Simulate someone's voice

Question

I'm gonna create a game for mobile devices with Unity Engine which is like "Tom Talking Cat" game but with some differences. In face, I'm not gonna just do some effects on user voice. I want to simulate someone's voice. Actually I want to make user voice like a specified voice of a human.
So, would you please help me how to do that ??!!
Thanks !

I tried to make user's voice like another person (in unity with C#), but I couldn't !! — Mahdi, Aug 22 '15 at 10:03
You are going to need to show us some code, what effort you put, what results you got. — Alexandre Borela, Aug 22 '15 at 10:10
Unfortunately I couldn't write any code !!! In fact, I need help for how to write this stuff !! — Mahdi, Aug 22 '15 at 10:28

score 0 · Answer 1 · answered Aug 22 '15 at 12:26

0

You should start by reading documentation, unity have a Microphone class to help record some audio from identified input devices.

It will return an AudioClip that you can manipulate as you wish.

answered Aug 22 '15 at 12:26

Frohlich

953
7
17

You don't need to put effects on audio with this approach, it just capture the voice, I thought you wanted to capture and simulate your user's voice so you need to capture it and analysis frequency and aspect to reproduce. If you don't want to simulate someone specific voice but create an entirely new voice synthesizer maybe you need a little more research to understand how huge is this effort and try to find an open source or proprietary lib for voice sintetizator and focus on your game production instead of create this kind of feature. – Frohlich Aug 23 '15 at 13:10
I thought you didn't understand my question. I have a specified person's voice which I'm gonna make user's voice like his. – Mahdi Aug 24 '15 at 11:05

score 0 · Answer 2 · edited Apr 13 '17 at 12:47

I think the author would like to have access to the solutions of the following problems combined :

Program that listens to person A and splits her speech into content and voice components. The content component is being discarded and the voice component is being stored in a file or database in a suitable format.
Inputs person's B speech the same way except retains both components but substitute the voice component with the one stored in the file from person A. Finally combing both components into a speech that has the content from one person (live stream in this case) with the voice stored in a file or a database.

I think what you want to do is to make a person talk with someone else's voice. Each of these problems is a scientific endeavor and only recently computer science started approaching voice recognition / generation solutions that are more or less usable.

I don't think you would be able to find a single library that can do this but here is a good read of the current state of the matter

https://dsp.stackexchange.com/questions/7833/how-to-mimic-copy-fake-someones-voice

Simulate someone's voice

2 Answers2