I am using Google's speech api for translating speech to text. I am using the http sync recognition. I am sending an opus file coded as base 64 like this:
{
"config": {
"encoding": "OGG_OPUS",
"sampleRateHertz": 16000,
"languageCode": "he-IL",
"enableWordTimeOffsets": false
},
"audio": {
"content": "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"
}
}
This is the result:
{
"results": [
{
"alternatives": [
{
"transcript": "סעדיה פז",
"confidence": 1
}
]
}
]
}
The confidence
should not be 1!
Until couple of days ago, the confidence for this message and some others would be somewhere in the range between 0 to 1.
But now every request returns with a confidence
of 1.
Is there any way to get it back? Does the api changed? I counldn't find any change in google's documentations
Thanks!