Thank you everyone for your input. ~ and
yes! today I'm going to put the recordings I can access on a cd. I can't access anything but the custom words and phrases I added, all of the standard inventory I uploaded to Model Talker was gone once I'd notified them I was finished recording. That part of the program is what they use to create the synthetic voice. When I receive my voice from them, I will back that file up, probably put it on a flash drive.
Some computer related things I do well with, other concepts I'm clueless about. trfogey, you clarified the wording for me. Barry, yes, I could hear the Etriloquist sample voices, so you're saying that I don't need to download a SAPI program onto my computer (thank god as I tried and it didn't' seem to be there after I unzipped the files) I use a Windows XP operating system.
.... My son is still insisting that if I just continue making the custom recording of my voice, using
any program, that he will be able to get them off of this computer and onto the devices later. He pointed out that creating the synthesized voice is a separate issue than saving my recorded voice.
Barry, that isn't a bad idea about taking voice volunteers
This disease is so horrible when it robs us of something so personal as our speech, and it doesn't just take it from us ourselves, but also from our loved ones. I remember you saying you used to talk on the phone much of the time for a living, and others have made remarks about having been real "talkers". I've never been a chatterbox, but people always liked my voice and I'd get comments on how pleasant it was. Now I think I sound like I am a little old lady, complete with my hair in a bun, and a flask hidden under my granny dress!
About making the inventory recordings for Model Talker:
My speech became more taxed than my voice. The microphone by being a headset style, stayed at a consistent distance from my mouth, and I could speak very softly (which is the only way it was even possible for me to do this.) By softly I mean softly in a room that was damped down with all kinds of sound absorbing material. As this cut down on ambient noise I was able to use the most quiet of voices. While recording, there are meters on screen that show with each record attempt if the voice is too loud or soft, if the pitch was optimal, and if the pronunciation was clear.
Here are samples of phrases (I specifically remember these because they gave me so much trouble! ) You will see that some words are just nonsense.
Dale the didd.
Zith the zirth
For this
Rather they buy
Glass bus doesn't
Jerr the jape
For an evening
I found it interesting that I reacted with the stereotypical response many people have when they're talking to someone who can't understand them. I would talk louder when my pronunciation was off and a particular entry would be rejected. How funny right? Its like someone talking to a person who doesn't speak their language well, but instead of talking slower to them, we tend to talk louder.
When I would get frustrated, I would make myself take a moment to rest, maybe take a sip of water, but drinking makes my voice weaker, so it was counterproductive. Sometimes if I'd already tried and tried to get the computer to take a particular phrase, even after I'd gone on and later came back to it, what I'd do was take a break to upload what I'd already done. I also would remind myself how many of the members here on the forum had expressed regret that they did not have their own voice to use anymore, and that would spur me on!
More about the recording process:
First, you will be required to record a sample inventory of words and sentences, and upload them to Model Talker. MT staff analyzes them, and if they feel you can speak clearly enough to be successful, they then open the full inventory for you. ~ You just go back into that first file you made and submitted to them, you will have named it whatever you like. After MT notifies you that you can start with the actual recordings, you re-open that original file, give it a while to load, maybe even exit and go back in, and then the full inventory of words/phrases for you to say will then be in it.
Each time a person opens the MTR (model talker recorder) re- calibration will be done before the first recording of that session is begun. It isn't like one has to remember to do this, it will begin that stage as soon as the record button is clicked on the first time of any new session.
I had to say "aaaaahhhhh' for a few seconds (quietly) and then it records a few seconds of silence. Next was saying "pa" a few times for the calibration of consonants, then repeat a few phrases. If anything does not come out clear enough, it will prompt you to retry that segment.
After the calibration is complete, and I'd move on to the actual recordings, a frequent error that would pop up was one having to do with the recording failing to register with MT at all. It was not like I was too soft, or not enunciating clearly enough. The message would be that it was having trouble telling when I was starting or finishing what I was to say. I found that if I waited a full "beat" longer than was my instinct before starting speaking when I clicked on the record button, and also paused that extra length of time after I finished, and then released the record button, it almost always fixed that issue. Eventually I'd get into the rhythm, and it would move along nicely.
On a few occasions when this did not seem to be helping, and the program was rejecting almost every attempt, I would exit the model talker recorder (MTR) and go back in, let it recalibrate, and it fixed it.
Once I receive my synthetic voice from them, I will find a way to make a sample accessible to you guys. Not sure how to do that, but its a goal!
