SoundNet

Code for paper "SoundNet: Learning Sound Representations from Unlabeled Video" by Yusuf Aytar, Carl Vondrick, Antonio Torralba. NIPS 2016

We learn rich natural sound representations by capitalizing on large amounts of unlabeled sound data collected in the wild. We leverage the natural synchronization between vision and sound to learn an acoustic representation using two-million unlabeled videos. We propose a student-teacher training procedure which transfers discriminative visual knowledge from well established visual models (e.g. ImageNet and PlacesCNN) into the sound modality using unlabeled video as a bridge.

Requirements

• torch7
• torch7 audio (and sox)
• torch7 hdf5 (only for feature extraction)
• probably a GPU

Pretrained Model

We provide pre-trained models that are trained over 2,000,000 unlabeled videos.

Feature Extraction

Using our network, you can extract good features for natural sounds. You can use our provided script to extract features. First, create a text file where each line lists an audio file you wish to process. We use MP3 files, but most audio formats should be supported. Then, extract features into HDF5 files like so:

$ list=data.txt th extract_feat.lua

where data.txt is this text file. It will write HDF5 files to the location of the input files with the features. You can then do anything you wish with these features.

By default, it will extract the conv7 layer. You can extract other layers like so:

$ list=data.txt layer=24 th extract_feat.lua

Advanced

If you want to write your own feature extraction code, it is very easy in Torch:

net = torch.load('soundnet8_final.t7')

sound = audio.load('file.mp3')

if sound:size(2) > 1 then sound = sound:select(2,1) end -- select first channel (mono)
sound:mul(2^-23)                                        -- make range [-256, 256]
sound = sound:view(1, 1, -1, 1)                         -- shape to BatchSize x 1 x DIM x 1
sound = sound:cuda()                                    -- ship to GPU
sound = sound:view(1, 1, -1, 1):cuda()

net:forward(sound)
feat = net.modules[24].output:float()

Training

The code for training is in main_train.lua, which will train the 8 layer SoundNet model. You can also use main_train_small.lua to train the 5 layer SoundNet model. To start training, just do:

$ CUDA_VISIBLE_DEVICES=0 th main_train.lua

The code for loading the data is in data/donkey_audio.lua. The training code will launch several threads. Each thread reads a different subset of the dataset. It will read MP3 files into a raw waveform. For the labels, it reads a large binary file that stores the class probabilities computed from ImageNet and Places networks.

If you want to fine-tune SoundNet on your own dataset, you can use main_finetune.lua.

Data

We plan to release 2 million MP3s and their corresponding class probabilities soon. Stay tuned.

References

If you use SoundNet in your research, please cite our paper:

Learning Sound Representations from Unlabeled Video 
Yusuf Aytar, Carl Vondrick, Antonio Torralba
NIPS 2016