I need to make a voice changer...


FaeMinx

Rainbow Liberation Instigation
Joined
Dec 11, 2010
Messages
3,143
Age
43
Location
outside looking in side looking out
I know you can buy portable ones, but I need something as discrete and lightweight as possible... So I think I will have to make one from scratch.

What I am wanting to do is fit it into the snout of a mask I am making... Have a little mic positioned near my mouth, and have have it output to some speakers that I will build into some fake ears or something like that.

I would like to be able to program how the voice is changed with python (or whatever), so I need to be able to plug it into a computer, or it needs to be able to receive a microSD, or something like that.

Obviously as low power as possible, I would like to power the speakers separately, wired to a battery pack on my chest or elsewhere. So I would need small batteries in the snout.... Or alternatives?


Any and all ideas welcome... From general ideas, to python code, to specific circuitry and components required...

I'm all ears.

Included are a few pictures of the mask.
 

Attachments

  • IMG_20180220_141709650.jpg
    IMG_20180220_141709650.jpg
    1.2 MB · Views: 274
  • IMG_20180224_154237045.jpg
    IMG_20180224_154237045.jpg
    1.5 MB · Views: 275
  • IMG_20180224_052952569.jpg
    IMG_20180224_052952569.jpg
    1.2 MB · Views: 262
  • IMG_20180224_053000457.jpg
    IMG_20180224_053000457.jpg
    1.2 MB · Views: 248
  • IMG_20180224_053234647.jpg
    IMG_20180224_053234647.jpg
    969.5 KB · Views: 414
How are planning to suppress your orginal voice? Having that plus the altered one played a micro second later sounds just weird.

I would not build anything for that, have a mini computer in your pocket (like maybe a pandora!) and just lay the microphone cable up to your mouth.
If the mask is however all that you plan to wear get a pi zero.
A plain sound altering script that just pipes the microphone input through a sound lib may generate a lot of noises you dont want.
I wonder if a speech recognising software and fully generated speak outs could work around that.

Oh yea human voice can be surprisingly loud at times. So you need a powerfull amp too. Ive ordered a few mini amps for a project a while ago, some are as small as fingernails, they have 3W and it turns out this is not enough for human voices.
 
How are planning to suppress your orginal voice? Having that plus the altered one played a micro second later sounds just weird.

I would not build anything for that, have a mini computer in your pocket (like maybe a pandora!) and just lay the microphone cable up to your mouth.
If the mask is however all that you plan to wear get a pi zero.
A plain sound altering script that just pipes the microphone input through a sound lib may generate a lot of noises you dont want.
I wonder if a speech recognising software and fully generated speak outs could work around that.

Oh yea human voice can be surprisingly loud at times. So you need a powerfull amp too. Ive ordered a few mini amps for a project a while ago, some are as small as fingernails, they have 3W and it turns out this is not enough for human voices.
Not intending to suppress my voice... Rather I want to harmonize with it, so I will need to play around and find something that sounds good. Add reverb or whatever....

First prize would be to have a number of presets I can switch on the fly to have different voices for different characters.

I will have other parts of the suit where I can integrate devices, but obviously... I want this all to be as light as possible.

I would like to build it into the snout if possible.

Surely there are solutions for noise cancellation... Adding a second microphone, etc.

I am not looking for the easiest solution, but the *most complete* solution.

I'm also jumping in the deep end here, as I don't have the first clue what is required where.

Surely I only need an amp to drive the speakers?

Sent from my Moto G (4) using Tapatalk
 
Perhaps you should not integrate the full electronics into the snout but just the microphone and speakers and connect them to your phone.

So, ...ehhh...who are you kidnapping for ransom then?
 
Perhaps you should not integrate the full electronics into the snout but just the microphone and speakers and connect them to your phone.

So, ...ehhh...who are you kidnapping for ransom then?
Hmm... That's a possibility... I will look into it.

As to who I am kidnapping...
There has been this... *boy*.... Walking around trying to steal my life with his human ideas... Well, I have put up with it for far too long!

