02-23-2022 07:46 PM
Hello, so I am doing a voice recognition project that involves this premade code that I found on YouTube. This is the link where you can obtain it: https://www.youtube.com/watch?v=aEqa-t_TWiY
In order for the code to function like it's supposed to, there is an additional step and that step is to record your data and add them to the dictionary VI. The instructions are simple but I'm stuck on the first one. I've made attempts but it doesn't seem to be working. The person in the video provided the instructions in their video description.
STEP 1: Run the main VI once (Don’t LOOP over). Record the desired word. Now go to the Dictionary VI.
STEP2: In Match Input Matrix the processed MFCC of the spoken word will appear. Copy the data by rightclicking on it. Data Operations>>Copy data
Sorry if this is kind of a dumb question but what does it mean by running the main VI once? Does it mean to hit the arrow on the top left, speak the desired work, and then stop the VI? I'm only confused about this because the code runs on a while loop so it's ongoing, and it says don't loop over. And yes, I've already attempted it but I'm uncertain if I'm doing it right.
Rest of the steps (Which I should be able to figure out once I'm clear on the first step)
STEP3: If the word said was ‘LEFT’ then paste the contents from Match Input to LEFT1 by rightclicking>>Data Operations>> Paste Dat
STEP4: Now again rightclick>> DataOperation>> Make current value to default to save the MFCC
STEP5: Repeat this process for all the words and store them in appropriate locations.
02-23-2022 09:40 PM
@jackpliskin wrote:
STEP 1: Run the main VI once (Don’t LOOP over). Record the desired word. Now go to the Dictionary VI.
Sorry if this is kind of a dumb question but what does it mean by running the main VI once?
Do you have any LabVIEW experience? Have you given any thought to the Data Acquisition Task involved in doing Voice Recognition?
Just thinking about the "problem" makes me think that Voice Recognition involves analyzing sound. How is "sound" acquired and saved? Maybe you speak into a microphone and record some sort of waveform. How much? Ah, that's the question.
I think the suggestion is that you make a single recording of a single word, with a VI written to record the sound of "one word". How long should the recording be? Just "long enough" -- you clearly need a "Record one word" VI that is flexible enough to record ""and" as well as "algorithm" (which probably takes about 4-5 times as long to record.
If you are serious about pursuing this project, I strongly recommend that on your next Post, you attach some of your LabVIEW code so we can see what you are up to, and can tailor our suggestions to what we perceive is your level of understanding.
Bob Schor
02-24-2022 06:48 AM
@jackpliskin wrote:
Hello, so I am doing a voice recognition project that involves this premade code that I found on YouTube. This is the link where you can obtain it: https://www.youtube.com/watch?v=aEqa-t_TWiY
In order for the code to function like it's supposed to, there is an additional step and that step is to record your data and add them to the dictionary VI. The instructions are simple but I'm stuck on the first one. I've made attempts but it doesn't seem to be working. The person in the video provided the instructions in their video description.
STEP 1: Run the main VI once (Don’t LOOP over). Record the desired word. Now go to the Dictionary VI.
Sorry if this is kind of a dumb question but what does it mean by running the main VI once? Does it mean to hit the arrow on the top left, speak the desired work, and then stop the VI? I'm only confused about this because the code runs on a while loop so it's ongoing, and it says don't loop over.
Press the run button, not the continuous run button.
Basically, what he's doing in the first 9 sec. of the video.