Add non-intrusive DS reference implementation: code, data, checkpoints, recipe

.gitattributes

@@ -1,35 +1,3 @@
-*.7z filter=lfs diff=lfs merge=lfs -text
-*.arrow filter=lfs diff=lfs merge=lfs -text
-*.bin filter=lfs diff=lfs merge=lfs -text
-*.bz2 filter=lfs diff=lfs merge=lfs -text
-*.ckpt filter=lfs diff=lfs merge=lfs -text
-*.ftz filter=lfs diff=lfs merge=lfs -text
-*.gz filter=lfs diff=lfs merge=lfs -text
-*.h5 filter=lfs diff=lfs merge=lfs -text
-*.joblib filter=lfs diff=lfs merge=lfs -text
-*.lfs.* filter=lfs diff=lfs merge=lfs -text
-*.mlmodel filter=lfs diff=lfs merge=lfs -text
-*.model filter=lfs diff=lfs merge=lfs -text
-*.msgpack filter=lfs diff=lfs merge=lfs -text
-*.npy filter=lfs diff=lfs merge=lfs -text
-*.npz filter=lfs diff=lfs merge=lfs -text
-*.onnx filter=lfs diff=lfs merge=lfs -text
-*.ot filter=lfs diff=lfs merge=lfs -text
-*.parquet filter=lfs diff=lfs merge=lfs -text
-*.pb filter=lfs diff=lfs merge=lfs -text
-*.pickle filter=lfs diff=lfs merge=lfs -text
-*.pkl filter=lfs diff=lfs merge=lfs -text
+*.wav filter=lfs diff=lfs merge=lfs -text
 *.pt filter=lfs diff=lfs merge=lfs -text
-*.pth filter=lfs diff=lfs merge=lfs -text
-*.rar filter=lfs diff=lfs merge=lfs -text
-*.safetensors filter=lfs diff=lfs merge=lfs -text
-saved_model/**/* filter=lfs diff=lfs merge=lfs -text
-*.tar.* filter=lfs diff=lfs merge=lfs -text
-*.tar filter=lfs diff=lfs merge=lfs -text
-*.tflite filter=lfs diff=lfs merge=lfs -text
-*.tgz filter=lfs diff=lfs merge=lfs -text
-*.wasm filter=lfs diff=lfs merge=lfs -text
-*.xz filter=lfs diff=lfs merge=lfs -text
-*.zip filter=lfs diff=lfs merge=lfs -text
-*.zst filter=lfs diff=lfs merge=lfs -text
-*tfevents* filter=lfs diff=lfs merge=lfs -text
+*.flac filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1,4 @@
+__pycache__/
+*.pyc
+logs/
+.ipynb_checkpoints/

DATA_LICENSES.md

@@ -0,0 +1,35 @@
+# Bundled data — sources and licenses
+
+This repository bundles audio so the recipe is self-contained. Each dataset is
+redistributed under its original license, reproduced/attributed below. Because
+**ESC-50 is CC BY-NC**, the bundled artifact as a whole is for
+**non-commercial / academic** use.
+
+## Clean speech — LibriTTS (`data/clean/`)
+
+- Subset of LibriTTS: `train-clean-100` (training), `dev-clean` (dev),
+  `test-clean` (test). Each `*.wav` ships with its `*.normalized.txt` transcript.
+- Source: https://www.openslr.org/60/
+- License: **CC BY 4.0**.
+- Citation:
+  > H. Zen, V. Dang, R. Clark, Y. Zhang, R. J. Weiss, Y. Jia, Z. Chen, Y. Wu.
+  > "LibriTTS: A Corpus Derived from LibriSpeech for Text-to-Speech," Interspeech 2019.
+
+## Noise — ESC-50 (`data/noise/`)
+
+- Full ESC-50 (2000 clips, 5 folds). Folds 1–3 used for training noise, fold 4
+  for dev, fold 5 for test (disjoint). `meta/esc50.csv` retained.
+- Source: https://github.com/karolpiczak/ESC-50
+- License: **CC BY-NC 3.0** (non-commercial).
+- Citation:
+  > K. J. Piczak. "ESC: Dataset for Environmental Sound Classification,"
+  > ACM Multimedia 2015.
+
+## ASR model — Whisper (not bundled)
+
+- `openai/whisper-base.en` is downloaded from the Hugging Face Hub at runtime
+  (not redistributed here). License: MIT (OpenAI Whisper).
+
+## Code
+
+- The source code in `ds4se/` and `scripts/` is MIT-licensed — see `LICENSE`.

