Verify unseen music stress inputs against production decoding

tools/media-eval/prepare_music_validation.py

@@ -4,6 +4,7 @@ import hashlib
 import json
 from pathlib import Path
 import subprocess
+import struct
 import urllib.request
 
 
@@ -23,11 +24,13 @@ def probe(path, ffprobe):
     return tags
 
 
-def pcm(path, ffmpeg):
-    data = execute([str(ffmpeg), "-v", "error", "-nostdin", "-threads", "1", "-i", str(path), "-t", "30", "-ac", "1", "-ar", "16000", "-f", "f32le", "pipe:1"])
-    if not data or len(data) > 480000 * 4 or len(data) % 4:
+def pcm(path, ffmpeg, ffprobe):
+    metadata = json.loads(execute([str(ffprobe), "-v", "error", "-show_format", "-of", "json", str(path)]))
+    duration_ms = min(30000, int(float(metadata["format"]["duration"]) * 1000 + 0.5))
+    data = execute([str(ffmpeg), "-v", "error", "-nostdin", "-threads", "1", "-ss", "0.000", "-t", f"{duration_ms / 1000:.3f}", "-i", str(path), "-af", f"aresample=16000,atrim=end_sample={duration_ms * 16},asetpts=PTS-STARTPTS", "-ac", "1", "-ar", "16000", "-threads", "1", "-f", "s16le", "pipe:1"])
+    if not data or len(data) > 480000 * 2 or len(data) % 2:
         raise ValueError("Invalid bounded PCM")
-    return sha(data)
+    return sha(b"".join(struct.pack("<f", value[0] / 32768) for value in struct.iter_unpack("<h", data)))
 
 
 def main():
@@ -74,7 +77,7 @@ def main():
                     command.extend(["-metadata", f"{key}={tags[key]}"])
             command.append(str(path))
             execute(command)
-            pcm_hash = pcm(path, args.ffmpeg)
+            pcm_hash = pcm(path, args.ffmpeg, args.ffprobe)
             if original_pcm is None:
                 original_pcm = pcm_hash
             if pcm_hash != original_pcm: