Hi,
I am facing few issues related to the emotion classification. It is reporting
"encoding": "MP3",
"languageCode": "en-US",
"speakers": 1,
"audioType" :"callcenter", #To be experimented
"doVad": True

I recorded audio from headphone. Hence selected call center.

1. When I spoke angrily, it showed as "happy"
2. I give speakers to be 1. But it reports multiple speakers though there is none else. i.e. It takes as "-1" always.
3. The speech to text does not work for me. Thats ok. It still can be managed with 3rd party. But voice-to-emotion is the core of Deepaffects. Please check if it can be improved.