Yren Posted June 20, 2008 Report Share Posted June 20, 2008 guys, i m doing speech recognition using labview for my final year project. Can anyone guide me on this ? thanks a lot and appreciate it much. :worship: Quote Link to comment
LAVA 1.0 Content Posted June 20, 2008 Report Share Posted June 20, 2008 QUOTE (Yren @ Jun 19 2008, 04:20 PM) guys, i m doing speech recognition using labview for my final year project. Can anyone guide me on this ? thanks a lot and appreciate it much. :worship: Hello, Well... First question : have you ever used LabVIEW on other project or are you completely new ? Speech recognition... Ok, that's vast... what exactly do you want to achieve ? Could you tell more precisely your goals, I guess you don't need a LabVIEW program that writes what you speak-up On what part do you need help ? Maybe on acquiring sound ? Maybe on processing the acquire signal ? Maybe something else... Do you know what hardware/software you'll be able to use ? Do you need advice to choose them ? A well set problem is half solved ! Help us to help you. Quote Link to comment
Thang Nguyen Posted June 20, 2008 Report Share Posted June 20, 2008 QUOTE (Yren @ Jun 19 2008, 09:20 AM) guys, i m doing speech recognition using labview for my final year project. Can anyone guide me on this ? thanks a lot and appreciate it much. :worship: You should at least start doing something before asking the question. I also feel interesting about your topic, but you should provide more information like Antonie questions. Quote Link to comment
Yren Posted June 20, 2008 Author Report Share Posted June 20, 2008 thanks for the replies, I m quite new to Labview currently but i m having my labview tutorial class so i m still a learner. Ok, here is what i plan to do in my project. i would like to control a simple dc motor in terms of speed and direction with speech recognition using labview. For example, when a input human voice = "faster", the dc motor will drive faster. I think i can input the voice from my integrated mic of my laptop by using the "acquire sound" (programming > graphics and sounds > sound > input > acquire sound) and display it in a waveform graph. But, how m i able to save the pattern of certain voice so that i can compare it with my new input voice ? or this is the wrong way to do it ? any other suggestion on the signal processing? Currently i m having the NI DAQ usb-6008. Anyone can give me a guide on this ? Thanks a lot guys. Quote Link to comment
Francois Normandin Posted June 20, 2008 Report Share Posted June 20, 2008 Hi Yren, Speech recognition is complex. Good luck! I'm not sure how well you can have LabVIEW decipher words from spectra acquired this way. If I were to attempt this, I'd probably try using a software that already does it, put the words it detects in a buffer and then read the buffer with LabVIEW. Then you can have your program trigger on specific words and actuate your motor... I might be completely wrong about this, but that must take real good signal analysis. I suspect most speech recognition softwares utilize existing libraries of subroutines. Maybe some of these are freely available and documented? You could definitely access those from within LabVIEW using ActiveX or dlls. Quote Link to comment
Phillip Brooks Posted June 20, 2008 Report Share Posted June 20, 2008 There was a presentation at last year's NI Developer Day that demonstrated using the voice recognition capabilities of Windows Vista with LabVIEW. http://www.ni.com/swf/flv/labview/us/vista/vr/ Quote Link to comment
Yren Posted June 21, 2008 Author Report Share Posted June 21, 2008 saw that video before.. is it possible to get the VI being used in the video ? Quote Link to comment
Rolf Kalbermatter Posted July 1, 2008 Report Share Posted July 1, 2008 QUOTE (Yren @ Jun 20 2008, 02:22 AM) saw that video before.. is it possible to get the VI being used in the video ? Its most likely just wrapping some .Net functionality built into Vista. Don't know more about it though. Rolf Kalbermatter Quote Link to comment
Norm Kirchner Posted July 1, 2008 Report Share Posted July 1, 2008 It most definitely is using some .NET code. But this would not be speech recognition using LabVIEW, as you mentioned is necessary for this class. Not really at least. If you are using LV and speech recognition is just part of what needs to be done in a bigger project, then using the built in windows speech to text ability is probably fair game. You are in luck though, because although the demo showcases vista you can do this in XP also. Quote Link to comment
LAVA 1.0 Content Posted August 1, 2008 Report Share Posted August 1, 2008 QUOTE (Yren @ Jun 19 2008, 11:31 AM) thanks for the replies, I m quite new to Labview currently but i m having my labview tutorial class so i m still a learner. Ok, here is what i plan to do in my project. i would like to control a simple dc motor in terms of speed and direction with speech recognition using labview. For example, when a input human voice = "faster", the dc motor will drive faster. I think i can input the voice from my integrated mic of my laptop by using the "acquire sound" (programming > graphics and sounds > sound > input > acquire sound) and display it in a waveform graph. But, how m i able to save the pattern of certain voice so that i can compare it with my new input voice ? or this is the wrong way to do it ? any other suggestion on the signal processing? Currently i m having the NI DAQ usb-6008. Anyone can give me a guide on this ? Thanks a lot guys. Here are two strategies you could take for your course (this thread is a month old, are you already done with class?). #1 -- Choose two words (one for faster, one for slower) that are uniquely different. You don't want to write a "speech recognition" package that understands many words, just the difference between two. I would recommend writing a labview program that can measure,save and display your microphone output, this way you can experiment and find two unique 'signatures'. There are many ways to process your raw microphone analog input once you get it on your computer,.. getting it into a frequency domain would be a start, that way you can statistically measure the difference between two chunks of analog audio data (words, phrase).. Does this make sense? #2 -- The easy way, measure and record the audio in chunks... say you hit a button and labview records your voice for 1 second, you could make a noise with increasing pitch(frequency) over that one second which would denote "faster" motor speed, and vice-versa. This would be simpler (I believe) as you would only need to track the derivative of the frequency over time. In other words, taking the 1 second of audio and chopping it up into 10 pieces (100 millisecond) chunks, and grabbing the frequency for each increment, and tracking the direction (increasing/decreasing).. Hope this helps! -Scott Quote Link to comment
mohankumarb Posted October 12, 2008 Report Share Posted October 12, 2008 i m also using labview 8.6 for speech processing ....... i need some example vi's to know about adaptive filters in addition to the examples provided in the labview8.6 itself..... i m also using labview 8.6 for speech processing ....... i need some example vi's to know about adaptive filters in addition to the examples provided in the labview8.6 itself..... Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.