As of Version 1.20 WebbotLib contains a 'retro' speech synthesiser to speak English text and requires only a single hardware PWM pin. The manual contains more details as to how it is used.
Some other sound synthesisers are just like mp3 players in that you need to record any sounds in advance onto the hardware. But with WebbotLib you can just speak any English text you like.
Because you may need to make, or buy, a suitable amplifier to use this feature then I have created some example sound sample files so that you can hear the output before deciding to make the small investment.
These samples have been created directly from the software, rather than by using a microphone to record the output, and hence represent the best quality achievable (although voice pitch can be changed via the software). You may want to keep playing the clips just in case your PC is adding to any output problems.