i'm amateur software developer, and i'm writing a program that will "read" given text. i record phonemes, one-by-one, but because i don't record them in studio (actually i'm recording them at home, with sh*tty microphone), every .wav file has different amplitude, frequency, etc, and after gluing, it sounds terrible. what can i do?
Comments
Since you know what is wrong with your recordings, go and get so
Since you know what is wrong with your recordings, go and get some gear that works for the job.
A decent mic for this job for about 60 $, a frontend like this here [="http://www.thomann.de/gb/edirol_ua4fx.htm"]EDIROL UA-4FX - Thomann UK Cyberstore[/]="http://www.thomann…"]EDIROL UA-4FX - Thomann UK Cyberstore[/]
It is no professional equipment, but it has some usable features and should be easy to handle.
And as a DAW maybe [[url=http://="http://www.thomann…"]STEINBERG CUBASE ESSENTIAL 5 - Thomann UK Cyberstore[/]="http://www.thomann…"]STEINBERG CUBASE ESSENTIAL 5 - Thomann UK Cyberstore[/]
With those tools you can cut, level and equalize all samples with great results. A solution that cost nothing, but sounds good, done with a no
budget consumer gear is only just maybe possible if you are a trained Audio guy. And he would be quite unhappy...
For the successs of your project, try to get some better gear and check the possibilties of it. Your recording task becomes a breeze...
If you must do it this way, try recording all the phonemes one a
If you must do it this way, try recording all the phonemes one after the other with short gaps in between in the same session. Use your DAW to chop the resulting file into multiple files each containing one phoneme. It will still be terrible, but there are different degrees of terribleness.
There are many commercial and free text-to-speech programs available. I would spend a little time looking to see if there is one that would be suitable for your purposes.