LICENSE

@@ -0,0 +1,26 @@
+MIT License
+
+Copyright (c) 2026 the authors of this re-implementation
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
+NOTE: This MIT license covers the source CODE in this repository only. The
+bundled audio DATA under data/ is redistributed under separate licenses — see
+DATA_LICENSES.md. In particular ESC-50 is CC BY-NC, so the repository as a whole
+is intended for non-commercial / academic use.

README.md

@@ -1,3 +1,180 @@
 ---
-license: mit
+license: cc-by-nc-4.0
+language:
+- en
+library_name: pytorch
+tags:
+- speech-enhancement
+- automatic-speech-recognition
+- noise-robust-asr
+- distortion-suppression
+- bsrnn
+- whisper
 ---
+
+# Non-intrusive Distortion Suppression for Noise-Robust ASR (DS / DS4BSRNN)
+
+Reference implementation and **self-contained open reproduction** of:
+
+> Wei Wang, Siyi Zhao, Yanmin Qian.
+> **Advancing Non-intrusive Suppression on Enhancement Distortion for Noise
+> Robust ASR.** ICASSP 2025.
+
+This repository provides the method's source code, pretrained checkpoints, **and
+the audio data**, so the full training and evaluation pipeline runs
+out-of-the-box. The paper's headline tables are measured on a private in-house
+Mandarin corpus (Table I) and licensed CHiME-4 + DNS data (Table III), which
+cannot be redistributed. This release therefore reproduces the **method and its
+trends on fully public data** — LibriTTS (speech) + ESC-50 (noise) with a frozen
+Whisper ASR — so every number here is directly runnable and verifiable. The
+qualitative conclusions match the paper: SE enhancement degrades ASR, OA
+recovers it, and the adaptive **DS / DS4BSRNN** beat the fixed-coefficient OA
+baseline (the more variable the conditions, the larger the margin).
+
+The only runtime download is the frozen Whisper ASR from the HF Hub (see
+*Environment*).
+
+---
+
+## What is implemented
+
+- **Decoupled DS module** (`ds4se/ds_module.py`, Fig. 1b + Algorithm 1):
+  band-splits the original `X` and enhanced `X̃` complex spectrograms, runs `L`
+  Time/Band-RNN groups, and a linear+sigmoid head emits per-(band, frame)
+  coefficients `S ∈ [0,1]`. Algorithm 1 interpolates `X̂ = S·X + (1−S)·X̃`.
+- **Coupled DS4BSRNN** (`ds4se/ds4bsrnn.py`, Fig. 2): reuses the BSRNN SE's
+  internal hidden representation `H` instead of a separate band-split — fewer
+  parameters.
+- **BSRNN SE backbone** (`ds4se/bsrnn.py`): self-contained, mirrors ESPnet's
+  `BSRNNSeparator`, exposes the hidden `H`.
+- **Non-intrusive training** (`scripts/train_ds_real.py`): SE and ASR are
+  **frozen**; only the DS module is trained on the ASR (cross-entropy) loss; SE
+  gradients are detached. **Scheduled DS coefficients** (warmup bias freeze)
