- __m128i tcim = _mm_shufflehi_epi16(_mm_shufflelo_epi16(subtcm, _MM_SHUFFLE(3, 1, 3, 1)), _MM_SHUFFLE(3, 1, 3, 1)), pix1, pix2, pix3, pix4, fracm;
- tcim = _mm_madd_epi16(_mm_add_epi16(tcim, _mm_setr_epi32(0, 0x10000, 0, 0x10000)), scalem);
- pix1 = _mm_unpacklo_epi8(_mm_loadl_epi64((const __m128i *)&pixelbase[_mm_cvtsi128_si32(tcim)]), _mm_setzero_si128());
- pix2 = _mm_unpacklo_epi8(_mm_loadl_epi64((const __m128i *)&pixelbase[_mm_cvtsi128_si32(_mm_shuffle_epi32(tcim, _MM_SHUFFLE(1, 1, 1, 1)))]), _mm_setzero_si128());
- pix3 = _mm_unpacklo_epi8(_mm_loadl_epi64((const __m128i *)&pixelbase[_mm_cvtsi128_si32(_mm_shuffle_epi32(tcim, _MM_SHUFFLE(2, 2, 2, 2)))]), _mm_setzero_si128());
- pix4 = _mm_unpacklo_epi8(_mm_loadl_epi64((const __m128i *)&pixelbase[_mm_cvtsi128_si32(_mm_shuffle_epi32(tcim, _MM_SHUFFLE(3, 3, 3, 3)))]), _mm_setzero_si128());
- fracm = _mm_srli_epi16(subtcm, 1);
+ __m128i tci = _mm_shufflehi_epi16(_mm_shufflelo_epi16(subtc, _MM_SHUFFLE(3, 1, 3, 1)), _MM_SHUFFLE(3, 1, 3, 1)), pix1, pix2, pix3, pix4, fracm;
+ tci = _mm_madd_epi16(_mm_add_epi16(tci, _mm_setr_epi32(0, 0x10000, 0, 0x10000)), tcoffset);
+ pix1 = _mm_unpacklo_epi8(_mm_loadl_epi64((const __m128i *)&pixelbase[_mm_cvtsi128_si32(tci)]), _mm_setzero_si128());
+ pix2 = _mm_unpacklo_epi8(_mm_loadl_epi64((const __m128i *)&pixelbase[_mm_cvtsi128_si32(_mm_shuffle_epi32(tci, _MM_SHUFFLE(1, 1, 1, 1)))]), _mm_setzero_si128());
+ pix3 = _mm_unpacklo_epi8(_mm_loadl_epi64((const __m128i *)&pixelbase[_mm_cvtsi128_si32(_mm_shuffle_epi32(tci, _MM_SHUFFLE(2, 2, 2, 2)))]), _mm_setzero_si128());
+ pix4 = _mm_unpacklo_epi8(_mm_loadl_epi64((const __m128i *)&pixelbase[_mm_cvtsi128_si32(_mm_shuffle_epi32(tci, _MM_SHUFFLE(3, 3, 3, 3)))]), _mm_setzero_si128());
+ fracm = _mm_srli_epi16(subtc, 1);