Google Text to Speech provides excellent speech quality. A limitation is that the Pi needs to be online to use it (which mine always is, so it’s not really a drawback).

First, the ALSA sound driver needs to be enabled. It may not be enabled by default, as it’s an experimental driver.

sudo apt-get update
sudo apt-get upgrade
sudo apt-get install alsa-utils
sudo modprobe snd_bcm2835

Also install mplayer, if you haven’t done so already:

sudo apt-get install mplayer

Add the following line to /etc/mplayer/mplayer.conf:

nolirc=yes

Since your speakers are probably connected to the headphones jack socket you may want to force the output to it:

amixer cset numid=3 1

The 1 refers to the headphones, whereas 0 is auto and 2 is HDMI.

A useful script to control the sound volume can be found here: http://www.dronkert.net/rpi/vol.html We’ll set it to max:

vol 100

Next we’ll create a simple shell script to make our call to Google Text to Speech and pass the audio result to mplayer, named ‘say’.

say:

#!/bin/sh
IFS=+
sudo /usr/bin/mplayer -ao alsa -really-quiet -noconsolecontrols "http://translate.google.com/translate_tts?tl=en&q=$*"

Now the Raspberry Pi can say whatever needs to bee said:

say "Lights on"
say "Hello. Camera is running and you are being recorded. I can see you're really upset about this."