diff --git a/vp8/common/alloccommon.c b/vp8/common/alloccommon.c
index 869f6e20f2dd0106b61cd3a119a237e3ad162c73..fb9aa2e2504c3f708b0f54a2508a7ca21fc963a2 100644
--- a/vp8/common/alloccommon.c
+++ b/vp8/common/alloccommon.c
@@ -188,7 +188,7 @@ void vp8_setup_version(VP8_COMMON *cm)
void vp8_create_common(VP8_COMMON *oci)
{
vp8_machine_specific_config(oci);
- vp8_default_coef_probs(oci);
+
vp8_init_mbmode_probs(oci);
vp8_default_bmode_probs(oci->fc.bmode_prob);
diff --git a/vp8/common/default_coef_probs.h b/vp8/common/default_coef_probs.h
new file mode 100755
index 0000000000000000000000000000000000000000..0d195636bcb70b8b2f372ffa91cf5c35b7b9ecca
--- /dev/null
+++ b/vp8/common/default_coef_probs.h
@@ -0,0 +1,188 @@
+/*
+ * Copyright (c) 2010 The WebM project authors. All Rights Reserved.
+ *
+ * Use of this source code is governed by a BSD-style license
+ * that can be found in the LICENSE file in the root of the source
+ * tree. An additional intellectual property rights grant can be found
+ * in the file PATENTS. All contributing project authors may
+ * be found in the AUTHORS file in the root of the source tree.
+*/
+
+
+/*Generated file, included by entropy.c*/
+
+
+static const vp8_prob default_coef_probs [BLOCK_TYPES]
+ [COEF_BANDS]
+ [PREV_COEF_CONTEXTS]
+ [ENTROPY_NODES] =
+{
+ { /* Block Type ( 0 ) */
+ { /* Coeff Band ( 0 )*/
+ { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 },
+ { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 },
+ { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 1 )*/
+ { 253, 136, 254, 255, 228, 219, 128, 128, 128, 128, 128 },
+ { 189, 129, 242, 255, 227, 213, 255, 219, 128, 128, 128 },
+ { 106, 126, 227, 252, 214, 209, 255, 255, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 2 )*/
+ { 1, 98, 248, 255, 236, 226, 255, 255, 128, 128, 128 },
+ { 181, 133, 238, 254, 221, 234, 255, 154, 128, 128, 128 },
+ { 78, 134, 202, 247, 198, 180, 255, 219, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 3 )*/
+ { 1, 185, 249, 255, 243, 255, 128, 128, 128, 128, 128 },
+ { 184, 150, 247, 255, 236, 224, 128, 128, 128, 128, 128 },
+ { 77, 110, 216, 255, 236, 230, 128, 128, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 4 )*/
+ { 1, 101, 251, 255, 241, 255, 128, 128, 128, 128, 128 },
+ { 170, 139, 241, 252, 236, 209, 255, 255, 128, 128, 128 },
+ { 37, 116, 196, 243, 228, 255, 255, 255, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 5 )*/
+ { 1, 204, 254, 255, 245, 255, 128, 128, 128, 128, 128 },
+ { 207, 160, 250, 255, 238, 128, 128, 128, 128, 128, 128 },
+ { 102, 103, 231, 255, 211, 171, 128, 128, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 6 )*/
+ { 1, 152, 252, 255, 240, 255, 128, 128, 128, 128, 128 },
+ { 177, 135, 243, 255, 234, 225, 128, 128, 128, 128, 128 },
+ { 80, 129, 211, 255, 194, 224, 128, 128, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 7 )*/
+ { 1, 1, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
+ { 246, 1, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
+ { 255, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 }
+ }
+ },
+ { /* Block Type ( 1 ) */
+ { /* Coeff Band ( 0 )*/
+ { 198, 35, 237, 223, 193, 187, 162, 160, 145, 155, 62 },
+ { 131, 45, 198, 221, 172, 176, 220, 157, 252, 221, 1 },
+ { 68, 47, 146, 208, 149, 167, 221, 162, 255, 223, 128 }
+ },
+ { /* Coeff Band ( 1 )*/
+ { 1, 149, 241, 255, 221, 224, 255, 255, 128, 128, 128 },
+ { 184, 141, 234, 253, 222, 220, 255, 199, 128, 128, 128 },
+ { 81, 99, 181, 242, 176, 190, 249, 202, 255, 255, 128 }
+ },
+ { /* Coeff Band ( 2 )*/
+ { 1, 129, 232, 253, 214, 197, 242, 196, 255, 255, 128 },
+ { 99, 121, 210, 250, 201, 198, 255, 202, 128, 128, 128 },
+ { 23, 91, 163, 242, 170, 187, 247, 210, 255, 255, 128 }
+ },
+ { /* Coeff Band ( 3 )*/
+ { 1, 200, 246, 255, 234, 255, 128, 128, 128, 128, 128 },
+ { 109, 178, 241, 255, 231, 245, 255, 255, 128, 128, 128 },
+ { 44, 130, 201, 253, 205, 192, 255, 255, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 4 )*/
+ { 1, 132, 239, 251, 219, 209, 255, 165, 128, 128, 128 },
+ { 94, 136, 225, 251, 218, 190, 255, 255, 128, 128, 128 },
+ { 22, 100, 174, 245, 186, 161, 255, 199, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 5 )*/
+ { 1, 182, 249, 255, 232, 235, 128, 128, 128, 128, 128 },
+ { 124, 143, 241, 255, 227, 234, 128, 128, 128, 128, 128 },
+ { 35, 77, 181, 251, 193, 211, 255, 205, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 6 )*/
+ { 1, 157, 247, 255, 236, 231, 255, 255, 128, 128, 128 },
