Take the 2-minute tour ×
Game Development Stack Exchange is a question and answer site for professional and independent game developers. It's 100% free, no registration required.

I'm interested in adding to my game either recorded human or synthesized speech. I'm interested in what game devs have to say regarding the value of recorded human voice over voice synthesis. Thanks in advance!


Points to consider:

  1. EDIT: First and foremost, the game is meant for kids (i.e. when you see an image of an apple, a voice says "a is for apple") - here it is on Appbrain (currently sans audio...)
  2. I searched the responses for Is it worthwhile recording your own or buying stock sound effects? - I'm basically wondering the same thing but for voice-specific audio.
  3. Do you find value or any benefit in opting for recorded human voice?
  4. What are some good voice synthesis or text-to-speech tools (preferably free)? I'm currently trying out Fruity Loops, which seems to get the job done (I input a phrase and it generates synthesized speech) however some words do not sound as realistic/convincing as others.
  5. The platform is Android (i.e. memory constraints).
share|improve this question
1  
lol hey gary. didn't know you were active in making games! –  Stephen Furlani Nov 1 '10 at 12:27
    
Hey Stephen! Small world! It's a sort of hobby right now. My first game was released as a beta version three weeks ago (soon to have sound as you can see). I'd say try it out but I've noticed your iphone updates on LinkedIn. ;) Best of luck with your app dev! –  gary Nov 1 '10 at 13:17
    
same to you, although I don't own a smartphone - just borrowing one! –  Stephen Furlani Nov 1 '10 at 14:43

5 Answers 5

up vote 5 down vote accepted

If you're aiming at kids, then I would say human voices are probably the best way to do it.

First off, there's the pronounciation issues. When I was a lot younger (age 8 or so), I had a lot of fun with some free 'Kids Encyclopedia' disk which had synthesized voice that spoke your name. It wasn't great, and I remember spending a while messing with how badly it pronounced stuff.

Second off, there is often a great difference between synthesized and human sounding. If it's for kids as young as 'a is for apple' level, then they might either a) be scared or b)not recognize something sounding so different from what they're used to; i.e. parents/guardians.

Lastly, the price of developing or buying a license for that kind of software is probably much higher than getting a voice actor to do it. Or getting yourself to do it. :)

share|improve this answer

I would rather read than have to listen to a synthesized voice.

Android where space is at a premium, I don't think any speech would worth the size penalty, but if you had to have speech, use a real human's voice.

The only reason to have synthesized voices is generating them live to save space, but on android were you are also limited on processing power, you will not want to generate speech live.

Edit:

For your game you might want to look into the build in voice support, or use one of those libraries, as it seems that your game will not require extreme performance, so using a bit of CPU for voice generation should be fine.

share|improve this answer
    
Good points. OK, the game is meant for kids. I should have added that... will add it now. –  gary Oct 30 '10 at 2:34

For the type of game you are creating, I would say that some form of audio is a must! From a development perspective, I would go with synthetic speech, as the words and sentences are fairly simple, reducing the chances of your audio sounding "robotic". One thing you'll have to pay very close attention to is how the speech system pronounces each word. If it pronounces any word in a sentence incorrectly, you'll have to either "con" the system into pronouncing words the right way, or pick new words. With your target demographic being toddlers, you cannot afford to have mispronounced words in your game.

share|improve this answer
    
I'm also considering pre-loading synth speech files (i.e. created in Fruity Loops, for example). This may be memory intensive but not processor intensive like text to speech. –  gary Oct 30 '10 at 2:55
    
@gary comtois After your edit indicating that the game was for toddlers, I made significant modifications to my answer. It's amazing how having children as a target demographic can impact development. :) –  Ari Patrick Oct 30 '10 at 3:07
1  
Understood! :) And I agree - mispronounced words won't be forgiven! However, I didn't realize "it's a kid's game" was such key info. As I said, some words/letters are less than perfect in certain styles of Fruity Loops' synthetic speech (i.e. synth human vs. synth child vs. robo-voice). I may have to rely on multiple styles in the game... Thanks again for the answer and insight. –  gary Oct 30 '10 at 3:15

Also consider if you want to port it to other languages. Though that will take so much more if you don't speak that language yourself.

Speech synthesizing for non-english are limited as I understand it. And if you don't speak that language yourself you need to hire someone to either validate the synthesized speach or record their speach.

However, if the synthesize engine is just as good in spanish as it is for english, then porting for that language will be very cheap.

share|improve this answer

As long as it is possible to use human voice, use it. Synthetized speech is only good in a few languages, probably only in japanese. If you do not want to port your program to other languages, then you can test the synthetized speech system and probably reduce the effort of developing the program.

The best choice might be to design it retargettable, so that you can switch between both systems dynamically.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.