{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.7.12","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"none","dataSources":[{"sourceId":44224,"databundleVersionId":5188730,"sourceType":"competition"}],"dockerImageVersionId":30407,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":false}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import pandas as pd\nimport numpy as np\nimport io\nimport os\nimport math\nimport tarfile\nimport multiprocessing\n\nimport scipy\nimport librosa\nimport boto3\nfrom botocore import UNSIGNED\nfrom botocore.config import Config\nimport requests\nimport matplotlib\nimport matplotlib.pyplot as plt\nimport time\nfrom IPython.display import Audio, display\nimport IPython.display as ipd\n\nimport torchaudio\nprint(\"Done !\")","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","execution":{"iopub.status.busy":"2023-03-21T21:56:00.99907Z","iopub.execute_input":"2023-03-21T21:56:00.999559Z","iopub.status.idle":"2023-03-21T21:56:01.009323Z","shell.execute_reply.started":"2023-03-21T21:56:00.999519Z","shell.execute_reply":"2023-03-21T21:56:01.007745Z"},"trusted":true},"outputs":[{"name":"stdout","text":"Done !\n","output_type":"stream"}],"execution_count":27},{"cell_type":"code","source":"TRAIN_AUDIO_PATH = \"/kaggle/input/birdclef-2023/train_audio\"\nMETADATA_TRAIN = \"/kaggle/input/birdclef-2023/train_metadata.csv\"","metadata":{"execution":{"iopub.status.busy":"2023-03-21T21:52:15.639409Z","iopub.execute_input":"2023-03-21T21:52:15.640412Z","iopub.status.idle":"2023-03-21T21:52:15.647065Z","shell.execute_reply.started":"2023-03-21T21:52:15.640366Z","shell.execute_reply":"2023-03-21T21:52:15.645745Z"},"trusted":true},"outputs":[],"execution_count":19},{"cell_type":"code","source":"train = pd.read_csv(METADATA_TRAIN)\ntrain.head()\nclass_labels = os.listdir(TRAIN_AUDIO_PATH)\nprint(f\"Number of class : {len(class_labels)} \")","metadata":{"execution":{"iopub.status.busy":"2023-03-21T21:52:15.648304Z","iopub.execute_input":"2023-03-21T21:52:15.649186Z","iopub.status.idle":"2023-03-21T21:52:15.777917Z","shell.execute_reply.started":"2023-03-21T21:52:15.64914Z","shell.execute_reply":"2023-03-21T21:52:15.776621Z"},"trusted":true},"outputs":[{"name":"stdout","text":"Number of class : 264 \n","output_type":"stream"}],"execution_count":20},{"cell_type":"code","source":"# Construct file path by concatenating fold and file name\ntrain['relative_path'] = '/' + train['filename'].astype(str) \ntrain.head()","metadata":{"execution":{"iopub.status.busy":"2023-03-21T21:52:15.780197Z","iopub.execute_input":"2023-03-21T21:52:15.780585Z","iopub.status.idle":"2023-03-21T21:52:15.80919Z","shell.execute_reply.started":"2023-03-21T21:52:15.78055Z","shell.execute_reply":"2023-03-21T21:52:15.807675Z"},"trusted":true},"outputs":[{"execution_count":21,"output_type":"execute_result","data":{"text/plain":" primary_label secondary_labels type latitude longitude \\\n0 abethr1 [] ['song'] 4.3906 38.2788 \n1 abethr1 [] ['call'] -2.9524 38.2921 \n2 abethr1 [] ['song'] -2.9524 38.2921 \n3 abethr1 [] ['song'] -2.9524 38.2921 \n4 abethr1 [] ['call', 'song'] -2.9524 38.2921 \n\n scientific_name common_name author \\\n0 Turdus tephronotus African Bare-eyed Thrush Rolf A. de By \n1 Turdus tephronotus African Bare-eyed Thrush James Bradley \n2 Turdus tephronotus African Bare-eyed Thrush James Bradley \n3 Turdus tephronotus African Bare-eyed Thrush James