| /* |
| * This file is part of FFmpeg. |
| * |
| * FFmpeg is free software; you can redistribute it and/or |
| * modify it under the terms of the GNU Lesser General Public |
| * License as published by the Free Software Foundation; either |
| * version 2.1 of the License, or (at your option) any later version. |
| * |
| * FFmpeg is distributed in the hope that it will be useful, |
| * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| * Lesser General Public License for more details. |
| * |
| * You should have received a copy of the GNU Lesser General Public |
| * License along with FFmpeg; if not, write to the Free Software |
| * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
| */ |
| |
| #include "libavutil/aarch64/asm.S" |
| |
| const factors, align=4 |
| .float 1.0, -1.0, 1.0, -1.0 |
| endconst |
| |
| const phi_noise_0, align=4 |
| .float 1.0, 0.0, 1.0, 0.0 |
| endconst |
| |
| const phi_noise_1, align=4 |
| .float 0.0, 1.0, 0.0, -1.0 |
| .float 0.0, -1.0, 0.0, 1.0 |
| endconst |
| |
| const phi_noise_2, align=4 |
| .float -1.0, 0.0, -1.0, 0.0 |
| endconst |
| |
| const phi_noise_3, align=4 |
| .float 0.0, -1.0, 0.0, 1.0 |
| .float 0.0, 1.0, 0.0, -1.0 |
| endconst |
| |
| function ff_sbr_sum64x5_neon, export=1 |
| add x1, x0, #64*4 |
| add x2, x0, #128*4 |
| add x3, x0, #192*4 |
| add x4, x0, #256*4 |
| mov x5, #64 |
| 1: ld1 {v0.4s}, [x0] |
| ld1 {v1.4s}, [x1], #16 |
| fadd v0.4s, v0.4s, v1.4s |
| ld1 {v2.4s}, [x2], #16 |
| fadd v0.4s, v0.4s, v2.4s |
| ld1 {v3.4s}, [x3], #16 |
| fadd v0.4s, v0.4s, v3.4s |
| ld1 {v4.4s}, [x4], #16 |
| fadd v0.4s, v0.4s, v4.4s |
| st1 {v0.4s}, [x0], #16 |
| subs x5, x5, #4 |
| b.gt 1b |
| ret |
| endfunc |
| |
| function ff_sbr_sum_square_neon, export=1 |
| movi v0.4s, #0 |
| 1: ld1 {v1.4s}, [x0], #16 |
| fmla v0.4s, v1.4s, v1.4s |
| subs w1, w1, #2 |
| b.gt 1b |
| faddp v0.4s, v0.4s, v0.4s |
| faddp v0.4s, v0.4s, v0.4s |
| ret |
| endfunc |
| |
| function ff_sbr_neg_odd_64_neon, export=1 |
| mov x1, x0 |
| movi v5.4s, #1<<7, lsl #24 |
| ld2 {v0.4s, v1.4s}, [x0], #32 |
| eor v1.16b, v1.16b, v5.16b |
| ld2 {v2.4s, v3.4s}, [x0], #32 |
| .rept 3 |
| st2 {v0.4s, v1.4s}, [x1], #32 |
| eor v3.16b, v3.16b, v5.16b |
| ld2 {v0.4s, v1.4s}, [x0], #32 |
| st2 {v2.4s, v3.4s}, [x1], #32 |
| eor v1.16b, v1.16b, v5.16b |
| ld2 {v2.4s, v3.4s}, [x0], #32 |
| .endr |
| eor v3.16b, v3.16b, v5.16b |
| st2 {v0.4s, v1.4s}, [x1], #32 |
| st2 {v2.4s, v3.4s}, [x1], #32 |
| ret |
| endfunc |
| |
| function ff_sbr_qmf_pre_shuffle_neon, export=1 |
| add x1, x0, #60*4 |
| add x2, x0, #64*4 |
| mov x3, #-16 |
| mov x4, #-4 |
| movi v6.4s, #1<<7, lsl #24 |
| ld1 {v0.2s}, [x0], #8 |
| st1 {v0.2s}, [x2], #8 |
| .rept 7 |
| ld1 {v1.4s}, [x1], x3 |
| ld1 {v2.4s}, [x0], #16 |
| eor v1.16b, v1.16b, v6.16b |
| rev64 v1.4s, v1.4s |
| ext v1.16b, v1.16b, v1.16b, #8 |
| st2 {v1.4s, v2.4s}, [x2], #32 |
| .endr |
| add x1, x1, #8 |
| ld1 {v1.2s}, [x1], x4 |
| ld1 {v2.2s}, [x0], #8 |
| ld1 {v1.s}[3], [x1] |
| ld1 {v2.s}[2], [x0] |
| eor v1.16b, v1.16b, v6.16b |
| rev64 v1.4s, v1.4s |
| st2 {v1.2s, v2.2s}, [x2], #16 |
| st2 {v1.s, v2.s}[2], [x2] |
| ret |
| endfunc |
| |
| function ff_sbr_qmf_post_shuffle_neon, export=1 |
| add x2, x1, #60*4 |
| mov x3, #-16 |
| mov x4, #32 |
| movi v6.4s, #1<<7, lsl #24 |
| 1: ld1 {v0.4s}, [x2], x3 |
| ld1 {v1.4s}, [x1], #16 |
| eor v0.16b, v0.16b, v6.16b |
| rev64 v0.4s, v0.4s |
| ext v0.16b, v0.16b, v0.16b, #8 |
| st2 {v0.4s, v1.4s}, [x0], #32 |
| subs x4, x4, #4 |
| b.gt 1b |
| ret |
| endfunc |
| |
| function ff_sbr_qmf_deint_neg_neon, export=1 |
| add x1, x1, #56*4 |
| add x2, x0, #60*4 |
| mov x3, #-32 |
| mov x4, #32 |
| movi v2.4s, #1<<7, lsl #24 |
| 1: ld2 {v0.4s, v1.4s}, [x1], x3 |
| eor v0.16b, v0.16b, v2.16b |
| rev64 v1.4s, v1.4s |
| ext v1.16b, v1.16b, v1.16b, #8 |
| st1 {v0.4s}, [x2] |
| st1 {v1.4s}, [x0], #16 |
| sub x2, x2, #16 |
| subs x4, x4, #4 |
| b.gt 1b |
| ret |
| endfunc |
| |
| function ff_sbr_qmf_deint_bfly_neon, export=1 |
| add x2, x2, #60*4 |
| add x3, x0, #124*4 |
| mov x4, #64 |
| mov x5, #-16 |
| 1: ld1 {v0.4s}, [x1], #16 |
| ld1 {v1.4s}, [x2], x5 |
| rev64 v2.4s, v0.4s |
| ext v2.16b, v2.16b, v2.16b, #8 |
| rev64 v3.4s, v1.4s |
| ext v3.16b, v3.16b, v3.16b, #8 |
| fadd v1.4s, v1.4s, v2.4s |
| fsub v0.4s, v0.4s, v3.4s |
| st1 {v0.4s}, [x0], #16 |
| st1 {v1.4s}, [x3], x5 |
| subs x4, x4, #4 |
| b.gt 1b |
| ret |
| endfunc |
| |
| function ff_sbr_hf_gen_neon, export=1 |
| sxtw x4, w4 |
| sxtw x5, w5 |
| movrel x6, factors |
| ld1 {v7.4s}, [x6] |
| dup v1.4s, v0.s[0] |
| mov v2.8b, v1.8b |
| mov v2.s[2], v7.s[0] |
| mov v2.s[3], v7.s[0] |
| fmul v1.4s, v1.4s, v2.4s |
| ld1 {v0.d}[0], [x3] |
| ld1 {v0.d}[1], [x2] |
| fmul v0.4s, v0.4s, v1.4s |
| fmul v1.4s, v0.4s, v7.4s |
| rev64 v0.4s, v0.4s |
| sub x7, x5, x4 |
| add x0, x0, x4, lsl #3 |
| add x1, x1, x4, lsl #3 |
| sub x1, x1, #16 |
| 1: ld1 {v2.4s}, [x1], #16 |
| ld1 {v3.2s}, [x1] |
| fmul v4.4s, v2.4s, v1.4s |
| fmul v5.4s, v2.4s, v0.4s |
| faddp v4.4s, v4.4s, v4.4s |
| faddp v5.4s, v5.4s, v5.4s |
| faddp v4.4s, v4.4s, v4.4s |
| faddp v5.4s, v5.4s, v5.4s |
| mov v4.s[1], v5.s[0] |
| fadd v4.2s, v4.2s, v3.2s |
| st1 {v4.2s}, [x0], #8 |
| sub x1, x1, #8 |
| subs x7, x7, #1 |
| b.gt 1b |
| ret |
| endfunc |
| |
| function ff_sbr_hf_g_filt_neon, export=1 |
| sxtw x3, w3 |
| sxtw x4, w4 |