+ { 121, 141, 235, 255, 225, 227, 255, 255, 128, 128, 128 },
+ { 45, 99, 188, 251, 195, 217, 255, 224, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 7 )*/
+ { 1, 1, 251, 255, 213, 255, 128, 128, 128, 128, 128 },
+ { 203, 1, 248, 255, 255, 128, 128, 128, 128, 128, 128 },
+ { 137, 1, 177, 255, 224, 255, 128, 128, 128, 128, 128 }
+ }
+ },
+ { /* Block Type ( 2 ) */
+ { /* Coeff Band ( 0 )*/
+ { 253, 9, 248, 251, 207, 208, 255, 192, 128, 128, 128 },
+ { 175, 13, 224, 243, 193, 185, 249, 198, 255, 255, 128 },
+ { 73, 17, 171, 221, 161, 179, 236, 167, 255, 234, 128 }
+ },
+ { /* Coeff Band ( 1 )*/
+ { 1, 95, 247, 253, 212, 183, 255, 255, 128, 128, 128 },
+ { 239, 90, 244, 250, 211, 209, 255, 255, 128, 128, 128 },
+ { 155, 77, 195, 248, 188, 195, 255, 255, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 2 )*/
+ { 1, 24, 239, 251, 218, 219, 255, 205, 128, 128, 128 },
+ { 201, 51, 219, 255, 196, 186, 128, 128, 128, 128, 128 },
+ { 69, 46, 190, 239, 201, 218, 255, 228, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 3 )*/
+ { 1, 191, 251, 255, 255, 128, 128, 128, 128, 128, 128 },
+ { 223, 165, 249, 255, 213, 255, 128, 128, 128, 128, 128 },
+ { 141, 124, 248, 255, 255, 128, 128, 128, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 4 )*/
+ { 1, 16, 248, 255, 255, 128, 128, 128, 128, 128, 128 },
+ { 190, 36, 230, 255, 236, 255, 128, 128, 128, 128, 128 },
+ { 149, 1, 255, 128, 128, 128, 128, 128, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 5 )*/
+ { 1, 226, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
+ { 247, 192, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
+ { 240, 128, 255, 128, 128, 128, 128, 128, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 6 )*/
+ { 1, 134, 252, 255, 255, 128, 128, 128, 128, 128, 128 },
+ { 213, 62, 250, 255, 255, 128, 128, 128, 128, 128, 128 },
+ { 55, 93, 255, 128, 128, 128, 128, 128, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 7 )*/
+ { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 },
+ { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 },
+ { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 }
+ }
+ },
+ { /* Block Type ( 3 ) */
+ { /* Coeff Band ( 0 )*/
+ { 202, 24, 213, 235, 186, 191, 220, 160, 240, 175, 255 },
+ { 126, 38, 182, 232, 169, 184, 228, 174, 255, 187, 128 },
+ { 61, 46, 138, 219, 151, 178, 240, 170, 255, 216, 128 }
+ },
+ { /* Coeff Band ( 1 )*/
+ { 1, 112, 230, 250, 199, 191, 247, 159, 255, 255, 128 },
+ { 166, 109, 228, 252, 211, 215, 255, 174, 128, 128, 128 },
+ { 39, 77, 162, 232, 172, 180, 245, 178, 255, 255, 128 }
+ },
+ { /* Coeff Band ( 2 )*/
+ { 1, 52, 220, 246, 198, 199, 249, 220, 255, 255, 128 },
+ { 124, 74, 191, 243, 183, 193, 250, 221, 255, 255, 128 },
+ { 24, 71, 130, 219, 154, 170, 243, 182, 255, 255, 128 }
+ },
+ { /* Coeff Band ( 3 )*/
+ { 1, 182, 225, 249, 219, 240, 255, 224, 128, 128, 128 },
+ { 149, 150, 226, 252, 216, 205, 255, 171, 128, 128, 128 },
+ { 28, 108, 170, 242, 183, 194, 254, 223, 255, 255, 128 }
+ },
+ { /* Coeff Band ( 4 )*/
+ { 1, 81, 230, 252, 204, 203, 255, 192, 128, 128, 128 },
+ { 123, 102, 209, 247, 188, 196, 255, 233, 128, 128, 128 },
+ { 20, 95, 153, 243, 164, 173, 255, 203, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 5 )*/
+ { 1, 222, 248, 255, 216, 213, 128, 128, 128, 128, 128 },
+ { 168, 175, 246, 252, 235, 205, 255, 255, 128, 128, 128 },
+ { 47, 116, 215, 255, 211, 212, 255, 255, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 6 )*/
+ { 1, 121, 236, 253, 212, 214, 255, 255, 128, 128, 128 },
+ { 141, 84, 213, 252, 201, 202, 255, 219, 128, 128, 128 },
+ { 42, 80, 160, 240, 162, 185, 255, 205, 128, 128, 128 }
+ },
+ { /* Coeff Band ( 7 )*/
+ { 1, 1, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
+ { 244, 1, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
+ { 238, 1, 255, 128, 128, 128, 128, 128, 128, 128, 128 }
+ }
+ }
+};
diff --git a/vp8/common/defaultcoefcounts.h b/vp8/common/defaultcoefcounts.h
deleted file mode 100644
index 293e74269f4479a0b1c695b783c306a07e4c14a6..0000000000000000000000000000000000000000
--- a/vp8/common/defaultcoefcounts.h
+++ /dev/null
@@ -1,26 +0,0 @@
-/*
- * Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- * Use of this source code is governed by a BSD-style license
- * that can be found in the LICENSE file in the root of the source
- * tree. An additional intellectual property rights grant can be found
- * in the file PATENTS. All contributing project authors may
- * be found in the AUTHORS file in the root of the source tree.
