Isolate voice and detect laugh/sreams
Budget: €250 – €750 EUR
We are trying to detect from large .mp3/.wav audios (4 hours average) the TOP 10 offsets in which someone screams or laughs.
The audios include some noise/music and a person talking. We want to know when the person is laughing or screaming.
The tasks are:
1. Receive a .mp3/.wav file (4 hours long average).
2. Isolate the person voice.
3. Detect the TOP 10 offsets where the person screams or laughs more.
4. Output the TOP 10 offsets.
The algorythm input is a .mp3/.wav (4 hours average) and the output must be the TOP 5 screaming audio offsets and the TOP 5 laugh offsets.
It is a must that the algorythm can analyze every 1 audio hour in at most 1 minute. This means the algorythm is at least 60x faster than the audio.
The algorythm will be deployed on our AWS.
The audios include some noise/music and a person talking. We want to know when the person is laughing or screaming.
The tasks are:
1. Receive a .mp3/.wav file (4 hours long average).
2. Isolate the person voice.
3. Detect the TOP 10 offsets where the person screams or laughs more.
4. Output the TOP 10 offsets.
The algorythm input is a .mp3/.wav (4 hours average) and the output must be the TOP 5 screaming audio offsets and the TOP 5 laugh offsets.
It is a must that the algorythm can analyze every 1 audio hour in at most 1 minute. This means the algorythm is at least 60x faster than the audio.
The algorythm will be deployed on our AWS.