+  avoid the trivial `S→0` collapse.
+- **Frozen Whisper ASR** (`ds4se/asr_whisper.py`): a differentiable log-mel
+  front-end wrapping `openai/whisper-base.en` for the ASR training loss and WER.
+
+## Results
+
+End-to-end pipeline: `noisy → frozen BSRNN SE → {OA | DS | DS4BSRNN} → frozen
+Whisper → WER`. Data: LibriTTS (clean) + ESC-50 (noise), simulated at varying
+SNR. **OA** (observation-adding) is the fixed-coefficient baseline `X̂ = s·X +
+(1−s)·X̃`, with `s` tuned on dev and applied to test.
+
+**Wide-SNR test (−15…25 dB), DS modules trained on 10k utts:**
+
+| Method | Test WER (%) | Params | vs OA |
+|---|---|---|---|
+| Noisy | 25.21 | — | — |
+| SE-enhanced (BSRNN) | 26.11 | — | — |
+| + OA (s=0.4, tuned on dev) | 20.43 | — | baseline |
+| + DS (decoupled, sub-band) | 19.83 | 28.7K | **−0.60** |
+| **+ DS4BSRNN (coupled, sub-band)** | **19.39** | **10.9K** | **−1.04** |
+
+**Narrow-SNR test (−5…20 dB):** OA 11.19 / DS 11.09 (margin 0.10). The DS-over-OA
+margin grows ~6× under wide SNR, because a *fixed* OA coefficient cannot track the
+per-utterance optimum when conditions vary, while the **adaptive** DS can — the
+paper's central point. Consistent with the paper, **DS4BSRNN is the best and most
+lightweight** (beats OA by 1.04 WER with ~⅓ the parameters of the decoupled DS).
+
+## Repository layout
+
+```
+ds4se/            # the method (BSRNN, DS, DS4BSRNN, STFT, Whisper wrapper, losses, data)
+scripts/          # prepare_manifests, train_se, train_ds_real, eval_wer, inspect_ds_coef, smoke_test
+ckpt/             # pretrained: se_bsrnn.pt, ds_subband_wide.pt, ds4bsrnn_wide.pt, ds_subband_v5.pt
+data/
+  clean/          # LibriTTS clips (train-clean-100 subset / dev-clean / test-clean) + transcripts
+  noise/          # ESC-50 (audio/ + meta/esc50.csv)
+  manifests/      # train.json (10k) / dev.json (150) / test.json (300), RELATIVE paths
+run_all.sh        # one-command end-to-end recipe
+```
+
+## Environment
+
+```bash
+pip install -r requirements.txt
+```
+Tested with PyTorch 2.9 (CUDA 12.x) on a single 80 GB GPU. The frozen ASR
+(`openai/whisper-base.en`, ~290 MB) is downloaded from the HF Hub on first use;
+for an **offline** machine, pre-fetch it (e.g. `huggingface-cli download
+openai/whisper-base.en`) and set `HF_HUB_OFFLINE=1`.
+
+## Reproduce the reported numbers (pretrained checkpoints, no training)
+
+```bash
+# Wide-SNR evaluation: OA tuned on dev -> applied to test; DS + DS4BSRNN
+python scripts/eval_wer.py --se ckpt/se_bsrnn.pt \
+    --ds ckpt/ds_subband_wide.pt --ds4 ckpt/ds4bsrnn_wide.pt \
+    --snr_lo -15 --snr_hi 25
+```
+This uses only in-repo data + checkpoints and prints the table above.
+
+## Full training recipe (from scratch)
+
+```bash
