free text-to-song software?

I know there's expensive-ass text-to-song software out there. Are there any that are freeware or open source? I don't have a singing voice so I need the computer to do it for me.

there is a university one called flinger (based on edinburgh unis festival tts) which i have somewhere and should be about if you google for it - trouble is you have to compile it yourself and i couldn't get the bugger to work in windows. it apparently SHOULD compile in windows but i'm fucked if i could even get hello world to compile. i think it demands you use GCC which took me hours to even get half working. it also probably wants weird input in a complicated form, still - better than my regular method of cutting to syllable length and individually tuning dozens of samples done from regular text to speech...

you have to 'register' but it comes up right away, its free for non-commercial.

i think there is a new version which i have not tried - i made my attempts to get it working at least a year ago so some of this may not apply now. there might also be a rough graphical interface now, i believe someone was working on one.

The version of flinger I have uses a midi file with lyrics and a melody line to generate the singing, it's quite good but some words come out weirdly.. For example I tried doing 'cut the midrange drop the bass' in it, and the word 'bass' came out as 'mmbass', hehe.

Eh, Flinger still is not free-free when you have to give them contact information and they put special rules and limitations on what you can and cannot do with it.

I guess there's nothing yet out there. I'll try using a regular speech-synthesis program (plenty of free-free ones of those out there), split the words into separate samples, and pitch-shift them.

shit talker

google it.

  my usernames always really suck said:

Eh, Flinger still is not free-free when you have to give them contact information and they put special rules and limitations on what you can and cannot do with it.

Well, really you could just use a spam-catcher email address and such, like I did.

Also, how will anyone be able to tell that you used Flinger and not another program? Voice synthesis is pretty same-y in general.

  pselodux. said:

  my usernames always really suck said:

Eh, Flinger still is not free-free when you have to give them contact information and they put special rules and limitations on what you can and cannot do with it.

 

Well, really you could just use a spam-catcher email address and such, like I did.

 

Also, how will anyone be able to tell that you used Flinger and not another program? Voice synthesis is pretty same-y in general.

Not really, I can tell the difference between AnalogX, FL Studio, Microsoft Sam, IBM TTS, and Yamaha Vocoloid pretty damn easily.

Just do it yourself by using a normal speech synth and pitching it in a sampler. If you're using synthesised voices you shouldn't care about it sounding like shit.

just record microsoft sam. shitty i know but cheap

  felttipen said:

just record microsoft sam. shitty i know but cheap

is microsoft sam built into windows? i don't think it is but if so how do i access that

i've always used shitty programs before. analog x is wicked though, mostly because you can really fuck with the sound

i downloaded one of these proggies and it made sound come out of one speaker only and fixt when i uninstalled it, wtf

http://www.research.att.com/%7Ettsweb/tts/demo.php

go to that site, u can download whatever u write. i made this

http://www.research.att.com/%7Ettsweb/tts/demo.ph