Chipmunks Doing Text To Speech

[briesmith]

Why do the text to speech phrases sound like Pinky and Perky or the Chipmunks?

 

Is there some kind of speed control parameter I'm not setting?

 

I've attached the script and log files in case that helps.

GetResults.vgs

1013_1941_vgEngine.zip

[SupportTeam]

Could you please .ZIP up and post the follwing two files from your system:

 

C:\Program Files\VoiceGuide\TTS\Common\ThankYou.wav

C:\Program Files\VoiceGuide\temp\tts_1_1.wav

 

As both sound files are played in same module they both need to be in the same format.

I think that your TTS engine would create the sound file in 8kHz 8bit PCM format. The ThankYou.wav file needs to be in same format.

 

If ThankYou.wav and TTS were in separate modules then they can have different formats.

[briesmith]

Here they are, both of them (Pinky AND Perky).

tts_1_1.wav

ThankYou.wav

[briesmith]

It's also intriguing that Pinky - or is it Perky? - has an English accent (from the Audrey voice) while the other one is definitely an American porker.

[SupportTeam]

ThankYou.wav is recorded in 11kHz format.

TTS is generating files in 8kHz format.

 

Sound files that are used in same module need to be in same format.

 

So the ThankYou.wav needs to be converted/recorded to 8kHz

[briesmith]

OK. Thanks for the info.

[briesmith]

What about the different accents?

[SupportTeam]

Perhaps the pre-recorded .WAV file was generated by a different TTS engine? Or maybe it was even recorded by a human?

[briesmith]

Separating the speech parts into different modules is no big deal but is the TTS engine that generates the text-to-speech on the fly within VG scripts adjustable? Are there are other encoding formats/densities available?

[SupportTeam]

This depends on the TTS engine installed. You would need to consult the documentation of the TTS engines installed on system.

Some of them allow you to specify XML options/tags within the text being spoken.

 

The encoding format is usually not changeable.