a simple process. If the text window supports any of the standard formats for rich text edit or text framework services (current format supported by Select-and-Say in DNS 9-9.5), then SAPI is able to pass the information back to DNS as to the location of all text visible on the screen. That is, any text that is visible is accessible by DNS through SAPI. Select-and-Say has nothing to do with a stream of data being collected. It also has nothing to do with the last 100 utterances. All of that is related to the Acoustic Model, which has only indirect connections to Select-and-Say. In short, if DNS is able to query SAPI as to the map of text in a text window that is visible (i.e., note that text that has scrolled off the screen is not accessible via this process until you scroll back to that text). As regards some of the sub features of Select-and-Say, select-and-correct or select and re-dictate, don't function at all because the underlying format of the text window doesn't support them. Microsoft Excel exhibits this problem if you attempt to correct text in a cell after the text entry is closed by using the enter key or move keys. In Microsoft Excel it works as long as you are initially dictating either into the cell itself or into the formula bar. So, Microsoft Excel is partially Select-and-Say enabled, but not fully. Another example an Excel is that if you pause during dictation, move to another location, and then move back, DNS does not add a space between the previous dictation and the new utterance. Depending upon the underlying text edit format for a particular text window, some features of Select-and-Say work reasonably well. Others do not. Nevertheless, if the little green light goes off, then Select-and-Say is not supported either in that window were under that condition. Again, Microsoft Excel is a good example. Regardless, the overall functionality of Select-and-Say depends on SAPI being able to access, map, and pass the location of text in the visible portion of the document back to DNS. If and when DNS has access to this information, then it doesn't have to guess. <p align="center" class="MsoNormal" style="margin: 0in 0in 0pt; line-height: 150%; text-align: center">KnowBrainer, Inc. Support Staff – Tom Hamilton A Nuance Gold Certified Endorsed Vendor
Yes for dictation of text. No for selection/correction of dictated text. Thanks, Tom for your valuable input. PainDoc