I am bursting the illusion(bubble) of his existence, and revealing the truth of MiceElf! Muhahahaha.....

I am replacing what made him human, with my own little creations... One... piece... at... a... time.


P.s. MiceElf are not bubbles. Do not try to pop them, else ye shall pop yerself!

.... and realize you have been MiceElf all along, so all for the Squeek eh?

Sent from my Moto G (4) using Tapatalk
 
Nice... http://fakevoice.com/ has been hacked with a cool effect... pitty. (note: could have side effects, as it also want to start macromedia flash. It is not needed for displaying the cool mouse effect)

Ok, Alternative 1 (will set you back 10 dollars):
Just search amazon or ali baba "toy voice changer" lots of cheap alternatives.

Alternative 2:

The other way is just a vocoder plug-in into a portable computer (aka laptop in your backpack) that does the realtime translation... keeping the laptop cool is a challenge...

note: I have no idea about vocoders... but maybe this one does something (preferably in realtime, not sure. Linux)

http://www.sirlab.de/linux/download_vocoder.html
 
Last edited:
why reinvent the wheel? there are plenty of toy voice changers that will do . my kids had one. when removed from the plastic megaphone type casing the board and switches is very small. the one my kids had did octave up, octave down robot, and a few other effects like reverb. if i remember it had 4 switches which different combos gave different effects. and all for less than a raspPi.

on secon thoughts most of thenm are very distorted so you would have to use a better amp and speaker than the toy comes with..

there are actually phone apps for voice changing too...
 
Why not just breathe some helium or smoke a lot of cigarettes ? It's the easiest way to change your voice !! No battery required !! xD
 
Why not just breathe some helium or smoke a lot of cigarettes ? It's the easiest way to change your voice !! No battery required !! xD
By that logic, whacking yourself in the crotch really hard also does the trick and costs nothing whatsoever. ;)
 
By that logic, whacking yourself in the crotch really hard also does the trick and costs nothing whatsoever. ;)

Exactly xD but choosing the treatment depends on the wanted result :cool:

I have seen a lot of smartphone apps to play with the voice, maybe is cheaper to get a low range smartphone and put an app in it
 
You all dont deserve to be in this forum.
The next one who mentions smartphone or premade solutions gets a free crotch adjustment.
 
Look here guys... I certainly will concider all these... Mickey Mouse solutions...

But I'm looking to *make* something decent... To me, having to use a smartphone is just...

Seriously, you expect me to produce top quality fursuits in the future, with all the bells and whistles and then tell the client...

"Sorry, will not function unless you have a smartphone... Oh, and uh... You need to download this app... What? After a forced update the app doesn't work properly! God damnit!"

Seriously guys, I'm looking for a pro solution... I want something that sounds amazing... Not some cheap toy with godaweful distortion, or having to rely on a laptop in a backpack orca phone in a pocket.

I want the person to be able to charge up the suit, put it on, and be good to go untill the batteries run out.

I want to be able to manually adjust how the voice is changed... I don't just want 3 or 4 presets... I want to be able to make infinite presets of my own.

Look guys, I came to you because honestly if it means designing a circuit board, populating it, and writing the software from scratch... That's what I will do! (With help of course).

I'm sorry I haven't investigated the pi zero or Arduino solutions yet, I'm very busy and had only 1 evening at home before having to return to my workshop (where there is no internet)... But I will, this weekend.

The reason I am suspicious of those solutions, is because they are designed for a multitude of applications, and I am sure that a solution designed for voice modulation *only* would be far more compact, maybe even more energy efficient.

So I am not discounting those solutions... If they end up being small and light enough.... It will make my life a lot easier.

But, I guess my question really is, and I hope some of you have the know how to answer me...

If I were to make the best damn voice changer from scratch, how would I go about it?

What components do I need?
How do I put them together?
What would the circuit look like?
What software would I need?

I say it must be Python, because then I know I will *eventually* be able to get my head around writing the software... Any other language could be beyond me.

Think of this as a 'new potential product' being discussed...

