There are different ways to produce text using speech. Most people have encountered real-time speech recognition and voice command recognition in many of the products they use, but these can also be applied to organisation-level solutions. On the other hand, offline speech recognition is best suited to situations where experts can be freed from the writing process to focus on other, more vital tasks. These various recognition methods are not mutually exclusive and the same solution can support several different use cases.
Real-time speech recognition − speech into text in real time
Real-time speech recognition is the conversion of speech into text that occurs in almost real time. Here, the speech recognition output is displayed on the speaker's own screen, thus allowing them to check the text and make any necessary changes immediately. Editing a high-quality speech recognition result is fast, and the desired information is preserved immediately. Lingsoft's clients make use of real-time speech recognition in for example the patient documentation process in health care services.
Offline speech recognition − speech recognition as part of the production process
Offline speech recognition takes place in the background of the dictation process. When a doctor has dictated their patient's information, an audio file is submitted for recognition, and the text that is produced using speech recognition is checked and revised later to ensure that it matches the audio file. In many cases, offline speech recognition is used to enhance centralised text processing either within an organisation or when it has been outsourced. Lingsoft utilises offline speech recognition as part of its medical transcription, transcription and programme subtitling services.
Voice command recognition – controlling a user interface with speech
Voice command recognition can be used to control various device or software functions. These functions are defined in advance and the speech recognition model is optimised to recognise the appropriate command words or phrases. Lingsoft's clients primarily utilise voice command recognition in automated telephone services, where callers use spoken commands as part of the service. Voice commands are also used in solutions that are based on offline and real-time speech recognition, such as medical transcription services or programme subtitling. In addition to continuous speech recognition, voice commands can also be used to change lines, delete words or change fonts, for example.