Author(s): Mausmita Devi, Laba Kr. Thakuria, Purnendu Bikash Acharjee, Pranhari Talukdar

Speech recognition or which is more commonly known as automatic speech recognition is the process of converting an acoustic waveform into the text form which is known to the user. So, ASR also sometim es can be defined as the speech to text conversion process [2]. Application of speech recognition system more commonly defined in terms of three tasks: signal modeling, network searching, and language understanding [8]. But this paper only concern with the ASR to understand the Assamese language only. Any how the main aim of ASR is to build an interface for computers which is truly useful for users. But, it still has several problems that prevent it from being used exclusively as a method of transcription. Some of problems of ASR are the accent problem of the user or it may be for the wrong pronunciation of the word. Some problems may be caused because of the homonyms. Homophones are the words that sound similar but have different orthography. In Assamese language there are number of words are acts like homophones. One of the basic speech recognition problems may also be caused due to quality of hardware being used. This work represents an attempt to discuss the probable issues related to build ASR of Assamese language.