+bash run_all.sh        # SE -> DS (decoupled) -> DS4BSRNN -> eval, on the bundled data
+```
+or step by step:
+```bash
+# (0) manifests are shipped; regenerate only if starting from a full LibriTTS:
+#     python scripts/prepare_manifests.py --libritts_root <LibriTTS> \
+#         --n_train 10000 --n_dev 150 --n_test 300 --relative_to .
+
+# (1) train the BSRNN speech-enhancement front-end (SI-SDR loss)
+python scripts/train_se.py --epochs 40 --num_channel 64 --num_layer 6 \
+    --out ckpt/se_bsrnn.pt
+
+# (2) non-intrusive DS training against frozen Whisper (wide SNR, unlocked adaptivity)
+python scripts/train_ds_real.py --se ckpt/se_bsrnn.pt --updates 8000 \
+    --warmup 200 --mode sub-band --lr 1e-2 --cosine --weight_decay 1e-5 \
+    --s_tv 0.0 --bias_init 1.0 --dev_items 150 --snr_lo -15 --snr_hi 25 \
+    --out ckpt/ds_subband_wide.pt
+python scripts/train_ds_real.py --se ckpt/se_bsrnn.pt --coupled --updates 8000 \
+    --warmup 200 --mode sub-band --lr 1e-2 --cosine --weight_decay 1e-5 \
+    --s_tv 0.0 --bias_init 1.0 --dev_items 150 --snr_lo -15 --snr_hi 25 \
+    --out ckpt/ds4bsrnn_wide.pt
+
+# (3) evaluate
+python scripts/eval_wer.py --se ckpt/se_bsrnn.pt \
+    --ds ckpt/ds_subband_wide.pt --ds4 ckpt/ds4bsrnn_wide.pt \
+    --snr_lo -15 --snr_hi 25
+
+# (optional) inspect the learned DS coefficients (adaptivity diagnostic)
+python scripts/inspect_ds_coef.py --ds ckpt/ds_subband_wide.pt
+```
+
+A quick self-check that exercises the core mechanisms without training:
+```bash
+python scripts/smoke_test.py
+```
+
+## Open-reproduction setup vs the paper's headline tables
+
+| Aspect | Paper (Tables I / III) | This open reproduction |
+|---|---|---|
+| Data | private in-house Mandarin / licensed CHiME-4 + DNS | public LibriTTS + ESC-50 (simulated) |
+| ASR back-end | Paraformer / large Whisper | frozen `openai/whisper-base.en` |
+| Reported quantity | absolute CER/WER | runnable WER reproducing the same trends |
+
+A few implementation details not fully specified in the paper text were fixed
+explicitly here: a 23-band 16 kHz split, DS band-embedding dim N=16, and the
+scheduled-bias `bias_init`.
+
+## Data & licenses
+
+The bundled audio is redistributed under its original licenses — see
+[`DATA_LICENSES.md`](DATA_LICENSES.md). **ESC-50 is CC BY-NC**, so the bundled
+artifact as a whole is for **non-commercial / academic** use. The repository
+*code* is MIT-licensed (see `LICENSE`).
+
+## Citation
+
+```bibtex
+@inproceedings{wang2025advancing,
+  title     = {Advancing Non-intrusive Suppression on Enhancement Distortion for Noise Robust ASR},
+  author    = {Wang, Wei and Zhao, Siyi and Qian, Yanmin},
+  booktitle = {ICASSP},
+  year      = {2025}
+}
+```

