Topics

[espeak-ng:master] new issue: Setting buffer size to 50ms leads to truncated mbrola-fr4 output #github


espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng:master] New Issue Created by sthibaul:
#772 Setting buffer size to 50ms leads to truncated mbrola-fr4 output

Hello,

With the default parameters,

properly speaks the provided word. But with a buffer size set to 50ms, the sound is completely cut. That can be reproduced by doing the following change, to make `espeak-ng` explicitly request 50ms buffer:

diff --git a/src/espeak-ng.c b/src/espeak-ng.c index 1634cc1d..5ce3f55c 100644 --- a/src/espeak-ng.c +++ b/src/espeak-ng.c @@ -600,7 +600,7 @@ int main(int argc, char **argv) } } else { // play the sound output - result = espeak_ng_InitializeOutput(PLAYBACK_MODE, 0, devicename[0] ? devicename : NULL); + result = espeak_ng_InitializeOutput(PLAYBACK_MODE, 50, devicename[0] ? devicename : NULL); samplerate = espeak_ng_GetSampleRate(); }

```

espeak-ng -v mb-fr4 Bonjour

will almost not pronounce anything. Passing a longer sentence gets more output, but still really not much, it seems there is a bogus computation somewhere there.

This was noticed with espeak-ng 1.49 and 1.50, with the mb-fr4 voice only, not the integrated fr voice.