A lot of challenges remain despite the rapid advancement of speaker identification technology for non-electronic disguised or physically disguised voices. This is one of the most difficult components of voice disguised that has yet to be overcome by the experts. With a normal or ideal voice, the identification process is simpler and leads to a more conclusive result. A problem arises when a suspect is identified using speech samples that have been altered to hide their identity. There are both unintended and intentional disguises in this sample content. This way a person speaks can be temporarily altered by stress, anger, concern, anxiousness, or sorrow. It is common for people to disguise themselves when they get unidentified, payment or threat phone calls. The speaker changes their voice purposefully fear of being discovered. This manuscript aims to identify a speaker for multimedia applications using a non-electronically disguised voice. In speech signal processing applications requiring non-electronic disguised voice under physical speech fluctuation, it is challenging to identify the speaker. The primary goal of this study was to regulate whether or not it was conceivable to prompt precise sentiments under disguised speech circumstances. Auditory and spectrographic investigations were used to determine the impact of different disguises on personal identity and speaker recognition performances.
Keywords: Voice disguise, MFCC, Speaker identification, Feature extraction