- */
-
-#ifndef __DEFAULTCOEFCOUNTS_H
-#define __DEFAULTCOEFCOUNTS_H
-
-#include "entropy.h"
-
-extern const unsigned int vp8_default_coef_counts[BLOCK_TYPES]
- [COEF_BANDS]
- [PREV_COEF_CONTEXTS]
- [MAX_ENTROPY_TOKENS];
-
-extern const unsigned int vp8_default_coef_counts_8x8[BLOCK_TYPES]
- [COEF_BANDS]
- [PREV_COEF_CONTEXTS]
- [MAX_ENTROPY_TOKENS];
-
-#endif
\ No newline at end of file
diff --git a/vp8/common/entropy.c b/vp8/common/entropy.c
index ca37aab22abd44b54245dbc4795c3f7333601eb0..34b54b9a77139ec92589b8b281cea2e0c0f2fe3d 100644
--- a/vp8/common/entropy.c
+++ b/vp8/common/entropy.c
@@ -15,6 +15,7 @@
#include "string.h"
#include "blockd.h"
#include "onyxc_int.h"
+#include "vpx_mem/vpx_mem.h"
#define uchar unsigned char /* typedefs can clash */
#define uint unsigned int
@@ -193,36 +194,13 @@ vp8_extra_bit_struct vp8_extra_bits[12] =
#endif
{ 0, 0, 0, 0}
};
-#include "defaultcoefcounts.h"
+
+#include "default_coef_probs.h"
void vp8_default_coef_probs(VP8_COMMON *pc)
{
- int h = 0;
-
- do
- {
- int i = 0;
-
- do
- {
- int k = 0;
-
- do
- {
- unsigned int branch_ct [ENTROPY_NODES] [2];
- vp8_tree_probs_from_distribution(
- MAX_ENTROPY_TOKENS, vp8_coef_encodings, vp8_coef_tree,
- pc->fc.coef_probs[h][i][k],
- branch_ct,
- vp8_default_coef_counts[h][i][k],
- 256, 1);
-
- }
- while (++k < PREV_COEF_CONTEXTS);
- }
- while (++i < COEF_BANDS);
- }
- while (++h < BLOCK_TYPES);
+ vpx_memcpy(pc->fc.coef_probs, default_coef_probs,
+ sizeof(default_coef_probs));
#if CONFIG_T8X8
h = 0;
do
@@ -250,7 +228,6 @@ void vp8_default_coef_probs(VP8_COMMON *pc)
#endif
}
-
void vp8_coef_tree_initialize()
{
init_bit_trees();
diff --git a/vp8/common/x86/idctllm_sse2.asm b/vp8/common/x86/idctllm_sse2.asm
index 34a7e18aea727ec7175f57de50d3832d4a13df2a..83d3765ff8e1ec51de76f7c124c59488530515d7 100644
--- a/vp8/common/x86/idctllm_sse2.asm
+++ b/vp8/common/x86/idctllm_sse2.asm
@@ -11,7 +11,7 @@
%include "vpx_ports/x86_abi_support.asm"
-;void idct_dequant_0_2x_sse2
+;void vp8_idct_dequant_0_2x_sse2
; (
; short *qcoeff - 0
; short *dequant - 1
@@ -21,8 +21,8 @@
; int blk_stride - 5
; )
-global sym(idct_dequant_0_2x_sse2)
-sym(idct_dequant_0_2x_sse2):
+global sym(vp8_idct_dequant_0_2x_sse2)
+sym(vp8_idct_dequant_0_2x_sse2):
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 6
@@ -97,8 +97,8 @@ sym(idct_dequant_0_2x_sse2):
pop rbp
ret
-global sym(idct_dequant_full_2x_sse2)
-sym(idct_dequant_full_2x_sse2):
+global sym(vp8_idct_dequant_full_2x_sse2)
+sym(vp8_idct_dequant_full_2x_sse2):
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 7
@@ -353,7 +353,7 @@ sym(idct_dequant_full_2x_sse2):
pop rbp
ret
-;void idct_dequant_dc_0_2x_sse2
+;void vp8_idct_dequant_dc_0_2x_sse2
; (
; short *qcoeff - 0
; short *dequant - 1
@@ -362,8 +362,8 @@ sym(idct_dequant_full_2x_sse2):
; int dst_stride - 4
; short *dc - 5
; )
-global sym(idct_dequant_dc_0_2x_sse2)
