Fix NEON optimized loop filter in libtheora

Closes #3
FNA-XNA · Feb 2, 2023 · 4d36f69 · 4d36f69
1 parent 3ca5349
commit 4d36f69
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/lib/theora/arm-intrinsics/armloop.c b/lib/theora/arm-intrinsics/armloop.c
@@ -12,10 +12,10 @@ static inline uint8x8x2_t loop_filter_neon(uint8x8x4_t pix, uint16_t lim2u) {
     int16x8_t r = vreinterpretq_s16_u16(vsubl_u8(pix.val[0], pix.val[3]));
     r = vaddq_s16(r, r2);
     r = vaddq_s16(r, vshlq_n_s16(r2, 1));
-    r = vshrq_n_s16(r, 3);
+    r = vrshrq_n_s16(r, 3);
 
     uint16x8_t absR = vreinterpretq_u16_s16(vabsq_s16(r));
-    int16x8_t sgnR = vrshrq_n_s16(r, 15);
+    int16x8_t sgnR = vshrq_n_s16(r, 15);
 
     uint16x8_t absF = vminq_u16(absR, vqsubq_u16(lim2, absR)); // |f| = MIN(|R|,MAX(2L-|R|, 0));
     int16x8_t f = veorq_s16(vaddq_s16(sgnR, vreinterpretq_s16_u16(absF)), sgnR);