International Journal of Computer Science Issues

More than a traditional journal...

Making the machine to hear as a human is one of the emerging technology of current technical world. If we can make machines to hear as humans, then we can use them to easily distinguish speech from music and background noises, to separate out the speech and music for special treatment, to know from direction sounds are coming, to learn which noises are typical and which are noteworthy. These machines should be able to listen and react in real time, to take appropriate action on hearing noteworthy events, to participate in ongoing activities, whether in factories, in musical performances, or in phone conversations. The existing auditory models for automatic speech recognition (ASR) has not been entirely successful, due to the highly evolved state of ASR system technologies , which are finely tuned to existing representations and to how phonetic properties of speech are manifest in those representations. One particularly promising area of machine hearing research is computational auditory scene analysis (CASA). To the extent that we can analyze sound scenes into separate meaningful components, we can achieve an advantage in tasks involving processing of those components separately. Separating speech from interference is one such application. This paper deals with the retrieval of the sound from the text queries using CASA and PAMIR algorithms with Pole-Zero filter cascade peripheral model. This paper work on content-based sound ranking system that uses biologically inspired auditory features and successfully learns a matching between acoustics and known text.