-sym(idct_dequant_dc_0_2x_sse2):
+global sym(vp8_idct_dequant_dc_0_2x_sse2)
+sym(vp8_idct_dequant_dc_0_2x_sse2):
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 7
@@ -438,8 +438,8 @@ sym(idct_dequant_dc_0_2x_sse2):
pop rbp
ret
-global sym(idct_dequant_dc_full_2x_sse2)
-sym(idct_dequant_dc_full_2x_sse2):
+global sym(vp8_idct_dequant_dc_full_2x_sse2)
+sym(vp8_idct_dequant_dc_full_2x_sse2):
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 7
diff --git a/vp8/common/x86/loopfilter_sse2.asm b/vp8/common/x86/loopfilter_sse2.asm
index 4efff7eb584d227a16b9dabe66094ffbf0814b68..295609c58167a1cb6018b0f3366bd6f467217ee0 100644
--- a/vp8/common/x86/loopfilter_sse2.asm
+++ b/vp8/common/x86/loopfilter_sse2.asm
@@ -1395,8 +1395,8 @@ sym(vp8_loop_filter_simple_horizontal_edge_sse2):
neg rax
; calculate mask
- movdqu xmm1, [rsi+2*rax] ; p1
- movdqu xmm0, [rdi] ; q1
+ movdqa xmm1, [rsi+2*rax] ; p1
+ movdqa xmm0, [rdi] ; q1
movdqa xmm2, xmm1
movdqa xmm7, xmm0
movdqa xmm4, xmm0
@@ -1406,8 +1406,8 @@ sym(vp8_loop_filter_simple_horizontal_edge_sse2):
pand xmm1, [GLOBAL(tfe)] ; set lsb of each byte to zero
psrlw xmm1, 1 ; abs(p1-q1)/2
- movdqu xmm5, [rsi+rax] ; p0
- movdqu xmm4, [rsi] ; q0
+ movdqa xmm5, [rsi+rax] ; p0
+ movdqa xmm4, [rsi] ; q0
movdqa xmm0, xmm4 ; q0
movdqa xmm6, xmm5 ; p0
psubusb xmm5, xmm4 ; p0-=q0
@@ -1449,7 +1449,7 @@ sym(vp8_loop_filter_simple_horizontal_edge_sse2):
psubsb xmm3, xmm0 ; q0-= q0 add
pxor xmm3, [GLOBAL(t80)] ; unoffset
- movdqu [rsi], xmm3 ; write back
+ movdqa [rsi], xmm3 ; write back
; now do +3 side
psubsb xmm5, [GLOBAL(t1s)] ; +3 instead of +4
@@ -1465,7 +1465,7 @@ sym(vp8_loop_filter_simple_horizontal_edge_sse2):
paddsb xmm6, xmm0 ; p0+= p0 add
pxor xmm6, [GLOBAL(t80)] ; unoffset
- movdqu [rsi+rax], xmm6 ; write back
+ movdqa [rsi+rax], xmm6 ; write back
; begin epilog
pop rdi
@@ -1507,17 +1507,17 @@ sym(vp8_loop_filter_simple_vertical_edge_sse2):
lea rdx, [rsi + rax*4]
lea rcx, [rdx + rax]
- movdqu xmm0, [rsi] ; (high 96 bits unused) 03 02 01 00
- movdqu xmm1, [rdx] ; (high 96 bits unused) 43 42 41 40
- movdqu xmm2, [rdi] ; 13 12 11 10
- movdqu xmm3, [rcx] ; 53 52 51 50
+ movd xmm0, [rsi] ; (high 96 bits unused) 03 02 01 00
+ movd xmm1, [rdx] ; (high 96 bits unused) 43 42 41 40
+ movd xmm2, [rdi] ; 13 12 11 10
+ movd xmm3, [rcx] ; 53 52 51 50
punpckldq xmm0, xmm1 ; (high 64 bits unused) 43 42 41 40 03 02 01 00
punpckldq xmm2, xmm3 ; 53 52 51 50 13 12 11 10
- movdqu xmm4, [rsi + rax*2] ; 23 22 21 20
- movdqu xmm5, [rdx + rax*2] ; 63 62 61 60
- movdqu xmm6, [rdi + rax*2] ; 33 32 31 30
- movdqu xmm7, [rcx + rax*2] ; 73 72 71 70
+ movd xmm4, [rsi + rax*2] ; 23 22 21 20
+ movd xmm5, [rdx + rax*2] ; 63 62 61 60
+ movd xmm6, [rdi + rax*2] ; 33 32 31 30
+ movd xmm7, [rcx + rax*2] ; 73 72 71 70
punpckldq xmm4, xmm5 ; 63 62 61 60 23 22 21 20
punpckldq xmm6, xmm7 ; 73 72 71 70 33 32 31 30
