PonycardM wrote:So no one thinks the voice lines idea is good?
I didnt say anything immediately because I wanted to give it a fair shot. I took a little lookie-loo into uberduck and renpy again...
bonberjean wrote:No really... that would increase File size and development time. The game already barely reaches the file size limit for free users, if you add voice files in the mix, you won't be able to download the game if you are a free user on the mega site
Mega, filesize- not unsolvable. That aspect didnt make me blink. Mega is one of like 6 decent options- Large filesize is only a problem for the android side of things (and we have several workarounds for that now) and those of you still using dialup. Besides, I'm sure it could be compressed down a large amount.
Besides, what? Wont be able to download the game? Mega allows any unknown IP address to upload or download up to 5 gb of data per day. Which isnt even close to a problem at this point as the 20.1 zip is 1.42 GB on windows, unless I'm missing something here or working on outdated info.
Development time though- well, I'll get to that.
Katria wrote:To say nothing of actually finding voice actors and actresses to record dialogue too.
I'd refer you to the quoted comment below, as finding voice actors wasnt actually suggested. Besides, money talks... so to speak. Puns intended.
PonycardM wrote:So I have been doing some thinking, do you think we could add voice lines though the use of uberduck.ai ?
I mean their are some good sounding voices on their, which could defiantly be used, but I'm asking more about programming the voice lines into the game.
Nothing about the idea of programming voice lines into the game bothers me in the slightest. I don't foresee any real issues.
Uberduck.ai though, leaves much to be desired, in my honest opinion. Voice lines are read largely in monotone, at a strict vocal speed that doesnt ebb and flow like normal speech. Even the more decent sounding voices do this. And with those I still hear altogether too much leftover distortion from the synthesis. I'm talking like little dubstep sounds in the "T"s and "S" sounds. The voices are perfectly comprehendible, but they dont have much
life, at least for me. Not in their current state.
If thats all I had to work with, I'd do some experimentation with some of the vocal clips in Audacity or maybe some ableton plugins, to stretch and pitch bend everything to artificially add life to the vocals, and then I'd EQ it pretty heavily to minimize the dubstep. And you can only pitch bend and stretch so much before you introduce too much distortion.
Personally I see a lot of development time in that. Uberduck doesnt even do this very well with other AI voices, like the google assistant voice, or alexa, for example. Vocals are much flatter (as in, monotone) and are overall distorted. For the amount of time I'd end up putting into it, I do believe I'd have wasted more time than the money would be worth to hire someone commissioning their voice.
Possible? Yes. Difficult? No. Would it be cool as fuck? Absolutely, yes. Is uberduck.ai the way to go? Probably not. Am I going to continue looking into it? Eh, sure, if you want. Point me to some of your favorite voices, maybe things will look better.
Other thing is I mean, I am a bit of a perfectionist, so maybe I'm going over the top, but personally I find the voices as they are off-putting.