ckpt/ds4bsrnn_wide.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0a76711b60dd5af84d549f343f678d79b03602a26edc5ea107c684c8ee7ab16b
+size 51185

ckpt/ds_subband_v5.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:384767af95009e417d678326b1b7fb0192ce7ef3abf4ebbd5e9aaa37c9b44cfa
+size 152707

ckpt/ds_subband_wide.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43ea4746341815fbb787badce2ddec0147ac11715ef1474fc6ea7034dee68a6a
+size 152991

ckpt/se_bsrnn.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa579de413a7498da65659e64fbbd3246e1fb596e66586e01947cfe22d1b7645
+size 15835929

data/clean/dev-clean/1272/141231/1272_141231_000019_000000.normalized.txt

@@ -0,0 +1 @@
+"Out!" was the single snarled word of the response.

data/clean/dev-clean/1272/141231/1272_141231_000019_000000.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:402f4751d237a8809623bf0d043ffbe91c5c4c29626363bf9221001072699505
+size 164204

data/clean/dev-clean/1272/141231/1272_141231_000024_000005.normalized.txt

@@ -0,0 +1 @@
+Just thrust and parry, and victory to the stronger.

data/clean/dev-clean/1272/141231/1272_141231_000024_000005.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:45f2d03555410efd471c881f41c5d51f40846ac5cb56867488590cc4536745e5
+size 166604

data/clean/dev-clean/1272/141231/1272_141231_000026_000007.normalized.txt

@@ -0,0 +1 @@
+Self hypnosis was an accepted tool.

data/clean/dev-clean/1272/141231/1272_141231_000026_000007.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:454961b5d0217c4ec523d52be0123b07b9a8edcd3dc21d747e261d065e7e5c8f
+size 112364

data/clean/dev-clean/1272/141231/1272_141231_000027_000007.normalized.txt

@@ -0,0 +1 @@
+This is physically impossible when conscious.

data/clean/dev-clean/1272/141231/1272_141231_000027_000007.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2c29b29abaea14641c5c391a7ce22a01ee647676668f8cc19b332818b05a6924
+size 131084

data/clean/dev-clean/1272/141231/1272_141231_000031_000000.normalized.txt

@@ -0,0 +1 @@
+When the buzzer sounded he pulled his foil from his second's startled grasp, and ran forward.

data/clean/dev-clean/1272/141231/1272_141231_000031_000000.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:edc8ad70c493ef6dfeda33ad044239d03ffc0146241a4443d7dd43fa7c54f440
+size 274124

data/clean/dev-clean/1462/170138/1462_170138_000006_000001.normalized.txt

@@ -0,0 +1 @@
+When they entered the stage box on the left the first act was well under way, the scene being the interior of a cabin in the south of Ireland.

data/clean/dev-clean/1462/170138/1462_170138_000006_000001.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cebf60b4ee0df7ca0babd80d80d0d73bd4277a86cef528b9d8613daf7994d8aa
+size 368204

data/clean/dev-clean/1462/170138/1462_170138_000015_000006.normalized.txt

@@ -0,0 +1 @@
+I dare say it's quite true that there's never been any one else." Mainhall vouched for her constancy with a loftiness that made Alexander smile, even while a kind of rapid excitement was tingling through him.

data/clean/dev-clean/1462/170138/1462_170138_000015_000006.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4eb10d1459bc3a7009c4c0d381f41583a0a9526916289502f1668c6278de0746
+size 556364

data/clean/dev-clean/1462/170138/1462_170138_000022_000014.normalized.txt

@@ -0,0 +1 @@
+You've made things come your way, and I never thought you'd do it.

data/clean/dev-clean/1462/170138/1462_170138_000022_000014.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:748caa1de4aa4aed74dd4b99f319dc37424d86470654d4e7e4c664a9c3ea0e88
+size 132044

data/clean/dev-clean/1462/170142/1462_170142_000031_000001.normalized.txt

@@ -0,0 +1 @@
+Something of their troubling sweetness came back to Alexander, too.

data/clean/dev-clean/1462/170142/1462_170142_000031_000001.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:32cfd2b555cbfd1e7b85d8ce81a6dececb255dc1e4f039659ab28778079ba6f2
+size 174284

data/clean/dev-clean/1462/170142/1462_170142_000039_000000.normalized.txt

@@ -0,0 +1 @@
+"I can't," he said heavily.

data/clean/dev-clean/1462/170142/1462_170142_000039_000000.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ba57b70089b3fce773fa0c9230166d75ad787d72e16181400242b45ddab17f66
+size 106124

data/clean/dev-clean/1462/170142/1462_170142_000043_000001.normalized.txt

@@ -0,0 +1 @@
+"Yes, I know I'm cowardly. I'm afraid of myself.

