Christopher4187
- 22nd January 2014, 14:34
My daughter wants to make a language translator for her science project. She wants to make a Siri-like device but obviously not as complex, no search function and it translates English into different languages. This seems like a daunting task but we starting looking into it and ordered a few things. We have a recording/playback module (ISD-COB17240), which works well, but I'm at a loss as to how to interpret a word. For example, if the user says "hello," I don't know how (or what's the best way) to interpret that audio.
I know part of the fun about a science fair is learning but if the end result is going to be a failed project, I'd prefer she do something different. We're trying to make this project "not so complicated" and at first glance this is what I was thinking:
MAINLOOP:
IF VALID_WORD=1 THEN
LED1=1
GOSUB CHECK_AUD
LED1=0
ENDIF
GOTO MAINLOOP
CHECK_AUD:
SELECT CASE SPOKEN_WORD
CASE 2545
PLAY "HELLO"
CASE 3568
PLAY "GOODBYE"
CASE 5834
PLAY "LEFT"
CASE 12202
PLAY "RIGHT"
END SELECT
RETURN
Obviously this isn't the PBP code but it's where I was thinking of going initially. Perhaps using the count function or a vibration sensor to record "known" words and then use the select case statement to playback the audio. It seems like good translators take into account a number of variables like accents and the speed at which a person says a word but that isn't the scope of our project. With a limited number of "known" words, the translation part should be easier and provide reliable results.
Any advice?
I know part of the fun about a science fair is learning but if the end result is going to be a failed project, I'd prefer she do something different. We're trying to make this project "not so complicated" and at first glance this is what I was thinking:
MAINLOOP:
IF VALID_WORD=1 THEN
LED1=1
GOSUB CHECK_AUD
LED1=0
ENDIF
GOTO MAINLOOP
CHECK_AUD:
SELECT CASE SPOKEN_WORD
CASE 2545
PLAY "HELLO"
CASE 3568
PLAY "GOODBYE"
CASE 5834
PLAY "LEFT"
CASE 12202
PLAY "RIGHT"
END SELECT
RETURN
Obviously this isn't the PBP code but it's where I was thinking of going initially. Perhaps using the count function or a vibration sensor to record "known" words and then use the select case statement to playback the audio. It seems like good translators take into account a number of variables like accents and the speed at which a person says a word but that isn't the scope of our project. With a limited number of "known" words, the translation part should be easier and provide reliable results.
Any advice?