Text-To-Speech on Raspberry Pi3

Hi

Anybody have experience with getting Text-To-Speech running in OpenFrameworks on the Raspberry PI?

I have found https://github.com/kashimAstro/ofxEspeakSynth by @kashim which I guess would that work for Raspberry? But the voice is very synthetic (here is a sample: http://espeak.sourceforge.net/samples/raven.ogg ) , so was looking for more natural sounding options?

Any input much appreciated!
Best
Andreas

1 Like

Hi,

if you install the “espeak-data” package,
(install on debian-based os: apt-get install espeak-date )

you can play with different voices and languages, check out mbrola voice of espeak
are the voices called with mb-* prefix.

usually you are in this path:
/usr/lib/x86_64-linux-gnu/espeak-data/voices/

if you play with the parameters of struct ESParam of ofxEspeakSynth,
you can have different results.

ofxEspeakSynth::ESParam es;
es.language="en_dict";      <------------  check different language 
es.flags=espeakCHARS_8BIT;
es.gender=1;
es.age=25;
es.rate=espeakPITCH;
spk.setup(es);

list language only en:

en_dict
en1_phtrans
en
en-us
mb-af1-en
mb-de4-en
mb-de5-en
mb-en1
mb-fr1-en
mb-fr4-en
mb-gr2-en
mb-hu1-en
mb-nl2-en
mb-pl1-en
mb-ro1-en
mb-sw1-en
mb-sw2-en
en-n
en-rp
en-sc
en-wi
en-wm

alternative advice: festival pack.

good day

Thanks a lot, will check it out :slight_smile:

1 Like