@@ -1540,17 +1540,17 @@ sym(vp8_loop_filter_simple_vertical_edge_sse2):
lea rdx, [rsi + rax*4]
lea rcx, [rdx + rax]
- movdqu xmm4, [rsi] ; 83 82 81 80
- movdqu xmm1, [rdx] ; c3 c2 c1 c0
- movdqu xmm6, [rdi] ; 93 92 91 90
- movdqu xmm3, [rcx] ; d3 d2 d1 d0
+ movd xmm4, [rsi] ; 83 82 81 80
+ movd xmm1, [rdx] ; c3 c2 c1 c0
+ movd xmm6, [rdi] ; 93 92 91 90
+ movd xmm3, [rcx] ; d3 d2 d1 d0
punpckldq xmm4, xmm1 ; c3 c2 c1 c0 83 82 81 80
punpckldq xmm6, xmm3 ; d3 d2 d1 d0 93 92 91 90
- movdqu xmm0, [rsi + rax*2] ; a3 a2 a1 a0
- movdqu xmm5, [rdx + rax*2] ; e3 e2 e1 e0
- movdqu xmm2, [rdi + rax*2] ; b3 b2 b1 b0
- movdqu xmm7, [rcx + rax*2] ; f3 f2 f1 f0
+ movd xmm0, [rsi + rax*2] ; a3 a2 a1 a0
+ movd xmm5, [rdx + rax*2] ; e3 e2 e1 e0
+ movd xmm2, [rdi + rax*2] ; b3 b2 b1 b0
+ movd xmm7, [rcx + rax*2] ; f3 f2 f1 f0
punpckldq xmm0, xmm5 ; e3 e2 e1 e0 a3 a2 a1 a0
punpckldq xmm2, xmm7 ; f3 f2 f1 f0 b3 b2 b1 b0
diff --git a/vp8/common/x86/postproc_mmx.c b/vp8/common/x86/postproc_mmx.c
deleted file mode 100644
index 6b6321ace3d672617105085cfb36956262e7135b..0000000000000000000000000000000000000000
--- a/vp8/common/x86/postproc_mmx.c
+++ /dev/null
@@ -1,1508 +0,0 @@
-/*
- * Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- * Use of this source code is governed by a BSD-style license
- * that can be found in the LICENSE file in the root of the source
- * tree. An additional intellectual property rights grant can be found
- * in the file PATENTS. All contributing project authors may
- * be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#include <math.h>
-#include <stdlib.h>
-#include "vpx_scale/yv12config.h"
-#include "pragmas.h"
-
-#define VP8_FILTER_WEIGHT 128
-#define VP8_FILTER_SHIFT 7
-
-
-
-/* static constants */
-__declspec(align(16))
-const static short Blur[48] =
-{
-
- 16, 16, 16, 16, 16, 16, 16, 16,
- 16, 16, 16, 16, 16, 16, 16, 16,
- 64, 64, 64, 64, 64, 64, 64, 64,
- 16, 16, 16, 16, 16, 16, 16, 16,
- 16, 16, 16, 16, 16, 16, 16, 16,
- 0, 0, 0, 0, 0, 0, 0, 0,
-
-};
-#define RD __declspec(align(16)) __int64 rd = 0x0040004000400040;
-#define R4D2 __declspec(align(16)) __int64 rd42[2] = {0x0004000400040004,0x0004000400040004};
-
-#ifndef RELOCATEABLE
-const static RD;
-const static R4D2;
-#endif
-
-
-/* external references */
-extern double vp8_gaussian(double sigma, double mu, double x);
-extern short vp8_rv[];
-extern int vp8_q2mbl(int x) ;
-
-
-
-void vp8_post_proc_down_and_across_mmx
-(
- unsigned char *src_ptr,
- unsigned char *dst_ptr,
- int src_pixels_per_line,
- int dst_pixels_per_line,
- int rows,
- int cols,
- int flimit
-)
-{
-#ifdef RELOCATEABLE
- RD
- R4D2
-#endif
-
- __asm
- {
- push ebx
- lea ebx, Blur
- movd mm2, flimit
- punpcklwd mm2, mm2
- punpckldq mm2, mm2
-
- mov esi, src_ptr
- mov edi, dst_ptr
-
- mov ecx, DWORD PTR rows
- mov eax, src_pixels_per_line ;
- destination pitch?
- pxor mm0, mm0 ;
- mm0 = 00000000
-
- nextrow:
-
- xor edx, edx ;
-
- clear out edx for use as loop counter
- nextcol:
-
- pxor mm7, mm7 ;
-
- mm7 = 00000000
- movq mm6, [ebx + 32 ] ;
- mm6 = kernel 2 taps
- movq mm3, [esi] ;
- mm4 = r0 p0..p7
- punpcklbw mm3, mm0 ;
- mm3 = p0..p3
- movq mm1, mm3 ;
- mm1 = p0..p3
- pmullw mm3, mm6 ;
- mm3 *= kernel 2 modifiers
-
- movq mm6, [ebx + 48] ;
- mm6 = kernel 3 taps
- movq mm5, [esi + eax] ;
- mm4 = r1 p0..p7
- punpcklbw mm5, mm0 ;
- mm5 = r1 p0..p3
- pmullw mm6, mm5 ;
- mm6 *= p0..p3 * kernel 3 modifiers
- paddusw mm3, mm6 ;
- mm3 += mm6
-
- ;
- thresholding
- movq mm7, mm1 ;
- mm7 = r0 p0..p3
- psubusw mm7, mm5 ;
- mm7 = r0 p0..p3 - r1 p0..p3
- psubusw mm5, mm1 ;
- mm5 = r1 p0..p3 - r0 p0..p3
- paddusw mm7, mm5 ;
- mm7 = abs(r0 p0..p3 - r1 p0..p3)
- pcmpgtw mm7, mm2
-
- movq mm6, [ebx + 64 ] ;
- mm6 = kernel 4 modifiers
- movq mm5, [esi + 2*eax] ;
- mm4 = r2 p0..p7
- punpcklbw mm5, mm0 ;
- mm5 = r2 p0..p3
- pmullw mm6, mm5 ;
- mm5 *= kernel 4 modifiers
- paddusw mm3, mm6 ;
- mm3 += mm5
-
- ;
- thresholding
- movq mm6, mm1 ;
- mm6 = r0 p0..p3
- psubusw mm6, mm5 ;
- mm6 = r0 p0..p3 - r2 p0..p3
- psubusw mm5, mm1 ;
- mm5 = r2 p0..p3 - r2 p0..p3
- paddusw mm6, mm5 ;
- mm6 = abs(r0 p0..p3 - r2 p0..p3)
- pcmpgtw mm6, mm2
- por mm7, mm6 ;
- accumulate thresholds
-
-
- neg eax
- movq mm6, [ebx ] ;
- kernel 0 taps
- movq mm5, [esi+2*eax] ;
- mm4 = r-2 p0..p7
- punpcklbw mm5, mm0 ;
- mm5 = r-2 p0..p3
- pmullw mm6, mm5 ;
- mm5 *= kernel 0 modifiers
- paddusw mm3, mm6 ;
- mm3 += mm5
-
- ;
- thresholding
- movq mm6, mm1 ;
- mm6 = r0 p0..p3
- psubusw mm6, mm5 ;
- mm6 = p0..p3 - r-2 p0..p3
- psubusw mm5, mm1 ;
- mm5 = r-2 p0..p3 - p0..p3
- paddusw mm6, mm5 ;
- mm6 = abs(r0 p0..p3 - r-2 p0..p3)
- pcmpgtw mm6, mm2
- por mm7, mm6 ;
- accumulate thresholds
-
- movq mm6, [ebx + 16] ;
- kernel 1 taps
- movq mm4, [esi+eax] ;
- mm4 = r-1 p0..p7
- punpcklbw mm4, mm0 ;
- mm4 = r-1 p0..p3
- pmullw mm6, mm4 ;
- mm4 *= kernel 1 modifiers.
- paddusw mm3, mm6 ;
- mm3 += mm5
-
- ;
- thresholding
- movq mm6, mm1 ;
- mm6 = r0 p0..p3
- psubusw mm6, mm4 ;
- mm6 = p0..p3 - r-2 p0..p3
- psubusw mm4, mm1 ;
- mm5 = r-1 p0..p3 - p0..p3
- paddusw mm6, mm4 ;
- mm6 = abs(r0 p0..p3 - r-1 p0..p3)
- pcmpgtw mm6, mm2
- por mm7, mm6 ;
- accumulate thresholds
-
-
- paddusw mm3, rd ;
- mm3 += round value
- psraw mm3, VP8_FILTER_SHIFT ;
- mm3 /= 128
-
- pand mm1, mm7 ;
- mm1 select vals > thresh from source
- pandn mm7, mm3 ;
- mm7 select vals < thresh from blurred result
- paddusw mm1, mm7 ;
- combination
-
- packuswb mm1, mm0 ;
- pack to bytes
-
- movd [edi], mm1 ;
- neg eax ;
- pitch is positive
-
-
- add esi, 4
- add edi, 4
- add edx, 4
-
- cmp edx, cols
- jl nextcol
- // done with the all cols, start the across filtering in place
- sub esi, edx
- sub edi, edx
-
-
- push eax
- xor edx, edx
- mov eax, [edi-4];
-
- acrossnextcol:
- pxor mm7, mm7 ;
- mm7 = 00000000
- movq mm6, [ebx + 32 ] ;
- movq mm4, [edi+edx] ;
- mm4 = p0..p7
- movq mm3, mm4 ;
- mm3 = p0..p7
- punpcklbw mm3, mm0 ;
- mm3 = p0..p3
- movq mm1, mm3 ;
- mm1 = p0..p3
- pmullw mm3, mm6 ;
- mm3 *= kernel 2 modifiers
-
- movq mm6, [ebx + 48]
- psrlq mm4, 8 ;
- mm4 = p1..p7
- movq mm5, mm4 ;
- mm5 = p1..p7
- punpcklbw mm5, mm0 ;
- mm5 = p1..p4
- pmullw mm6, mm5 ;
- mm6 *= p1..p4 * kernel 3 modifiers
- paddusw mm3, mm6 ;
- mm3 += mm6
-
- ;
- thresholding
- movq mm7, mm1 ;
- mm7 = p0..p3
- psubusw mm7, mm5 ;
- mm7 = p0..p3 - p1..p4
- psubusw mm5, mm1 ;
- mm5 = p1..p4 - p0..p3
- paddusw mm7, mm5 ;
- mm7 = abs(p0..p3 - p1..p4)
- pcmpgtw mm7, mm2
-
- movq mm6, [ebx + 64 ]
- psrlq mm4, 8 ;
- mm4 = p2..p7
- movq mm5, mm4 ;
- mm5 = p2..p7
- punpcklbw mm5, mm0 ;
- mm5 = p2..p5
- pmullw mm6, mm5 ;
- mm5 *= kernel 4 modifiers
- paddusw mm3, mm6 ;
- mm3 += mm5
-
- ;
- thresholding
- movq mm6, mm1 ;
- mm6 = p0..p3
- psubusw mm6, mm5 ;
- mm6 = p0..p3 - p1..p4
- psubusw mm5, mm1 ;
- mm5 = p1..p4 - p0..p3
- paddusw mm6, mm5 ;
- mm6 = abs(p0..p3 - p1..p4)
- pcmpgtw mm6, mm2
- por mm7, mm6 ;
- accumulate thresholds
-
-
- movq mm6, [ebx ]
- movq mm4, [edi+edx-2] ;
- mm4 = p-2..p5
- movq mm5, mm4 ;
- mm5 = p-2..p5
- punpcklbw mm5, mm0 ;
- mm5 = p-2..p1
- pmullw mm6, mm5 ;
- mm5 *= kernel 0 modifiers
- paddusw mm3, mm6 ;
- mm3 += mm5
-
- ;
- thresholding
- movq mm6, mm1 ;
- mm6 = p0..p3
- psubusw mm6, mm5 ;
- mm6 = p0..p3 - p1..p4
- psubusw mm5, mm1 ;
- mm5 = p1..p4 - p0..p3
- paddusw mm6, mm5 ;
- mm6 = abs(p0..p3 - p1..p4)
- pcmpgtw mm6, mm2
- por mm7, mm6 ;
- accumulate thresholds
-
- movq mm6, [ebx + 16]
- psrlq mm4, 8 ;
- mm4 = p-1..p5
- punpcklbw mm4, mm0 ;
- mm4 = p-1..p2
- pmullw mm6, mm4 ;
- mm4 *= kernel 1 modifiers.
- paddusw mm3, mm6 ;
- mm3 += mm5
-
- ;
- thresholding
- movq mm6, mm1 ;
- mm6 = p0..p3
- psubusw mm6, mm4 ;
- mm6 = p0..p3 - p1..p4
- psubusw mm4, mm1 ;
- mm5 = p1..p4 - p0..p3
- paddusw mm6, mm4 ;
- mm6 = abs(p0..p3 - p1..p4)
- pcmpgtw mm6, mm2
- por mm7, mm6 ;
- accumulate thresholds
-
- paddusw mm3, rd ;
- mm3 += round value
- psraw mm3, VP8_FILTER_SHIFT ;
- mm3 /= 128
-
- pand mm1, mm7 ;
- mm1 select vals > thresh from source
- pandn mm7, mm3 ;
- mm7 select vals < thresh from blurred result
- paddusw mm1, mm7 ;
- combination
-
- packuswb mm1, mm0 ;
- pack to bytes
- mov DWORD PTR [edi+edx-4], eax ;
- store previous four bytes
- movd eax, mm1
-
- add edx, 4
- cmp edx, cols
- jl acrossnextcol;
-
- mov DWORD PTR [edi+edx-4], eax
- pop eax
-
- // done with this rwo
- add esi, eax ;
- next line
- mov eax, dst_pixels_per_line ;
- destination pitch?
- add edi, eax ;
- next destination
- mov eax, src_pixels_per_line ;
- destination pitch?
-
- dec ecx ;
- decrement count
- jnz nextrow ;
- next row
- pop ebx
-
- }
-}
-
-
-
-void vp8_post_proc_down_and_across_xmm
-(
- unsigned char *src_ptr,
- unsigned char *dst_ptr,
- int src_pixels_per_line,
- int dst_pixels_per_line,
- int rows,
- int cols,
- int flimit
-)
-{
-#ifdef RELOCATEABLE
- R4D2
-#endif
-
- __asm
- {
- movd xmm2, flimit
- punpcklwd xmm2, xmm2
- punpckldq xmm2, xmm2
- punpcklqdq xmm2, xmm2
-
- mov esi, src_ptr
- mov edi, dst_ptr
-
- mov ecx, DWORD PTR rows
- mov eax, src_pixels_per_line ;
- destination pitch?