| mov x5, #40*2*4 |
| add x1, x1, x4, lsl #3 |
| 1: ld1 {v0.2s}, [x1], x5 |
| ld1 {v1.s}[0], [x2], #4 |
| fmul v2.4s, v0.4s, v1.s[0] |
| st1 {v2.2s}, [x0], #8 |
| subs x3, x3, #1 |
| b.gt 1b |
| ret |
| endfunc |
| |
| function ff_sbr_autocorrelate_neon, export=1 |
| mov x2, #38 |
| movrel x3, factors |
| ld1 {v0.4s}, [x3] |
| movi v1.4s, #0 |
| movi v2.4s, #0 |
| movi v3.4s, #0 |
| ld1 {v4.2s}, [x0], #8 |
| ld1 {v5.2s}, [x0], #8 |
| fmul v16.2s, v4.2s, v4.2s |
| fmul v17.2s, v5.2s, v4.s[0] |
| fmul v18.2s, v5.2s, v4.s[1] |
| 1: ld1 {v5.d}[1], [x0], #8 |
| fmla v1.2s, v4.2s, v4.2s |
| fmla v2.4s, v5.4s, v4.s[0] |
| fmla v3.4s, v5.4s, v4.s[1] |
| mov v4.d[0], v5.d[0] |
| mov v5.d[0], v5.d[1] |
| subs x2, x2, #1 |
| b.gt 1b |
| fmul v19.2s, v4.2s, v4.2s |
| fmul v20.2s, v5.2s, v4.s[0] |
| fmul v21.2s, v5.2s, v4.s[1] |
| fadd v22.4s, v2.4s, v20.4s |
| fsub v22.4s, v22.4s, v17.4s |
| fadd v23.4s, v3.4s, v21.4s |
| fsub v23.4s, v23.4s, v18.4s |
| rev64 v23.4s, v23.4s |
| fmul v23.4s, v23.4s, v0.4s |
| fadd v22.4s, v22.4s, v23.4s |
| st1 {v22.4s}, [x1], #16 |
| fadd v23.2s, v1.2s, v19.2s |
| fsub v23.2s, v23.2s, v16.2s |
| faddp v23.2s, v23.2s, v23.2s |
| st1 {v23.s}[0], [x1] |
| add x1, x1, #8 |
| rev64 v3.2s, v3.2s |
| fmul v3.2s, v3.2s, v0.2s |
| fadd v2.2s, v2.2s, v3.2s |
| st1 {v2.2s}, [x1] |
| add x1, x1, #16 |
| faddp v1.2s, v1.2s, v1.2s |
| st1 {v1.s}[0], [x1] |
| ret |
| endfunc |
| |
| .macro apply_noise_common |
| sxtw x3, w3 |
| sxtw x5, w5 |
| movrel x7, X(ff_sbr_noise_table) |
| add x3, x3, #1 |
| 1: and x3, x3, #0x1ff |
| add x8, x7, x3, lsl #3 |
| add x3, x3, #2 |
| ld1 {v2.4s}, [x0] |
| ld1 {v3.2s}, [x1], #8 |
| ld1 {v4.2s}, [x2], #8 |
| ld1 {v5.4s}, [x8] |
| mov v6.16b, v2.16b |
| zip1 v3.4s, v3.4s, v3.4s |
| zip1 v4.4s, v4.4s, v4.4s |
| fmla v6.4s, v1.4s, v3.4s |
| fmla v2.4s, v5.4s, v4.4s |
| fcmeq v7.4s, v3.4s, #0 |
| bif v2.16b, v6.16b, v7.16b |
| st1 {v2.4s}, [x0], #16 |
| subs x5, x5, #2 |
| b.gt 1b |
| .endm |
| |
| function ff_sbr_hf_apply_noise_0_neon, export=1 |
| movrel x9, phi_noise_0 |
| ld1 {v1.4s}, [x9] |
| apply_noise_common |
| ret |
| endfunc |
| |
| function ff_sbr_hf_apply_noise_1_neon, export=1 |
| movrel x9, phi_noise_1 |
| and x4, x4, #1 |
| add x9, x9, x4, lsl #4 |
| ld1 {v1.4s}, [x9] |
| apply_noise_common |
| ret |
| endfunc |
| |
| function ff_sbr_hf_apply_noise_2_neon, export=1 |
| movrel x9, phi_noise_2 |
| ld1 {v1.4s}, [x9] |
| apply_noise_common |
| ret |
| endfunc |
| |
| function ff_sbr_hf_apply_noise_3_neon, export=1 |
| movrel x9, phi_noise_3 |
| and x4, x4, #1 |
| add x9, x9, x4, lsl #4 |
| ld1 {v1.4s}, [x9] |
| apply_noise_common |
| ret |
| endfunc |
