No abstract

This is a repository copy of Analysing acoustic model changes for active learning in automatic speech recognition.

This chapter reports the results after two years of deployment of PreLingua, a free computer-based tool for voice therapy, in different educational institutions. PreLingua gathers a set of activities that use speech processing techniques and an adapted interface to train patients who present speech development delays or special voice needs in the environment of special education. Its visual interface is especially designed for children with cognitive disabilities and maps relevant voice parameters like intensity, vocal onset, durations of sounds, fundamental frequency, and formant frequencies to visually attractive graphics. Reports of successful results of the use of PreLingua have been gathered in several countries by audiologists, speech therapists, and other professionals in the fields of voice therapy, and also, in other fields such as early stimulation, mutism, and attention-deficit disorders. This chapter brings together the experiences of these professionals on the use of the tool and how the use of an interface paradigm that maps acoustic features directly to visual elements in a screen can provide improvements in voice disorders in patients with cognitive and speech delays.