I have been considering the following:
What bitrate and sample rate should I put all speech files within my game. I have been considering 22khz at around 64/96kbps bitrate. I came to the conclusion that it being only speech it would need to be much higher.
But here is the catch, just wondering if people would prefer having the full 44khz sample rate rather than 22khz?
Most people are unable to tell the difference between 22khz and 44khz.
It's just speech, which unlike a song, does not have that many frequencies in it. I'm sure 22khz or even less should be fine.
Yeah I decided on 22khz, it's what I was originally going to choose. Just thought I'd get other peoples opinions.
I used mono 22khz for the dialog in Limey Lizard. The main issue you'll have to deal with is the quality of the microphones the people who are doing the lines use as well as their living conditions. Boyd's computer is really, really loud, so when I got his lines I had to literally filter and re-filter them until they sounded reasonably clear. You should definitely worry about quality of the recording source more than frequency.
Thanks for the heads-up in contracting Boyd's voice acting services. ;)
Cheers for that ProgZmax..
A note on recording audio:
*Get them to do it somewhere they don't feel embarrassed speaking loudly.
*Make sure they have a guard on their microphone (little felt thing) otherwise you'll have lousy, noisy recordings.
*Have the microphone several inches away from the face if possible.
*If they can do it with a laptop rather than their computer, all the better. This will help with the first point and also reduce ambient noise.
Also, lowering the frequency to intentionally degrade the quality is a neat trick to make a voice sound like it's coming from a radio/communication device.
Haha, no need for those types of tips but thank you anyway. I studied sound engineering at uni, haha.. good to see people so eager to help me out though :)
I'm sorry, but if you studied this then why did you ask about the sample rate etc. when you probably know what the standard is and how much a human ear can recognize and distinguish imperfections, bitrate and voices? And yeah, I got that you were looking for opinions but if you're trained for that then you'd probably know more.
I suggest seeing how audio books are sampled, that's a more or less monotone speech similar to what will be used in most games. The few ones I have are 22kHz at 64 kbps and some in 32kHz at 32kbps. I guess it will be noticable if the characters yell or sing, you should try recording that and see what fits best.
No I was just curious to see what people preferred. Personally for me I can tell the difference between 22khz and 44khz. No matter what the medium is. But that's just me, I was just wondering what other people thought. The other thing is I didn't really delve into mp3 technology when doing sound engineering because its not something you regularly use if at all ever during the recording process.
I whip my voice actors. Back to your cells bitches!!! *ahem*
Well, layabout is right, no one (normal folk) can tell the difference of 44 and 22 frequency. However as Progz said its vitally important to make sure the man who's recording the lines or woman can make a clean recording. It's usually a better choice of a worse but clean voice than a better and noisy.
I would have to agree Dualnames, but so far all the actors I have, have done great jobs with noise supression :) Otherwise I would get out a noise gate and put some wonderful compression on it to get rid of it.
Here's a sample of music at both
44KHz
(http://tindeck.com/image/msgv/stats.png) (http://tindeck.com/listen/msgv)
and
22KHz
(http://tindeck.com/image/mjaa/stats.png) (http://tindeck.com/listen/mjaa)
Music should definitely be at a higher bitrate, while it's not that noticeable with speech. Note it's mainly the hihats that are affected. This would probably apply to any "S" sounds aswell, and it's not necessarily a bad thing!
@Haddas: Well it's not so much the "quality" (in a way that you don't get what music it is) but it just seems to be more supressed/muffeled. That's why mp3 is usually in 44, but still don't think it would affect the voice quality that much if it's 22. You'd notice if you were listening to the voices at the same time at 22 and 44 or if it was singing the lyrics for some rock song but other than that...
If one notices at what Haddas did, they will also notice that the size is the same! (actually the 22 Khz is 1 kb bigger! ;D).
The only reason to downgrade on the quality of voice acting and music is size. If there's no benefit, simply don't do it.
I cannot hear a single difference between them honestly.
Quote from: Layabout on Wed 27/05/2009 08:41:44
I cannot hear a single difference between them honestly.
Maybe time for a new pair of speakers? :D