asmpoly.s
上传用户:zhongxx05
上传日期:2007-06-06
资源大小:33641k
文件大小:10k
- ;/* ***** BEGIN LICENSE BLOCK *****
- ; * Version: RCSL 1.0/RPSL 1.0
- ; *
- ; * Portions Copyright (c) 1995-2002 RealNetworks, Inc. All Rights Reserved.
- ; *
- ; * The contents of this file, and the files included with this file, are
- ; * subject to the current version of the RealNetworks Public Source License
- ; * Version 1.0 (the "RPSL") available at
- ; * http://www.helixcommunity.org/content/rpsl unless you have licensed
- ; * the file under the RealNetworks Community Source License Version 1.0
- ; * (the "RCSL") available at http://www.helixcommunity.org/content/rcsl,
- ; * in which case the RCSL will apply. You may also obtain the license terms
- ; * directly from RealNetworks. You may not use this file except in
- ; * compliance with the RPSL or, if you have a valid RCSL with RealNetworks
- ; * applicable to this file, the RCSL. Please see the applicable RPSL or
- ; * RCSL for the rights, obligations and limitations governing use of the
- ; * contents of the file.
- ; *
- ; * This file is part of the Helix DNA Technology. RealNetworks is the
- ; * developer of the Original Code and owns the copyrights in the portions
- ; * it created.
- ; *
- ; * This file, and the files included with this file, is distributed and made
- ; * available on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
- ; * EXPRESS OR IMPLIED, AND REALNETWORKS HEREBY DISCLAIMS ALL SUCH WARRANTIES,
- ; * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY, FITNESS
- ; * FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
- ; *
- ; * Technology Compatibility Kit Test Suite(s) Location:
- ; * http://www.helixcommunity.org/content/tck
- ; *
- ; * Contributor(s):
- ; *
- ; * ***** END LICENSE BLOCK ***** */
- AREA |.text|, CODE, READONLY
- PCM RN r0
- VB1 RN r1
- COEF RN r2
- VLO RN r0 ; must push PCM ptr to stack during inner looop
- VHI RN r3 ; temp variable
- SUM1LL RN r4
- SUM1LH RN r5
- SUM2LL RN r6
- SUM2LH RN r7
- SUM1RL RN r8
- SUM1RH RN r9
- SUM2RL RN r10
- SUM2RH RN r11
- CF1 RN r12
- CF2 RN r14
- SIGN RN r12 ; used for clipping - after discarding CF1
- MAXPOS RN r14 ; used for clipping - after discarding CF2
- I RN r12 ; overlay loop counter with CF1, SIGN
- GBLA RNDVAL
- RNDVAL SETA (1 << ((32 - 12) + (6 - 1)))
- ; C64TOS - clip 64-bit accumulator to short (no rounding)
- ; xl, xh = value (lo 32, hi 32)
- ; input assumed to have 6 fraction bits
- ; sign = temp variable to use for sign