Bradley \n4 Turdus tephronotus African Bare-eyed Thrush James Bradley \n\n license rating \\\n0 Creative Commons Attribution-NonCommercial-Sha... 4.0 \n1 Creative Commons Attribution-NonCommercial-Sha... 3.5 \n2 Creative Commons Attribution-NonCommercial-Sha... 3.5 \n3 Creative Commons Attribution-NonCommercial-Sha... 5.0 \n4 Creative Commons Attribution-NonCommercial-Sha... 4.5 \n\n url filename \\\n0 https://www.xeno-canto.org/128013 abethr1/XC128013.ogg \n1 https://www.xeno-canto.org/363501 abethr1/XC363501.ogg \n2 https://www.xeno-canto.org/363502 abethr1/XC363502.ogg \n3 https://www.xeno-canto.org/363503 abethr1/XC363503.ogg \n4 https://www.xeno-canto.org/363504 abethr1/XC363504.ogg \n\n relative_path \n0 /abethr1/XC128013.ogg \n1 /abethr1/XC363501.ogg \n2 /abethr1/XC363502.ogg \n3 /abethr1/XC363503.ogg \n4 /abethr1/XC363504.ogg ","text/html":"
\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n
primary_labelsecondary_labelstypelatitudelongitudescientific_namecommon_nameauthorlicenseratingurlfilenamerelative_path
0abethr1[]['song']4.390638.2788Turdus tephronotusAfrican Bare-eyed ThrushRolf A. de ByCreative Commons Attribution-NonCommercial-Sha...4.0https://www.xeno-canto.org/128013abethr1/XC128013.ogg/abethr1/XC128013.ogg
1abethr1[]['call']-2.952438.2921Turdus tephronotusAfrican Bare-eyed ThrushJames BradleyCreative Commons Attribution-NonCommercial-Sha...3.5https://www.xeno-canto.org/363501abethr1/XC363501.ogg/abethr1/XC363501.ogg
2abethr1[]['song']-2.952438.2921Turdus tephronotusAfrican Bare-eyed ThrushJames BradleyCreative Commons Attribution-NonCommercial-Sha...3.5https://www.xeno-canto.org/363502abethr1/XC363502.ogg/abethr1/XC363502.ogg
3abethr1[]['song']-2.952438.2921Turdus tephronotusAfrican Bare-eyed ThrushJames BradleyCreative Commons Attribution-NonCommercial-Sha...5.0https://www.xeno-canto.org/363503abethr1/XC363503.ogg/abethr1/XC363503.ogg
4abethr1[]['call', 'song']-2.952438.2921Turdus tephronotusAfrican Bare-eyed ThrushJames BradleyCreative Commons Attribution-NonCommercial-Sha...4.5https://www.xeno-canto.org/363504abethr1/XC363504.ogg/abethr1/XC363504.ogg
\n
"},"metadata":{}}],"execution_count":21},{"cell_type":"code","source":"path_class_1 = os.path.join(TRAIN_AUDIO_PATH, class_labels[0])\nprint(f\"Class 1: {path_class_1}\")\naudio_class_1_path = os.path.join(path_class_1, os.listdir(path_class_1)[0])\nprint(f\"Audio 1 of class 1: {audio_class_1_path}\")\nprint(\"Class 1: \", path_class_1.split(\"/\")[-1])","metadata":{"execution":{"iopub.status.busy":"2023-03-21T21:52:15.810889Z","iopub.execute_input":"2023-03-21T21:52:15.811419Z","iopub.status.idle":"2023-03-21T21:52:15.82384Z","shell.execute_reply.started":"2023-03-21T21:52:15.811361Z","shell.execute_reply":"2023-03-21T21:52:15.822375Z"},"trusted":true},"outputs":[{"name":"stdout","text":"Class 1: /kaggle/input/birdclef-2023/train_audio/yetgre1\nAudio 1 of class 1: /kaggle/input/birdclef-2023/train_audio/yetgre1/XC247367.ogg\nClass 1: yetgre1\n","output_type":"stream"}],"execution_count":22},{"cell_type":"markdown","source":"# Hiểu dữ liệu âm thanh và chuyển đổi tín hiệu âm thanh thành biểu diễn hình ảnh bằng cách sử dụng biểu đồ quang phổ","metadata":{}},{"cell_type":"markdown","source":"## 1. Convert the sound to tensor","metadata":{}},{"cell_type":"code","source":"filename = \"/kaggle/input/birdclef-2023/train_audio/abethr1/XC128013.ogg\"\nwaveform, sample_rate = torchaudio.load(filepath=filename, num_frames=3)\nprint(f'waveform tensor with 3 frames: {waveform} \\n')\nwaveform, sample_rate = torchaudio.load(filepath=filename, num_frames=3, frame_offset =2)\nprint(f'waveform tensor with 2 frame_offsets: {waveform} \\n')\nwaveform, sample_rate = torchaudio.load(filepath=filename)\nprint(f'waveform tensor: {waveform}' )","metadata":{"execution":{"iopub.status.busy":"2023-03-21T21:53:44.261375Z","iopub.execute_input":"2023-03-21T21:53:44.262008Z","iopub.status.idle":"2023-03-21T21:53:44.544458Z","shell.execute_reply.started":"2023-03-21T21:53:44.261954Z","shell.execute_reply":"2023-03-21T21:53:44.542721Z"},"trusted":true},"outputs":[{"name":"stdout","text":"waveform tensor with 3 frames: tensor([[0., 0., 0.]]) \n\nwaveform tensor with 2 frame_offsets: tensor([[0., 0., 0.]]) \n\nwaveform tensor: tensor([[ 0.0000e+00, 0.0000e+00, 0.0000e+00, ..., -9.1553e-05,\n 9.4604e-04, 9.1553e-04]])\n","output_type":"stream"}],"execution_count":23},{"cell_type":"markdown","source":"2. Plot the waweform","metadata":{}},{"cell_type":"code","source":"def plot_audio(filename):\n waveform, sample_rate = torchaudio.load(filename)\n\n print(\"Shape of waveform: {}\".format(waveform.size()))\n print(\"Sample rate of waveform: {}\".format(sample_rate))\n\n plt.figure()\n plt.plot(waveform.t().numpy())\n\n return waveform, sample_rate","metadata":{"execution":{"iopub.status.busy":"2023-03-21T21:54:38.667149Z","iopub.execute_input":"2023-03-21T21:54:38.667624Z","iopub.status.idle":"2023-03-21T21:54:38.676244Z","shell.execute_reply.started":"2023-03-21T21:54:38.667568Z","shell.execute_reply":"2023-03-21T21:54:38.67467Z"},"trusted":true},"outputs":[],"execution_count":24},{"cell_type":"code","source":"waveform, sample_rate = plot_audio(filename)\nipd.Audio(waveform.numpy(), rate=sample_rate)","metadata":{"execution":{"iopub.status.busy":"2023-03-21T21:56:09.357039Z","iopub.execute_input":"2023-03-21T21:56:09.35752Z","iopub.status.idle":"2023-03-21T21:56:10.160256Z","shell.execute_reply.started":"2023-03-21T21:56:09.357469Z","shell.execute_reply":"2023-03-21T21:56:10.159195Z"},"trusted":true},"outputs":[{"name":"stdout","text":"Shape of waveform: torch.Size([1, 1459513])\nSample rate of waveform: 32000\n","output_type":"stream"},{"execution_count":28,"output_type":"execute_result","data":{"text/plain":"","text/html":"\n \n "},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"
","image/png":"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\n"},"metadata":{}}],"execution_count":28},{"cell_type":"markdown","source":"- **Spectrogram**: Create a spectrogram from a waveform.\n- **MFCC**: Create the Mel-frequency cepstrum coefficients from a waveform.\n- **MelSpectrogram**: Create Mel spectrograms from a waveform by using the `STFT` function in PyTorch.","metadata":{}},{"cell_type":"code","source":"def load_audio_files(path: str, label:str):\n\n dataset = []\n walker = sorted(str(p) for p in Path(path).glob(f'*.wav'))\n\n for i, file_path in enumerate(walker):\n path, filename = os.path.split(file_path)\n speaker, _ = os.path.splitext(filename)\n speaker_id, utterance_number = speaker.split(\"_nohash_\")\n utterance_number = int(utterance_number)\n \n # Load audio\n waveform, sample_rate = torchaudio.load(file_path)\n dataset.append([waveform, sample_rate, label, speaker_id, utterance_number])\n \n return dataset","metadata":{},"outputs":[],"execution_count":null}]}