ViSQOL v3: An Open Source Production Ready Objective Speech and Audio Metric.
Michael ChinenFelicia S. C. LimJan SkoglundNikita GureevFeargus O'GormanAndrew HinesPublished in: CoRR (2020)
Keyphrases
- open source
- audio visual
- audio stream
- audio signals
- emotion recognition
- broadcast news
- text to speech
- speech processing
- speech recognition
- source code
- open source software
- speaker identification
- audio features
- automatic transcription
- multi stream
- multimedia
- linear predictive coding
- cepstral features
- quality metrics
- spoken documents
- image quality metrics
- speech music discrimination
- metric learning
- signal processing
- visual speech
- human language
- digital audio
- audio recordings
- prosodic features
- audio video
- speaker verification
- case study
- distance metric
- metric space
- speech synthesis
- reduced reference
- feature vectors
- voice activity detection
- hidden markov models
- image quality
- visual information
- acoustic signals
- production system
- evaluation metrics
- speech signal
- spontaneous speech
- content based video retrieval
- digital video
- speaker recognition