Sent from my Moto G (4) using Tapatalk
 
What would the circuit look like?
Well this line puzzles me a bit.
Are you asking how to build an actual voice changing circuit completely from scratch? You can do that course, but if you want this programmable via software you are going far far away from your original project.
If you want to have anything that runs something like python scripts on it, accepts sd cards or store hundrets of presets you cant dodge using a full fledged mini computer as a core.
 
Just start with a quick and dirty prototype. That is why I suggested using a phone (or Pandora, Laptop, etc.) as a first step. Your python script won't care all that much about the underlying platform.
 
forget about rolling your own electronic circuit ... doing that will require some serious time investemnt to even get near anything "nice" sounding with tons of stuff to watch out for, also if you want it to be tuneable and flexible you will end up with a very big circuit ... unless you have a profound interest and time at your hands stay clear of analog electronics xD

there are some dedicated voice changer ICs on the market but they will need support ciruitry, amplifiers etc. again doable but takes tons of time and not as flexible

doing sampling and changing voice in realtime on a microcontroller needs some heavy duty algorithms (and Math) like FFT : https://mathblog.com/creating-cartoon-voices-with-math/

so you would probably end up needing a DSP with some audio circuitry ... doable but complex to learn and definitely not with Python

what you want pretty much lands you on a Raspberry, BeagleBone or similar running some Linux, having a sound system and Python available ... RaspBerry would need a cheap USB soundcard, not sure about the others
 
Arduino may not be bulky, if the result device is carefully planned.
First of all, prototype on Arduino module, let's take Uno because it's cheap and easy. Get the module with socket, not chip soldered in. After making prototype, get rid of the module from your project. I do it all time, I made lots of projects using single Arduino module, I'm only changing chips.
Most things in Arduino module are things you don't need in a final device. Computer interface? Can be external on 3 pins. Power supply? It's hard to power any useful stuff from the linear chip on the board. 3V3 regulator? err... it's hard to power an SD card from it.

So what I usually do? First, I build prototype on Arduino and breadboard. After having something nice, I design the power supply for my needs. You may need some battery, so consider locating the device away from the "mouthpiece" and using some cable. After re-designing power supply, I design the device and the whole Arduino gets minimized to:
- The microcontroller
- 4 capacitors
- 1 resistor
- 1 16MHz crystal.
This is much smaller than module.
However, there is one problem with Arduino: The awfully small computational power as for sound processing. Arduino is not a DSP and has small RAM. You can theoretically sample sound into it, then modify the samples some way and then put them using something like Covox (if you don't want PWM, it sounds like good ZX Spectrum), but this will cost a power and may even introduce delays, I think that even with running at overclocked 24MHz you may have problems with realtime processing.
The problem I always had with Arduino output using PWM+filters was the "metal", "noisy" characteristics of sound coming from it. It is just not possible to have a smooth wave. I finally made the situation a bit better by using Covox-like resistor ladder to pretend the sine signal.

There may be other solutions, like these Intel Linux small computers, DSP chips (I haven't tried them) or even using analog circuit, but such circuit needs to be very well designed and, better, simulated.
 
These are all very good responses, and I was getting excited about the Arduino untill I realized that for the effects I'm looking for, I will need a fairly beefy processor...

It looks like the next step for me is to do some research when I get home this weekend.

In the meantime, anyone know of something like Arduino but far more powerful?

Alternatively, and I like this idea... As I want to use Python...
What is the smallest computer you know of that could handle this effectively... If I have to build something into the suit, perhaps it can perform other functions as well... Like handle the animatronics for the ears and tail... And lighting effects...



Sent from my Moto G (4) using Tapatalk
 
had a google around and I think this one is hard to beat in size, power and staying low cost : https://www.raspberrypi-spy.co.uk/2017/03/amplified-voice-changer-using-a-raspberry-pi-zero/

- RaspberryPi Zero W (WLan)
- USB soundcard
- speaker(with amplifier or use an external amplifier) + microphone
- USB battery pack
- cables + adapter

as for anything more powerful than an Arduino ... you could look into some of the ARM based Microcontrollers from NXP or Microchip which have internal flash and run with minimal external components
 
Back
Top