Revert "decoder/mad: use MAD_F_MIN and MAD_F_MAX"

This reverts commit f7ed7446. It was a bad idea, because MAD_F_MIN and MAD_F_MAX do not represent the clamping limits, but the theoretical minimum and maximum values of the mad_fixed_t data type. Closes https://github.com/MusicPlayerDaemon/MPD/issues/617

Revert "decoder/mad: use MAD_F_MIN and MAD_F_MAX"
864d6f31 · Max Kellermann · f44c67de · 864d6f31 · 864d6f31
Commit 864d6f31 authored Aug 05, 2019 by Max Kellermann
Hide whitespace changes
Inline Side-by-side

Showing with 5 additions and 1 deletion

NEWS NEWS +2 -0

MadDecoderPlugin.cxx src/decoder/plugins/MadDecoderPlugin.cxx +3 -1

No files found.
--- a/NEWS
+++ b/NEWS
 ver 0.21.13 (not yet released)
+* decoder
+  - mad: fix crackling sound (0.21.12 regression)

 ver 0.21.12 (2019/08/03)
 * decoder

--- a/src/decoder/plugins/MadDecoderPlugin.cxx
+++ b/src/decoder/plugins/MadDecoderPlugin.cxx
@@ -79,12 +79,14 @@ static inline int32_t
 mad_fixed_to_24_sample(mad_fixed_t sample) noexcept
 {
 	static constexpr unsigned bits = 24;
+	static constexpr mad_fixed_t MIN = -MAD_F_ONE;
+	static constexpr mad_fixed_t MAX = MAD_F_ONE - 1;

 	/* round */
 	sample = sample + (1L << (MAD_F_FRACBITS - bits));

 	/* quantize */
-	return Clamp(sample, MAD_F_MIN, MAD_F_MAX)
+	return Clamp(sample, MIN, MAX)
 		>> (MAD_F_FRACBITS + 1 - bits);
 }