Closed meenakshi17media closed 4 years ago
Is it mandatory to use .wav file as an input ?
Yes, and the wav file must be in specific format.
can we use .mp3 or .flv audio files ?
You have to convert first with ffmpeg for example
is there any time duration of audio file defined for 1 input ?
No
is there any pre-processing steps for input audio file.
No
ok, if i directly convert .mp3 in .wav file by using online tools, it will not work right ?
ok, if i directly convert .mp3 in .wav file by using online tools, it will not work right ?
Yes
i am not able to use
vosk-model-small-en-us-0.3 | 36M | TBD | Lightweight wideband model for Android and RPi |
---|
this model in android giving error : Result %s elapsed %d milliseconds
I am giving .wav file converted by ffmpeg audio file
Make sure audio file has proper format - 16 khz 16 bit mono signed pcm.
In small model for android you don't have model.conf file, you mentioned in your document that :
conf/model.conf - provide default decoding beams and silence phones. you have to create this file yourself, it is not present in kaldi model
can you help me with this, so i can build model.conf file for android speech to text feature. And the solution you provided for STT its based pre-defined grammar or keyword or it will work without sending grammer also.
can you help me with this, i am an android person, don't have that much knowledge about kaldi and all models things.
Thanks in Advance
My questions are :
I passed a random 1 minute .mp3 audio file as an input but getting wrong results.