Do you have a lot to type, but so little time to type it? Do you drive long distances to work and back and want to take advantage of the extra time? Do you enjoy the long break to work where you can gather your thoughts, but aren’t able to get them onto paper?

Speech to Text v2.35 allows you to speak while it types what you say either in its own text editor or in MS Word. Your own voice commands turn different functions on and off. Immediately upon opening the Speech to Text program the Speech Recognition Engine appears where you create individual profiles and embark on a training program that will fine tune the engine to recognize your individual speech patterns.

Before downloading Speech to Text be sure to download and install the Microsoft Speech Recognition engine. First, it will lead you to configure your microphone through a series of interactive screens. Next you will be asked to create an unique profile. This profile goes through at least 1 speech recognition training. You read slowly and clearly what is on each successive screen. As the words are highlighted by the engine, you can continue reading the specific training text. It is advised that you go through at least 3 training sessions with the Speech Recognition engine. Remember to speak slowly and enunciate each and every word. You will find the engine keeping up with you this way. If you speed up, you will find the engine lagging way behind and you will have to repeat entire sentences.

Eventually, the Speak-to Text software will appear without the Speech Recognition engine popping up first. No problem, simply go to your Control Panel and double click on the Speech icon. Click on Train Profile and select one of the 8 sessions listed for training. Or, easier yet, click on the 3rd from last icon in the Icon menu.

To turn on the microphone go to Tools menu and select Microphone on/off. Click and you will see a check in front of the menu selection. Another way is to depress the slanted lollipop icon in the middle of the Icon menu.

Tools menu and select on Train word is a way to establish the pronunciation of individual words. Don't expect to use this extensively as a way to train the Speech Recognition engine. Sentence structure is where it is at.

Tools menu and select Setting bring a person to four tabs, General, Greeting, Formatting, and User Commands. Formatting allows you to force all numbers from 0 to 20 as numbers instead of text . You can also do Background noise filtering whereby you Enable filtering of background noise for dictation and/or Enable filtering of background noise for commands.

Additional features:
Save in Word
Open PDF and create PDF
Clear Speech Buffer
Listening for okay, cancel, apply, and tab
Create custom commands

It was enjoyable to have speech recognition by this software at least 60% of the time after the first three Speech Recognition engine training sessions during the very first use of this software. However, the real frustrations came when the engine did not recognize individual profiles that had already been created causing a person to redo the training sessions. It also seemed as if the more training sessions were attempted, the less recognition was apparent. Wake up became leg up and Okay was heard as FDA. Perhaps this inability to remember profiles and the training sessions already completed is limited to the SpeakToText trial version.

Here are some snapshots of SpeakToText:






To get your free trial go to:

http://www.coolsoftllc.com/main.asp

I didn't try this program myself, but I'm not sure how usable it is. Text to speech is one thing, but speech to text is a completely different one. Don't we all remember (I hope we do) what happened to Vista Speech Recognition System? If not, take a look at this video. It will show you perfectly well how even a giant like Microsoft can't deal with it properly

Yet again, I might be wrong, and this program might be a future must-have.

This program is not freeware, so it should not be posted in this section...moved accordingly.

These programs as with others will require some time to train it properly. I never heard of this program before, but it should be better than the default Microsoft speech recognition software. For those interested, you might also want to take a look at Dragon Naturally Speaking:

http://www.nuance.com/naturallyspeaking/

Prices will be higher for this speech to text program.

@pyost: Just saw that YouTube video you posted on the Microsoft screw up....ROFLMAO.

There were better Speech recognition softwares than what Microsoft had to offer. At least when I was very much into that technology. If I recall correctly, there was this Dragon Speech Recognition software which was very famous in our area. After going through lots of sample reading, the recognition accuracy reached commendable heights. It was then that I realized the problems the process has to face.

The amount of noise exponentiates the performance issues. Although rarely a factor, because of work areas being relatively calmer. Another factor is the quality and type of the microphone. A poor quality microphone has the same effect as noise. Headset Microphones are the best since they maintain the distance between the mouth and the Mic which is harder to accomplish on the desktop ones. These problem can be taken care of, but the others pertaining to our own voice are difficult to. Often, the software would recognize speech very well. Especially after it has just finished analyzing your voice based on the sample reading. Come back an hour later and it would all go bonkers. I guess it was due to the fact that the pitch, tone & style of my voice changed during the interval.

I am not sure if this happens to others, but this for me was the greatest drawback in such systems. As more and more research is done, we will head to the age when keyboards will become an extinct technology.



No No , You posted that thing before me, while I was denied by "Some Files are Missing Error".

This post has been edited by turbopowerdmaxsteel: Apr 6 2007, 09:41 PM

@turbopowerdmaxsteel: That's ok. I agree with you on the quality of the microphone. You don't need to get a top of the line microphone, but a decent one will not cost too much. Try going for a Sennheiser or maybe even a Logitech model. I would read some reviews (Google for them) on the model you wish to buy and see what other consumers have to say about them before buying one.

Well, this speech to text thing is really old now. Nothing new. That's what i know for the fact that i remember trying to use one around 4 or 5 years back. Failed hopelessly though.

These things are far from perfect. seriously. Typing is much faster, and quicker, and would/will have less errors!