At the hearing held recently due to the excommunication of the national farmers, record files and transcripts were presented as evidence. Some error points appeared in the published transcripts. If the part marked "if not" is correct in the transcript, /h-/ utterance should appear. However, analyzing the spectral characteristics of the recorded file shows that it forms five formants from low frequency component to 4kHz. This characteristic is closer to the spectral characteristic of /p-/ which is plosive sound than the spectral characteristic of /h-/ which is silent fricative. Therefore, according to spectrum analysis and context flow, "do not dig" rather than "do not" is correct. If the part of the transcript that says "The state will go with it" is correct, /g-/ utterance should appear. However, the analysis of the formant characteristics in the LPC spectrum of the recorded file shows that there are three formants from the low frequency component to 3 kHz, a valley near 3 kHz, and a small formant at 4 kHz. This characteristic is closer to the formant characteristic of /h-/ which is the fricative sound than the formant characteristic of the plosive sound /g-/. Thus, spectrum analysis and contextual flow are right, "Hook and I'll go with it" rather than "The state will go with it." In this paper, we have identified some errors in recorded files and transcripts from recent hearings. I hope that it will be helpful to accurately record the transcripts of the recording files, one of the most important evidence in the modern society, through scientific analysis.