search

yuntaeyang / TelME

18 stars 4 forks source link

Audio file loading error #2

Closed ChauncyCai closed 5 months ago

ChauncyCai commented 6 months ago

When running python MELD/inference.py, the make_batchs() function of utils.py in the MELD folder reports an error when running. The librosa.load() function loads the mp4 file instead of the wav file. How to deal with this?

ChauncyCai commented 6 months ago

soundfile.LibsndfileError: Error opening 'D:/workspace/PyTorchProject/TelME/dataset/MELD.Raw/output_repeated_splits_test/dia0_utt0.mp4': Format not recognised.

yuntaeyang commented 6 months ago

Hello.

I don't know exactly what the problem is if it's the difference in librosa's version There is no problem loading mp4 files with librosa.

It is recommended that you set the video folder the same as the Video_Path column in train_meld_emo.csv.

ChauncyCai commented 6 months ago

I checked the relevant information and found that the librosa.load() function can only read .wav

ChauncyCai commented 6 months ago

path = "D:\workspace\PyTorchProject\TelME\dataset\MELD.Raw\output_repeated_splits_test\dia0_utt0.mp4" path_1 = "D:/workspace/PyTorchProject/TelME/dataset/MELD.Raw/output_repeated_splits_test/dia0_utt0.mp4" audio, rate = librosa.load(path)

I tested it separately using the above code, and still got an error: soundfile.LibsndfileError: Error opening 'D:\workspace\PyTorchProject\TelME\dataset\MELD.Raw\output_repeated_splits_test\dia0_utt0.mp4': Format not recognised.

yuntaeyang commented 6 months ago
스크린샷 2024-05-09 오후 3 08 11
yuntaeyang commented 6 months ago

You can see that the mp4 file is loaded using librosa in the Google Co Lab environment. Your problem may be the librosa version.

ChauncyCai commented 6 months ago

thank you very much for your help, i will try it again,

Greenteagodie commented 5 months ago

Is the problem solved？

Dengdan-1999 commented 1 month ago

I don't know exactly what the problem is if it's the difference in librosa's version There is no problem loading mp4 files with librosa.

It is recommended that you set the video folder the same as the Video_Path column in train_meld_emo.csv.

hello. have you solved this problem. would you please give me some advice? I have the same issue with you and I have tried almost all other versions of librosa like 0.7.2, 0.8.1, 0.10.0, 0.10.1, 0.10.2, 0.10.2.post1. but still post the same error! can you tell me what your setting is? my versions of the packages are as follows: av==10.0.0 librosa==0.10.0 numpy==1.23.5 opencv-python==4.7.0.72 pandas==1.5.3 scikit-learn==1.2.2 scipy==1.10.1 torch==1.13.1 torchaudio==0.13.1 torchvision==0.14.1 transformers==4.27.2 what's wrong with it? so confused!