For the same purpose, the recorded audio file should be opened in a suitable audio player that displays the waveform. The waveform should be carefully examined to assess the quality of the recorded sample. If the top peaks of the waveform are missing or cut, as shown in Figure 3.11, the recording is considered to be in a clipped condition and is not suitable for speaker identification. Clipped audio samples do not contain proper formants in the voice spectrogram, which significantly affects the accuracy of speaker analysis.

Audio clipping depends on several key factors that cause the audio signal to exceed the maximum recording capacity, resulting in flattened waveform peaks and distortion. These factors include:

• Input Gain Settings: Setting the microphone or audio interface gain too high is the most common cause of clipping.



• Microphone Distance and Speaking Volume: Speaking too loudly or too close to the microphone can overload the microphone preamp and produce clipped recordings.



• System Thresholds (Analog vs. Digital): Digital clipping occurs when the signal exceeds 0 dBFS, creating harsh distortion, whereas analog clipping introduces a gradual and comparatively warmer saturation before distortion becomes noticeable.



• Multiple Gain Stages: Increasing the volume across multiple recording devices, mixers, or software plugins can collectively push the signal beyond the allowable limit, resulting in clipping.

The intensity of the recorded audio signal should remain within the normal range (the ideal level is approximately 5). If the audio intensity is too low or the recording contains excessive background noise, the waveform appears as illustrated in Figure 3.12. In such situations, the Signal-to-Noise Ratio (SNR) may not be sufficient for reliable speaker identification.

In both of the above situations, the audio sample should be re-recorded while following proper recording precautions. When the recording is performed correctly, the waveform appears as shown in Figure 3.13, indicating that the audio sample is suitable for forensic speaker identification.

After recording, the voice sample should be securely saved on a CD or memory card. The cryptographic hash values of the audio file(s) should be calculated and documented. These hash values must be enclosed along with the storage media in a properly sealed evidence package to maintain the integrity and authenticity of the digital evidence.