Wednesday, July 27, 2005


Ray Schroeder Blog Posting on SpeechExpert
RealPlayer – Free Version
Audio Version of this Post (5.27 MB)

Don’t continue reading this post until you downloaded the audio version. In order to listen to the audio version you need real player. I provided a link for the free version in case you don’t happen to have real player on your computer. It should play fine once you’ve got a player installed.

Yesterday I trashed Microsoft at the end of the post. Today I wanna actually congratulate them for their voice to text software which is available inside Microsoft word. I was inspired to play with this software after reading racial orders Blog posting which I saw on the Sloan see web site about a week ago. It’s interesting how 1ft. rigors and other I didn’t play a little bit with the speechexpert- software that he talked about but I decided that trying while Microsoft had to offer was probably going to be more useful and long run.

The novel thing that I’m doing here with a voice to tax software is that I’ve simultaneously recording a real media stream using the free version of real producer and also transcribing that voice stream two text using the word speech software. To the extent that we can get something of decent quality from the transcriptions point of view one now has something that will be accessible for all and actually I think something that a lot of other but should be trying.

On and off I played with voice recording for at least five or six years and putting streaming voice with either web pages or PowerPoint or something of that, and. The two issues I’ve had in the past with this our first how long would take to produce the contents and second whether in fact what I was doing was really excess of all for all. This stuff is getting much closer to actually doing the trick. As an aside it also might be useful as a way to actually produce first drafts of fundamentally printed material.

I don’t envision everybody doing this and media only but I think it is sufficiently easy that some people should try it and their experience might actually allow others to try it as well.

Let me explain what you’re hearing a part from my voice. Every time word here is a pause from the tape takes what is grayed out of my screen and converts it to printed text. At the same time it gets a little beeped to note that it has accomplished that task. So this is perhaps not as listenable as if I was just doing the voice recorder. You need to be the judge to find out whether it not you can envision listing to stuff like this on a regular basis.

You will also note that I am talking with a date with the punctuation commands. And, actually I didn’t do it correctly so it did the punctuation rather than say the words I need to get along better at that. My point is that if I’m going to be this as a first draft to any further Abell helped me a lot to have the punctuation already there. Consequently I am speaking with pauses rather than continuously. That means it may not be as listenable as it would otherwise the ultimate goal is for me to speak continuously and put the text to actually be created continuously.

We’re still somewhat far from their ultimate goal but where a lot closer than we were a few years ago.

What you see as far as text on this page is exactly as it was converted by word. I have not made any edits at all I have simply copied the word document into the Blog editor. With that copying one can then publish directly to the web. It may not be perfect but I thought you should see what the exact output of the word speech program was.

Yesterday I spent about ten or fifteen minutes training word two of my voice. Today I spent about another fifteen minutes with more training. Presumably the accuracy increases as the amount of training goes up. They also say that you need to be in a quiet place to do the recording. My office is on the street so there are street noise and this a mother noise as well. I think it’s not bad. I’m going to try to use this a little bit more and see if I can get proficient at

No comments: