[VB.NET] Compare words in sound files by the vocal wave.

I found a video on youtube that detects the graphic of an audio so I thought of creating a comparison program and not recognition where I have two audios with the same word but with two sizes 0:03 and 0:02 and compare the audios by the result of the picturebox and create a margin of error of 30% and maximum length for each then it would be possible to compare these two different audio files or they would always give me results that are different or always the same and what factors would lead to that The goal is just to compare and say they look but maybe they are not the same audios, the recognition has other metrics so it is a theoretical idea.

Url: https://www.youtube.com/watch?v=USip0-QxX8E

Concept:

Wave1.wav | 2kb
wave2.wav | 3kb

wave1 different of wave2

Wave1 sound wave exemple: (...\.\.\....\.\.\.\\..\..\****...)
Wave2 sound wave exemple: (...\.\.\....\.\.\.\\..\..\****...\..\..\.**)

the above concept helps to understand the simulated compliance are similar but one is bigger and the other smaller but excluding the margin and other possibilities it is possible to know with 70% chances that audios are equal ?

If you mean you want speech recognition: https://www.phon.ucl.ac.uk/courses/s...cognition.html.

Quote:

---

Originally Posted by Peter Swinkels

If you mean you want speech recognition: https://www.phon.ucl.ac.uk/courses/s...cognition.html.

---

I don't want to recognize voice in this common way I want to capture the representativeness example let's say that the bits of the letter Z are like this: 39087789 = Z so I want to see if in wave 2 there is somewhere in the comparison something like the represented value of bits of the audio frequency instead of voice by microphone voice by audio, in a way that I can define who Z is as an example file: wave1.wav = z so if wave2.wav has something like ou = z then it looks like is there something like that?

There are no bits for the spoken letter Z.
Every new recording of the letter by every person will show a different wave pattern.

Why would you want to do that? What kind of program are you writing?

A 64 bit OS:
https://www.vbforums.com/showthread....-with-assembly

Alright, if you seriously want to write a 64-bit OS, why are you even bothering with vb.net? You CANNOT write an OS with that. You're going to need assembly language and/or C. And an emulator to test it on. Also anything related to speech recognition is about the last thing I would worry about. I already asked the question in your other thread just now, but I will ask again just to be sure: why do you want to write an OS?

If you are more interested in studying how an OS works than actually building one, you might want to look at examining the Linux kernel or something similar.

Quote:

---

Originally Posted by LiwisJames

I found a video on youtube that detects the graphic of an audio so I thought of creating a comparison program and not recognition where I have two audios with the same word but with two sizes 0:03 and 0:02 and compare the audios by the result of the picturebox and create a margin of error of 30% and maximum length for each then it would be possible to compare these two different audio files or they would always give me results that are different or always the same and what factors would lead to that The goal is just to compare and say they look but maybe they are not the same audios, the recognition has other metrics so it is a theoretical idea.

Url: https://www.youtube.com/watch?v=USip0-QxX8E

Concept:

Wave1.wav | 2kb
wave2.wav | 3kb

wave1 different of wave2

Wave1 sound wave exemple: (...\.\.\....\.\.\.\\..\..\****...)
Wave2 sound wave exemple: (...\.\.\....\.\.\.\\..\..\****...\..\..\.**)

the above concept helps to understand the simulated compliance are similar but one is bigger and the other smaller but excluding the margin and other possibilities it is possible to know with 70% chances that audios are equal ?

---

Quote:

---

Originally Posted by LiwisJames

I don't want to recognize voice in this common way I want to capture the representativeness example let's say that the bits of the letter Z are like this: 39087789 = Z so I want to see if in wave 2 there is somewhere in the comparison something like the represented value of bits of the audio frequency instead of voice by microphone voice by audio, in a way that I can define who Z is as an example file: wave1.wav = z so if wave2.wav has something like ou = z then it looks like is there something like that?

---

Speech pattern recognition. Such software does exist. It's high end software though and not something that's going to be done with .NET ... at least not easily. It's also going to work on speech patterns and phrases, not individual letters, as that's too small of a sample.


-tg

Quote:

---

Originally Posted by techgnome

Speech pattern recognition. Such software does exist. It's high end software though and not something that's going to be done with .NET ... at least not easily. It's also going to work on speech patterns and phrases, not individual letters, as that's too small of a sample.


-tg

---

but there is the possibility of creating a high-end software in .NET core or something very close.