From c6e3291ee56b07792c9d8c7d7d49487b3042e01b Mon Sep 17 00:00:00 2001
From: Jean-Marie Lemercier <40563271+jmlemercier@users.noreply.github.com>
Date: Fri, 8 Dec 2023 09:30:39 +0100
Subject: [PATCH] Update create_wsj0_reverb.py to fix the sorting of WSJ0 files

---
 preprocessing/create_wsj0_reverb.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/preprocessing/create_wsj0_reverb.py b/preprocessing/create_wsj0_reverb.py
index d51f957..0a0ffe9 100644
--- a/preprocessing/create_wsj0_reverb.py
+++ b/preprocessing/create_wsj0_reverb.py
@@ -36,7 +36,7 @@ if __name__ == '__main__':
 
     splits = ['valid', 'train', 'test']
     dic_split = {"valid": "si_dt_05", "train": "si_tr_s", "test": "si_et_05"}
-    speech_lists = {split:glob(f"{os.path.join(args.wsj0_dir, dic_split[split])}/**/*.wav") for split in splits}
+    speech_lists = {split:sorted(glob(f"{os.path.join(args.wsj0_dir, dic_split[split])}/**/*.wav")) for split in splits}
     sample_rate = 16000
     output_dir = args.target_dir
 
@@ -115,4 +115,4 @@ if __name__ == '__main__':
             sf.write(os.path.join(reverberant_output_dir, output_filename), 1/output_scaling*reverberant, samplerate=sample_rate)
 
             if split == "test":
-                sf.write(os.path.join(unauralized_output_dir, output_filename), speech, samplerate=sample_rate)
\ No newline at end of file
+                sf.write(os.path.join(unauralized_output_dir, output_filename), speech, samplerate=sample_rate)
-- 
GitLab