- ; maxPos = 0x00007fff (takes 2 instr. to generate - calculating
- ; once and using repeatedly saves if you do several CTOS in a row)
- MACRO
- C64TOS $xl, $xh, $sign, $maxPos
- mov $xl, $xl, lsr #(20+6)
- orr $xl, $xl, $xh, lsl #(12-6)
- mov $sign, $xl, ASR #31
- cmp $sign, $xl, ASR #15
- eorne $xl, $sign, $maxPos
-
- MEND ; C64TOS
- ; MC0S - process 2 taps, 1 sample per channel (sample 0)
- ; x = vb1 offset
- MACRO
- MC0S $x
- ldr CF1, [COEF], #4
- ldr CF2, [COEF], #4
- ldr VLO, [VB1, #(4*($x))]
- ldr VHI, [VB1, #(4*(23 - $x))]
- smlal SUM1LL, SUM1LH, VLO, CF1
- ldr VLO, [VB1, #(4*(32 + $x))]
- rsb CF2, CF2, #0
- smlal SUM1LL, SUM1LH, VHI, CF2
- ldr VHI, [VB1, #(4*(32 + 23 - $x))]
-
- smlal SUM1RL, SUM1RH, VLO, CF1
- smlal SUM1RL, SUM1RH, VHI, CF2
- MEND ; MC0S
- ; MC1S - process 2 taps, 1 sample per channel (sample 16)
- ; x = vb1 offset
- MACRO
- MC1S $x
- ldr CF1, [COEF], #4
- ldr VLO, [VB1, #(4*($x))]
- ldr VHI, [VB1, #(4*(32 + $x))]
- smlal SUM1LL, SUM1LH, VLO, CF1
- smlal SUM1RL, SUM1RH, VHI, CF1
-
- MEND ; MC1S
- ; MC2S - process 2 taps, 2 samples per channel
- ; x = vb1 offset
- MACRO
- MC2S $x
- ; load data as far as possible in advance of using it
- ldr CF1, [COEF], #4
- ldr CF2, [COEF], #4
- ldr VLO, [VB1, #(4*($x))]
- ldr VHI, [VB1, #(4*(23 - $x))]
- smlal SUM1LL, SUM1LH, VLO, CF1
- smlal SUM2LL, SUM2LH, VLO, CF2
- rsb CF2, CF2, #0
- smlal SUM2LL, SUM2LH, VHI, CF1
- smlal SUM1LL, SUM1LH, VHI, CF2
- ldr VHI, [VB1, #(4*(32 + 23 - $x))]
- ldr VLO, [VB1, #(4*(32 + $x))]
- smlal SUM1RL, SUM1RH, VHI, CF2
- smlal SUM2RL, SUM2RH, VHI, CF1
- rsb CF2, CF2, #0
- smlal SUM1RL, SUM1RH, VLO, CF1
- smlal SUM2RL, SUM2RH, VLO, CF2
- MEND ; MC2S
- ; void PolyphaseStereo(short *pcm, int *vbuf, const int *coefBase)
- xmp3_PolyphaseStereo FUNCTION
- EXPORT xmp3_PolyphaseStereo
- stmfd sp!, {r4-r11, r14}
-
- ; clear out stack space for 2 local variables (4 bytes each)
- sub sp, sp, #8
- str PCM, [sp, #4] ; sp[1] = pcm pointer
- ; special case, output sample 0
- mov SUM1LL, #RNDVAL ; load rndVal (low 32)
- mov SUM1RL, #RNDVAL ; load rndVal (low 32)
- mov SUM1LH, #0
- mov SUM1RH, #0
- MC0S 0
- MC0S 1
- MC0S 2
- MC0S 3
- MC0S 4
- MC0S 5
- MC0S 6
- MC0S 7
- ldr PCM, [sp, #4] ; load pcm pointer
- mov MAXPOS, #0x7f00
- orr MAXPOS, MAXPOS, #0xff
-
- C64TOS SUM1LL, SUM1LH, SIGN, MAXPOS
- C64TOS SUM1RL, SUM1RH, SIGN, MAXPOS
- strh SUM1LL, [PCM, #(2*0)]
- strh SUM1RL, [PCM, #(2*1)]
- ; special case, output sample 16
- add COEF, COEF, #(4*(256-16)) ; coef = coefBase + 256 (was coefBase + 16 after MC0S block)
