380   return vreinterpretq_s16_u16(vmovl_u8(v));  in ConvertU8ToS16_NEON()
1286   const uint16x8_t L0 = vmovl_u8(vld1_u8(dst + 0 * BPS - 1));  in DC4_NEON()
1287   const uint16x8_t L1 = vmovl_u8(vld1_u8(dst + 1 * BPS - 1));  in DC4_NEON()
1288   const uint16x8_t L2 = vmovl_u8(vld1_u8(dst + 2 * BPS - 1));  in DC4_NEON()
1289   const uint16x8_t L3 = vmovl_u8(vld1_u8(dst + 3 * BPS - 1));  in DC4_NEON()
1432     const uint16x8_t B = vmovl_u8(A);  in DC8_NEON()
1444     const uint16x8_t L0 = vmovl_u8(vld1_u8(dst + 0 * BPS - 1));  in DC8_NEON()
1445     const uint16x8_t L1 = vmovl_u8(vld1_u8(dst + 1 * BPS - 1));  in DC8_NEON()
1446     const uint16x8_t L2 = vmovl_u8(vld1_u8(dst + 2 * BPS - 1));  in DC8_NEON()
1447     const uint16x8_t L3 = vmovl_u8(vld1_u8(dst + 3 * BPS - 1));  in DC8_NEON()
1448     const uint16x8_t L4 = vmovl_u8(vld1_u8(dst + 4 * BPS - 1));  in DC8_NEON()
1449     const uint16x8_t L5 = vmovl_u8(vld1_u8(dst + 5 * BPS - 1));  in DC8_NEON()
1450     const uint16x8_t L6 = vmovl_u8(vld1_u8(dst + 6 * BPS - 1));  in DC8_NEON()
1451     const uint16x8_t L7 = vmovl_u8(vld1_u8(dst + 7 * BPS - 1));  in DC8_NEON()
1526       const uint16x8_t L0 = vmovl_u8(vld1_u8(dst + (i + 0) * BPS - 1));  in DC16_NEON()
1527       const uint16x8_t L1 = vmovl_u8(vld1_u8(dst + (i + 1) * BPS - 1));  in DC16_NEON()
1528       const uint16x8_t L2 = vmovl_u8(vld1_u8(dst + (i + 2) * BPS - 1));  in DC16_NEON()
1529       const uint16x8_t L3 = vmovl_u8(vld1_u8(dst + (i + 3) * BPS - 1));  in DC16_NEON()
1530       const uint16x8_t L4 = vmovl_u8(vld1_u8(dst + (i + 4) * BPS - 1));  in DC16_NEON()
1531       const uint16x8_t L5 = vmovl_u8(vld1_u8(dst + (i + 5) * BPS - 1));  in DC16_NEON()
1532       const uint16x8_t L6 = vmovl_u8(vld1_u8(dst + (i + 6) * BPS - 1));  in DC16_NEON()
1533       const uint16x8_t L7 = vmovl_u8(vld1_u8(dst + (i + 7) * BPS - 1));  in DC16_NEON()