[vc_empty_space][vc_empty_space]
The effect of recording devices towards mfcc based speech features in a typical forensic scenario found in Indonesia
Mandasari M.I.a, Sudarsono A.S.a, Sarwono J.a, Firmanto A.D.a, Stevanus I.a
a Department of Engineering Physics, Institut Teknologi Bandung, Indonesia
[vc_row][vc_column][vc_row_inner][vc_column_inner][vc_separator css=”.vc_custom_1624529070653{padding-top: 30px !important;padding-bottom: 30px !important;}”][/vc_column_inner][/vc_row_inner][vc_row_inner layout=”boxed”][vc_column_inner width=”3/4″ css=”.vc_custom_1624695412187{border-right-width: 1px !important;border-right-color: #dddddd !important;border-right-style: solid !important;border-radius: 1px !important;}”][vc_empty_space][megatron_heading title=”Abstract” size=”size-sm” text_align=”text-left”][vc_column_text]Copyright © (2018) by International Institute of Acoustics & Vibration. All rights reserved.In real condition, the speech signal used in speaker recognition system are recorded by different types microphone. Specifically, in the forensic application in Indonesia, it is commonly found that the questioned speech sample is generally from a wire-tapped/GSM-taped conversation. Meanwhile, the reference speech sample is recorded in an investigation room by using a handheld speech recorder. This differences in recording devices are often referred to as mismatch channel in the speaker recognition research field. Experiment on this research tries to find the effect of microphone variation to the features of speech based on Mel-Frequency Cepstrum Coefficient (MFCC). Analysis towards all the MFCC based features were conducted using the statistical approach. Here, the Analysis of Variance (ANOVA) was done. The result shows that the variation of recording tools significantly (sig<0.05) affect the MFCC parameters. Further analysis was conducted by comparing the difference of the recorder to the reference recording tool (Shure SM-58 microphone). The T-Test shows that the maximum variables similarity between the reference microphone and the other recording tools is 38%. In general, the mobile phone is the most similar to the reference microphone for "U", "E", "O" vowels, the second similar to the reference for "A" vowel (32% similarity) and the third similar to the reference for "I" vowel (17% similarity). To conclude, based on the experiments conducted in this research, there is a significant difference between the MFCC based features recorded by various recording devices, i.e., microphone, telephone, and handheld recorders. Thus, it is essential for an automatic speaker recognition system to include channel compensation in it such that the effect of mismatch channel towards the system performance can be minimised.[/vc_column_text][vc_empty_space][vc_separator css=".vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}"][vc_empty_space][megatron_heading title="Author keywords" size="size-sm" text_align="text-left"][vc_column_text]Automatic speaker recognition,Channel compensation,Forensic applications,Mel frequency cepstrum coefficients,MFCC,Speaker recognition,Speaker recognition system,Statistical approach[/vc_column_text][vc_empty_space][vc_separator css=".vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}"][vc_empty_space][megatron_heading title="Indexed keywords" size="size-sm" text_align="text-left"][vc_column_text]MFCC,Sound recorder,Speaker recognition system[/vc_column_text][vc_empty_space][vc_separator css=".vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}"][vc_empty_space][megatron_heading title="Funding details" size="size-sm" text_align="text-left"][vc_column_text][/vc_column_text][vc_empty_space][vc_separator css=".vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}"][vc_empty_space][megatron_heading title="DOI" size="size-sm" text_align="text-left"][vc_column_text][/vc_column_text][/vc_column_inner][vc_column_inner width="1/4"][vc_column_text]Widget Plumx[/vc_column_text][/vc_column_inner][/vc_row_inner][/vc_column][/vc_row][vc_row][vc_column][vc_separator css=”.vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}”][/vc_column][/vc_row]