Post by thread:PIC generated phonemes?

Ok, I'm guessing that the answer to my question is, (a) PIC's don't have enough memory (b) PIC's aren't fast enough or (c) you're a morron, it's (d) Has anyone tried to get a PIC to generated phonemes in order to generate speech? Obviously I know nothing about this. I do know that it's hard to find chips that do this. I just thought that a pic hooked up as a slave with parallel inputs and a output to a LM386 to a speaker could work? You know: 1001 ooh 1010 aah :) Maybe status lines would be needed too, I don't know. Way beyond me. Anyone tried? Nathan Hendler Tucson, AZ USA

A SX (PIC clone) http://techref.massmind.org/scenix hooked up to a big flash memory http://techref.massmind.org/atmel/at45d081 which gives you storage for 1Mbyte of phoneme wav data and running a hand coded SPI routine to readout the data at better than 20kbps and and running a PWM out virtual peripheral to recreate the wav http://techref.massmind.org/io/pwm to a LM386 or other amp Could do it. Devote a few months of Saturdays and become famous. Actually, its already been done with far less. http://www.circuitcellar.com/pastissues/articles/richey110/text.htm There are other methods (LPC, bi and tri phones?) of joining sounds together that produce better speech from less data. http://www.speech.cs.cmu.edu/comp.speech/ --- James Newton spam_OUTjamesnewtonTakeThisOuTgeocities.com 1-619-652-0593 http://techref.massmind.org All the engineering secrets worth knowing {Original Message removed}

Nathan Hendler wrote: {Quote hidden}> > Ok, I'm guessing that the answer to my question is, (a) PIC's don't have > enough memory (b) PIC's aren't fast enough or (c) you're a morron, it's > (d) > > Has anyone tried to get a PIC to generated phonemes in order to generate > speech? Obviously I know nothing about this. I do know that it's hard to > find chips that do this. I just thought that a pic hooked up as a slave > with parallel inputs and a output to a LM386 to a speaker could work? You > know: > > 1001 ooh > 1010 aah > :) > > Maybe status lines would be needed too, I don't know. Way beyond me. > Anyone tried? > > Nathan Hendler > Tucson, AZ USA Nathan, Back in the early 90s I think, there was a text to speech chip called the SPO-AL-256. At least that's what I think it was called. The number I gave might have been the voice recognition chip available at the same time :) Anyways, it did exactly what you ask for. I doubt it is still available, but maybe you can find some information on it that may help? If that isn't the correct number, let me know and I might be able to find the correct one. Josh Koffman .....joshyKILLspam@spam@mb.sympatico.ca

{Quote hidden}> Ok, I'm guessing that the answer to my question is, (a) PIC's don't have > enough memory (b) PIC's aren't fast enough or (c) you're a morron, it's > (d) > > Has anyone tried to get a PIC to generated phonemes in order to generate > speech? Obviously I know nothing about this. I do know that it's hard to > find chips that do this. I just thought that a pic hooked up as a slave > with parallel inputs and a output to a LM386 to a speaker could work? You > know: > > 1001 ooh > 1010 aah > :) > > Maybe status lines would be needed too, I don't know. Way beyond me. > Anyone tried? > > Nathan Hendler > Tucson, AZ USA Nathan, Of course as soon as I sent the last email, a new idea occurred to me. I know of a friend who is using a voice storage chip to store numbers and then recall them at will. I don't know if the chip is fast enough, but if you stored individual phonemes, you can recall them in whatever order you wish. That way you could use the pic to remember the locations of the correct phoneme, and select it by putting a value on another port. I can ask him to get in touch with you if you wish. Hope this helps! Josh Koffman joshyKILLspammb.sympatico.ca

> From: "Josh Koffman" <.....joshyKILLspam.....MB.SYMPATICO.CA> > Nathan Hendler wrote: > > > > Ok, I'm guessing that the answer to my question is, (a) PIC's don't have > > enough memory (b) PIC's aren't fast enough or (c) you're a morron, it's > > (d) > > > > Has anyone tried to get a PIC to generated phonemes in order to generate > > speech? Obviously I know nothing about this. I do know that it's hard to > > find chips that do this. I just thought that a pic hooked up as a slave > > with parallel inputs and a output to a LM386 to a speaker could work? You > > know: > > > > 1001 ooh > > 1010 aah > > :) > > > > Maybe status lines would be needed too, I don't know. Way beyond me. {Quote hidden}> > Anyone tried? > > > > Nathan Hendler > > Tucson, AZ USA > > Nathan, > Back in the early 90s I think, there was a text to speech chip called > the SPO-AL-256. At least that's what I think it was called. The number I > gave might have been the voice recognition chip available at the same > time :) Anyways, it did exactly what you ask for. I doubt it is still > available, but maybe you can find some information on it that may help? > If that isn't the correct number, let me know and I might be able to > find the correct one. I think Oki still do a text-to-voice chip. .