data/clean/dev-clean/1462/170142/1462_170142_000043_000001.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3860b8525e5fd53c262d495dce87b27efd2488a9a7dd7516b38324110bf3a9ea
+size 155082

data/clean/dev-clean/1462/170145/1462_170145_000007_000000.normalized.txt

@@ -0,0 +1 @@
+"Of course I know, Bartley," she said at last, "that after this you won't owe me the least consideration.

data/clean/dev-clean/1462/170145/1462_170145_000007_000000.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c46242dd159b77763a54be59cd0c075c59cd6d53d12c87fef8a1392e3fef9f0b
+size 290924

data/clean/dev-clean/1462/170145/1462_170145_000008_000003.normalized.txt

@@ -0,0 +1 @@
+"Here, put your feet on this stool. You don't mean to say you walked down-and without overshoes!"

data/clean/dev-clean/1462/170145/1462_170145_000008_000003.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c38fd778b58ab137f01fd40afec773cedb2acf88ceb6c06183defc95f9734da9
+size 278924

data/clean/dev-clean/1462/170145/1462_170145_000009_000007.normalized.txt

@@ -0,0 +1 @@
+But you won't let me write to you, and I had to see you after that letter, that terrible letter you wrote me when you got home."

data/clean/dev-clean/1462/170145/1462_170145_000009_000007.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1292d5e08727ab33fc789845cc910f98296cb0ac613ebbad0622607a87fec05a
+size 298604

data/clean/dev-clean/174/168635/174_168635_000027_000004.normalized.txt

@@ -0,0 +1 @@
+To meet was to find each other.

data/clean/dev-clean/174/168635/174_168635_000027_000004.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1cfb01a5e648b7170cdc4c608a83b19a4181212a645e4c381eec72f3d2463f74
+size 109004

data/clean/dev-clean/174/50561/174_50561_000023_000004.normalized.txt

@@ -0,0 +1 @@
+The dancers once more encircle their weeping comrade, and now they are gowned in white and pink.

data/clean/dev-clean/174/50561/174_50561_000023_000004.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3fe2f83cf5fb38805c05f28450ff8f76e76d290a214943a974249d1692eed4c0
+size 265484

data/clean/dev-clean/1919/142785/1919_142785_000047_000002.normalized.txt

@@ -0,0 +1 @@
+In Jamaica it flowers about August or September, fading about the end of the year.

data/clean/dev-clean/1919/142785/1919_142785_000047_000002.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6c15470beec6845db1cd0753e1ad07a466800dfd82c133676cbb6912d21bbc0
+size 232842

data/clean/dev-clean/1919/142785/1919_142785_000120_000000.normalized.txt

@@ -0,0 +1 @@
+PESTLE AND MORTAR.--No cookery can be perfectly performed without the aid of the useful instruments shown in the engraving.

data/clean/dev-clean/1919/142785/1919_142785_000120_000000.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c162b79b9d8aeca27505fe43a6c4aa748c6eac9a4dbc180c4c6423f1fc000469
+size 342284

data/clean/dev-clean/1919/142785/1919_142785_000123_000008.normalized.txt

@@ -0,0 +1 @@
+For brown or white ragouts they should be firm, and when the quenelles are used very small, extreme delicacy will be necessary in their preparation.

data/clean/dev-clean/1919/142785/1919_142785_000123_000008.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff6219fdd1790bb5a5f7b5b8636fcf9175b9b605f6b45cb44b7ecbff99ab2943
+size 470446

data/clean/dev-clean/1988/147956/1988_147956_000002_000003.normalized.txt

@@ -0,0 +1 @@
+They made a gold ribbon across the prairie.

data/clean/dev-clean/1988/147956/1988_147956_000002_000003.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f8a30e86a1562186dbf0727e5d0929ba9383ce85ff329dfca5df1a8d4b5fbe43
+size 109964

data/clean/dev-clean/1988/147956/1988_147956_000003_000005.normalized.txt

@@ -0,0 +1 @@
+One son, Fuchs said, was well grown, and strong enough to work the land; but the father was old and frail and knew nothing about farming.