- add VB1, VB1, #(4*1024) ; vb1 = vbuf + 64*16
-
- mov SUM1LL, #RNDVAL ; load rndVal (low 32)
- mov SUM1RL, #RNDVAL ; load rndVal (low 32)
- mov SUM1LH, #0
- mov SUM1RH, #0
- MC1S 0
- MC1S 1
- MC1S 2
- MC1S 3
- MC1S 4
- MC1S 5
- MC1S 6
- MC1S 7
- ldr PCM, [sp, #4] ; load pcm pointer
- mov MAXPOS, #0x7f00
- orr MAXPOS, MAXPOS, #0xff
-
- C64TOS SUM1LL, SUM1LH, SIGN, MAXPOS
- C64TOS SUM1RL, SUM1RH, SIGN, MAXPOS
- strh SUM1LL, [PCM, #(2*(2*16+0))]
- strh SUM1RL, [PCM, #(2*(2*16+1))]
- ; main convolution loop: sum1L = samples 1, 2, 3, ... 15 sum2L = samples 31, 30, ... 17
- sub COEF, COEF, #(4*(264-16)) ; coef = coefBase + 16 (was coefBase + 264 after MC1S block)
- sub VB1, VB1, #(4*(1024-64)) ; vb1 = vbuf + 64 (was vbuf + 64*16 after MC1S block)
- mov I, #15 ; loop counter, count down
- add PCM, PCM, #(2*2) ; pcm+=2
-
- LoopPS
- str I, [sp, #0] ; sp[0] = i (loop counter)
- str PCM, [sp, #4] ; sp[1] = pcm (pointer to pcm buffer)
-
- mov SUM1LL, #RNDVAL ; load rndVal (low 32)
- mov SUM1RL, #RNDVAL ; load rndVal (low 32)
- mov SUM2LL, #RNDVAL ; load rndVal (low 32)
- mov SUM2RL, #RNDVAL ; load rndVal (low 32)
-
- mov SUM1LH, #0
- mov SUM1RH, #0
- mov SUM2LH, #0
- mov SUM2RH, #0
- MC2S 0
- MC2S 1
- MC2S 2
- MC2S 3
- MC2S 4
- MC2S 5
- MC2S 6
- MC2S 7
-
- add VB1, VB1, #(4*64) ; vb1 += 64
-
- ldr PCM, [sp, #4] ; load pcm pointer
- mov MAXPOS, #0x7f00
- orr MAXPOS, MAXPOS, #0xff
-
- C64TOS SUM1LL, SUM1LH, SIGN, MAXPOS
- C64TOS SUM1RL, SUM1RH, SIGN, MAXPOS
- C64TOS SUM2LL, SUM2LH, SIGN, MAXPOS
- C64TOS SUM2RL, SUM2RH, SIGN, MAXPOS
-
- ldr I, [sp, #0] ; load loop counter
- add CF2, PCM, I, lsl #3 ; CF2 = PCM + 4*i (short offset)
- strh SUM2LL, [CF2], #2 ; *(pcm + 2*2*i + 0)
- strh SUM2RL, [CF2], #2 ; *(pcm + 2*2*i + 1)
- strh SUM1LL, [PCM], #2 ; *(pcm + 0)
- strh SUM1RL, [PCM], #2 ; *(pcm + 1)
-
- subs I, I, #1
- bne LoopPS
- ; restore stack pointer
- add sp, sp, #8
- ldmfd sp!, {r4-r11, pc}
- ENDFUNC
- ;; MONO PROCESSING
- ; MC0M - process 2 taps, 1 sample (sample 0)
- ; x = vb1 offset
- MACRO
- MC0M $x
- ldr CF1, [COEF], #4
- ldr CF2, [COEF], #4
- ldr VLO, [VB1, #(4*($x))]
- ldr VHI, [VB1, #(4*(23 - $x))]
- rsb CF2, CF2, #0
- smlal SUM1LL, SUM1LH, VLO, CF1
- smlal SUM1LL, SUM1LH, VHI, CF2
- MEND ; MC0M
- ; MC1M - process 2 taps, 1 sample (sample 16)
- ; x = vb1 offset
- MACRO
- MC1M $x
- ldr CF1, [COEF], #4
- ldr VLO, [VB1, #(4*($x))]
- smlal SUM1LL, SUM1LH, VLO, CF1
-
- MEND ; MC1M
- ; MC2M - process 2 taps, 2 samples
- ; x = vb1 offset
- MACRO
- MC2M $x
- ; load data as far as possible in advance of using it
