This manually gated dataset contains all 33 transcript chunks from
@loicsumforissuchaninterestingcharacter😅👌🏽#streetballbanter#footballfans_QRBjkUAn1mQ. The audio column contains the original 48 kHz mono 96 kbps
MP3 chunks and the text column contains the aligned transcript.
SNR: 15.5324 dB
Mean spectral flatness: 0.000013
Noise floor: 0.1173024625
Peak frame RMS: 0.7013385296
Classified noisy: true… See the full description on the dataset page:
https://huggingface.co/datasets/Kppwdfgu1/loic-librosa-snr-chunks-20260722.
