Skip to content

Loading Audio Data

NWaveDataGen converts a folder of .wav files into spike-compatible, time-binned features ready for the nwave hardware frontend.

Expected folder structure:

data_for_nwave_commands/
├── yes/
│   ├── sample_001.wav
│   └── ...
└── no/
    ├── sample_001.wav
    └── ...

Each subdirectory is a class label. NWaveDataGen handles resampling, padding/trimming, and applying the H1 hardware filter bank.

import torch
import matplotlib.pyplot as plt
from nwavesdk import NWaveDataGen, NWaveDataloaderConfig

data_config = NWaveDataloaderConfig(
    batch_size=16,
    val_split=0.2,
    test_split=0.0,
    shuffle_train=True,
    random_state=42,
    num_workers=2,
)

dm = NWaveDataGen(
    data_parent="../data_for_nwave_commands",  # one subfolder per class
    sample_rate=16000,
    recording_duration_s=1.0,    # pad/trim all clips to 1 second
    sim_time_s=8e-3,             # 8 ms time bins → 125 timesteps for a 1-second clip
    dataloader_config=data_config,
    task="classification",
    return_filename=False,
)

loaders = dm.dataloaders()
train_loader = loaders["train"]
val_loader   = loaders["val"]

print(f"Train batches: {len(train_loader)}")
print(f"Val   batches: {len(val_loader)}")

2026-05-20 09:21:40,995 - root - WARNING - Using 13 valid freqs out of 16 for sr=16000Hz (Nyquist=8000.0Hz).



Classes (loading wavs):   0%|          | 0/2 [00:00<?, ?it/s]



Loading no:   0%|          | 0/3536 [00:00<?, ?it/s]



Loading yes:   0%|          | 0/3625 [00:00<?, ?it/s]



Filtering no:   0%|          | 0/3536 [00:00<?, ?it/s]



Filtering yes:   0%|          | 0/3625 [00:00<?, ?it/s]


Train batches: 359
Val   batches: 90

x, y = next(iter(train_loader))

# Visualise the filter-bank output for the first sample: frequency channels over time.
sample = x[0].T.numpy()  # (frequency_channels, timesteps)

fig, ax = plt.subplots(figsize=(10, 3))
ax.imshow(sample, aspect="auto", origin="lower", interpolation="nearest",
          extent=[0, x.shape[1], 0, x.shape[2]])
ax.set_xlabel("Timestep")
ax.set_ylabel("Frequency channel")
ax.set_title(f"Frontend output — label: {y[0].item()}")
plt.tight_layout()
plt.show()

png

Further reading