- ldr CF1, [COEF], #4
- ldr CF2, [COEF], #4
- ldr VLO, [VB1, #(4*($x))]
- ldr VHI, [VB1, #(4*(23 - $x))]
- smlal SUM1LL, SUM1LH, VLO, CF1
- smlal SUM2LL, SUM2LH, VLO, CF2
- rsb CF2, CF2, #0
- smlal SUM1LL, SUM1LH, VHI, CF2
- smlal SUM2LL, SUM2LH, VHI, CF1
- MEND ; MC2M
- ; void PolyphaseMono(short *pcm, int *vbuf, const int *coefBase)
- xmp3_PolyphaseMono FUNCTION
- EXPORT xmp3_PolyphaseMono
- stmfd sp!, {r4-r11, r14}
-
- ; clear out stack space for 4 local variables (4 bytes each)
- sub sp, sp, #8
- str PCM, [sp, #4] ; sp[1] = pcm pointer
- ; special case, output sample 0
- mov SUM1LL, #RNDVAL ; load rndVal (low 32)
- mov SUM1LH, #0
- MC0M 0
- MC0M 1
- MC0M 2
- MC0M 3
- MC0M 4
- MC0M 5
- MC0M 6
- MC0M 7
- ldr PCM, [sp, #4] ; load pcm pointer
- mov MAXPOS, #0x7f00
- orr MAXPOS, MAXPOS, #0xff
-
- C64TOS SUM1LL, SUM1LH, SIGN, MAXPOS
- strh SUM1LL, [PCM, #(2*0)]
- ; special case, output sample 16
- add COEF, COEF, #(4*(256-16)) ; coef = coefBase + 256 (was coefBase + 16 after MC0M block)
- add VB1, VB1, #(4*1024) ; vb1 = vbuf + 64*16
-
- mov SUM1LL, #RNDVAL ; load rndVal (low 32)
- mov SUM1LH, #0
- MC1M 0
- MC1M 1
- MC1M 2
- MC1M 3
- MC1M 4
- MC1M 5
- MC1M 6
- MC1M 7
- ldr PCM, [sp, #4] ; load pcm pointer
- mov MAXPOS, #0x7f00
- orr MAXPOS, MAXPOS, #0xff
-
- C64TOS SUM1LL, SUM1LH, SIGN, MAXPOS
- strh SUM1LL, [PCM, #(2*16)]
- ; main convolution loop: sum1L = samples 1, 2, 3, ... 15 sum2L = samples 31, 30, ... 17
- sub COEF, COEF, #(4*(264-16)) ; coef = coefBase + 16 (was coefBase + 264 after MC1M block)
- sub VB1, VB1, #(4*(1024-64)) ; vb1 = vbuf + 64 (was vbuf + 64*16 after MC1M block)
- mov I, #15 ; loop counter, count down
- add PCM, PCM, #(2) ; pcm++
-
- LoopPM
- str I, [sp, #0] ; sp[0] = i (loop counter)
- str PCM, [sp, #4] ; sp[1] = pcm (pointer to pcm buffer)
-
- mov SUM1LL, #RNDVAL ; load rndVal (low 32)
- mov SUM2LL, #RNDVAL ; load rndVal (low 32)
- mov SUM1LH, #0
- mov SUM2LH, #0
- MC2M 0
- MC2M 1
- MC2M 2
- MC2M 3
- MC2M 4
- MC2M 5
- MC2M 6
- MC2M 7
-
- add VB1, VB1, #(4*64) ; vb1 += 64
-
- ldr PCM, [sp, #4] ; load pcm pointer
- mov MAXPOS, #0x7f00
- orr MAXPOS, MAXPOS, #0xff
-
- C64TOS SUM1LL, SUM1LH, SIGN, MAXPOS
- C64TOS SUM2LL, SUM2LH, SIGN, MAXPOS
-
- ldr I, [sp, #0] ; load loop counter
- add CF2, PCM, I, lsl #2 ; CF2 = PCM + 2*i (short offset)
- strh SUM2LL, [CF2], #2 ; *(pcm + 2*i + 0)
- strh SUM1LL, [PCM], #2 ; *(pcm + 0) ; pcm++
-
- subs I, I, #1
- bne LoopPM
- ; restore stack pointer
- add sp, sp, #8
- ldmfd sp!, {r4-r11, pc}
- ENDFUNC
- END