Voice sources

[raneq]

You may be interested in the CC0-licensed Mozilla's Common Voice.

Each entry in the dataset consists of a unique MP3 and corresponding text file. Many of the 4,257 recorded hours in the dataset also include demographic metadata like age, sex, and accent that can help train the accuracy of speech recognition engines. The dataset currently consists of 3,401 validated hours in 40 languages, but we’re always adding more voices and languages. Take a look at our Languages page to request a language or start contributing.

Forvo too has a huge dataset of pronounced words and sentences, but they are not as eager as Mozilla to share it.

[peppelinux]

simple as possibile, customizable (let's prevent machine learning...)

export ESLANG=it

mkdir $ESLANG
cd $ESLANG

for i in {a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,A,B,C,D,E,F,G,H,I,J,K,L,M,N,O,P,Q,R,S,T,U,V,W,X,Y,Z,0,1,2,3,4,5,6,7,8,9}; do mkdir $i; espeak -s 70 -p 15 -v$ESLANG+f1 $i -w $i/orig_default.wav; ffmpeg -i $i/orig_default.wav -ar 8000 -ac 1 -acodec pcm_u8 $i/default.wav; rm $i/orig_default.wav; done

[peppelinux]

here https://github.com/lepture/captcha/pull/43

[lepture]

Added here:

https://captcha.lepture.com/audio/#voice-library