psibre

Results 94 comments of psibre

@moitreebasu1990 @pradiptadeb1990 Could you please use the experimental unitselection library to unpack the payload of the [attached waveform timeline](../files/978309/timeline_waveforms.2.zip)?

@MomnaSaeed > i couldn't understand which files i need to zip actually ? and from where ? All files starting with `100.` from the subdirectories of your voicebuilding directory.

Thanks -- all files look fine, and the audio format is also correct... To get your voice working, I would suggest that you perhaps start over with a clean project...

We've tried to inspect the waveform timeline you provided, and rebuilt a "mini" waveform timeline from the single-utterance data (`100`) you shared, both with stable v5.2 and our experimental unitselection...

Thanks for the console output. It certainly seems like the audio files are all being processed as expected, so it's a mystery to me at this point why the waveform...

To get your unit-selection voice built, I would suggest you make your data available, as well as the Urdu language component, so that we can try to reproduce the error.

Thanks for sharing the data. I've setup an [experimental voicebuilding project for your voice](/psibre/voice-ur-moona). Running tasks such as `generateAllophones` makes it quite clear that there are many things going wrong....

In your `ur.txt`, you have hyphens between *all* of the phonemes! So for example, ``` بابا b-A-b-A ``` is transcribed in a way that tells MaryTTS that this word consists...

Thanks for your feedback! Please keep in mind that some of this code was written more than 15 years ago by people with limited programming experience, and even today, the...

AFAIA, reflection is used mainly via the service loader mechanism for classpath discovery. MaryTTS was designed to work with new language and voice components "installed" by copying the jar files...