![]() | ![]() | ![]() |
| | |||||||
| Robotics Chat Specific to discussions about robots and the making of. |
| | LinkBack | Thread Tools | Display Modes |
| | (permalink) |
| first of all, izzit possible? i am gonna do this robot for my school project. can anyone help me...this is my first robotic project i really wanna hear my robot talk! plz...help me.... my progress is still 0% any info would be GOOD!! | |
| |
| | (permalink) | |
| Quote:
| ||
| |
| | (permalink) |
| Someone here once recommended Winbond's text-to-speech IC for a similar project. It has an embedded text->phoenem converter and it can be programmed through an SPI interface which also accepts ASCII input and system commands (datasheets). If you're in the U.S., they usually go for about $36. | |
| |
| | (permalink) |
| First off, nothing is impossible. Time and buget are the only limitations to achieving anything. I am assuming that, because it is a school project, time and buget are both limited. I see this project as having three steps: Convert speech-to-text Analyse what was said and an appropriate response Convert text-to-speech I think a computer is the only way to build this project. The second step is the hardest and if you get this the other two should be easy. Just think what goes through your mind when someone asks you a question? You would have to program the robot to think like this. | |
| |
| | (permalink) |
| You know, this thing may very well sound like it CAN NOT be done. But... if you think about it, people do write dictionaries. So if you could find a way to put the entire english dictionary in some program, then figure out a way to actually comprehend all of those words, I think you could really get somewhere with it. After all, the text-to-speach and speach to text has already been done. Anyway, this would be the kind of thing that some genius that's been working on this for decades would come up with. You'd see it all over the news too. It would be extremely, extremely time-consuming to do. It wouldn't even be extremely hard. But, I don't think it would be best to aim for something like this. Maybe you can go for the same type of goal but using commands instead. So it would be a robot that would understand commands or select phrases. Would you are looking for, is an android. Do you have any smaller goals you might want to aim for instead? Maybe we can help you through some of those. :wink: Rain
__________________ When life gives you lemons... make a battery. | |
| |
| | (permalink) |
| I say it all depends on the level of interaction you're looking for. Any level of communication beyond key phrases and commands would be an exceptionally daunting tast, so shooting for some HAL 9000-type of sophistication probably wouldn't make the best school project. A basic text-to-speech-converting bot wouldn't take nearly as much resources as long as you limit it's vocubulary to something reasonable. Anyway, a similar project was used using a SPO256A device here not too long ago. Also, Tigerbotics has a synthesiser called SpeakJet though the MarkIII online store. | |
| |
| | (permalink) |
| hey guys...thanx a lot for d little bit of here and there.. my heart was pumping real fast when i was readin all of ur replies..lol hoping dat there's really hope for me to continue with my project | |
| |
| | (permalink) |
| hey guys...thanx a lot for d little bit of here and there.. my heart was pumping real fast when i was readin all of ur replies..lol hoping dat there's really hope for me to continue with my project really kinda worried after reading all ur replies..."to swing or not to swing" however, i think dat its easier if i use keyboard as the input as im afraid that there's some words that sounds the same.. and it will be quite tough for my robot to differentiate the words...rite ha? i think i need AI aftificial intelligent software.. can i write the program myself ha? i have my flow chart of the robot...any suggestions to add or correct me if im wrong... 1. input(microphone/keyboard?) 2. adc 3. microcontroller (speech-to-text) 4. lcd display 5. microcontroller (AI) 6. lcd display 7. microcontroller (text-to-speech) 8. dac 9. output (speaker) actually, i got this idea for this project from ALICE AI website. but i wanna hear her talk or shall i say read the text with robot kinda voice... that's all... k, do reply me back. any reply is very much apprreciated bubbye | |
| |
| | (permalink) |
| do a google on text to speech or this one is cool 8) http://www.research.att.com/projects/tts/demo.html this one has text to speech chips.. http://www.winbond.com/e-winbondhtm/partner/b_2_a.htm good luck | |
| |
| | (permalink) | |
| Quote:
__________________ I'm no electronics god, i just talk too much. | ||
| |
| | (permalink) | ||
| Quote:
To be truthful, with the very little I've thought about it, me too.
__________________ When life gives you lemons... make a battery. | |||
| |
| | (permalink) |
| well cindy, it seems u r pretty young for this kind of a project. but well, if u use readymade kits for particular operations then, it maybe a bit easier. u can make the task easier by asking the question by typing and the robot replies. u can feed few particular questions..... n predefined answers as sound files..... whenever one asks one of those questions .... it answers... a more intelligent approach would be..... using a software to match any random question from the keywords..... and there could be a buffer of few thousand answers in that particular topic..... still it is not very good. the robot may answer i don t know, sorry.... in case of a lot of answers.... other approaches, like understanding what u ask, n then answering it.... is a big project and will consume a lot of time, effort and money..... as well as expertise.... Subhasis
__________________ Irreverence and flippancy is essential to mantain sanity. | |
| |
| | (permalink) |
| hey guys, have you ever chatted with oliver at www.oliverbot.com ? it is a program that can carry an actual conversation with a human being. try it, it is kinda fun. I am not sure how this works, but to me it looks like it is a big database where certain input words will cause it to display different responses, and it works pretty well. if there is a way that the input could be speech (or speech to text) and the output is speech too (text to speech), then you've got a pretty social and interactive robot ! any ideas? | |
| |
| | (permalink) | |
| Quote:
| ||
| |
| | (permalink) |
| There's a company called sensory that makes speech recognition chips. They claim they can get really high recognition rates for a small vocabulary. | |
| |