- pxor xmm0, xmm0 ;
- mm0 = 00000000
-
- nextrow:
-
- xor edx, edx ;
-
- clear out edx for use as loop counter
- nextcol:
- movq xmm3, QWORD PTR [esi] ;
-
- mm4 = r0 p0..p7
- punpcklbw xmm3, xmm0 ;
- mm3 = p0..p3
- movdqa xmm1, xmm3 ;
- mm1 = p0..p3
- psllw xmm3, 2 ;
-
- movq xmm5, QWORD PTR [esi + eax] ;
- mm4 = r1 p0..p7
- punpcklbw xmm5, xmm0 ;
- mm5 = r1 p0..p3
- paddusw xmm3, xmm5 ;
- mm3 += mm6
-
- ;
- thresholding
- movdqa xmm7, xmm1 ;
- mm7 = r0 p0..p3
- psubusw xmm7, xmm5 ;
- mm7 = r0 p0..p3 - r1 p0..p3
- psubusw xmm5, xmm1 ;
- mm5 = r1 p0..p3 - r0 p0..p3
- paddusw xmm7, xmm5 ;
- mm7 = abs(r0 p0..p3 - r1 p0..p3)
- pcmpgtw xmm7, xmm2
-
- movq xmm5, QWORD PTR [esi + 2*eax] ;
- mm4 = r2 p0..p7
- punpcklbw xmm5, xmm0 ;
- mm5 = r2 p0..p3
- paddusw xmm3, xmm5 ;
- mm3 += mm5
-
- ;
- thresholding
- movdqa xmm6, xmm1 ;
- mm6 = r0 p0..p3
- psubusw xmm6, xmm5 ;
- mm6 = r0 p0..p3 - r2 p0..p3
- psubusw xmm5, xmm1 ;
- mm5 = r2 p0..p3 - r2 p0..p3
- paddusw xmm6, xmm5 ;
- mm6 = abs(r0 p0..p3 - r2 p0..p3)
- pcmpgtw xmm6, xmm2
- por xmm7, xmm6 ;
- accumulate thresholds
-
-
- neg eax
- movq xmm5, QWORD PTR [esi+2*eax] ;
- mm4 = r-2 p0..p7
- punpcklbw xmm5, xmm0 ;
- mm5 = r-2 p0..p3
- paddusw xmm3, xmm5 ;
- mm3 += mm5
-
- ;
- thresholding
- movdqa xmm6, xmm1 ;
- mm6 = r0 p0..p3
- psubusw xmm6, xmm5 ;
- mm6 = p0..p3 - r-2 p0..p3
- psubusw xmm5, xmm1 ;
- mm5 = r-2 p0..p3 - p0..p3
- paddusw xmm6, xmm5 ;
- mm6 = abs(r0 p0..p3 - r-2 p0..p3)
- pcmpgtw xmm6, xmm2
- por xmm7, xmm6 ;
- accumulate thresholds
-
- movq xmm4, QWORD PTR [esi+eax] ;
- mm4 = r-1 p0..p7
- punpcklbw xmm4, xmm0 ;
- mm4 = r-1 p0..p3
- paddusw xmm3, xmm4 ;
- mm3 += mm5
-
- ;
- thresholding
- movdqa xmm6, xmm1 ;
- mm6 = r0 p0..p3
- psubusw xmm6, xmm4 ;
- mm6 = p0..p3 - r-2 p0..p3
- psubusw xmm4, xmm1 ;
- mm5 = r-1 p0..p3 - p0..p3
- paddusw xmm6, xmm4 ;
- mm6 = abs(r0 p0..p3 - r-1 p0..p3)
- pcmpgtw xmm6, xmm2
- por xmm7, xmm6 ;
- accumulate thresholds
-
-
- paddusw xmm3, rd42 ;
- mm3 += round value
- psraw xmm3, 3 ;
- mm3 /= 8
-
- pand xmm1, xmm7 ;
- mm1 select vals > thresh from source
- pandn xmm7, xmm3 ;
- mm7 select vals < thresh from blurred result
- paddusw xmm1, xmm7 ;
- combination
-
- packuswb xmm1, xmm0 ;
- pack to bytes
- movq QWORD PTR [edi], xmm1 ;
-
- neg eax ;
- pitch is positive
- add esi, 8
- add edi, 8
-
- add edx, 8
- cmp edx, cols
-
- jl nextcol
-
- // done with the all cols, start the across filtering in place
- sub esi, edx
- sub edi, edx
-
- xor edx, edx
- movq mm0, QWORD PTR [edi-8];
-
- acrossnextcol:
- movq xmm7, QWORD PTR [edi +edx -2]
- movd xmm4, DWORD PTR [edi +edx +6]
-
- pslldq xmm4, 8
- por xmm4, xmm7
-
- movdqa xmm3, xmm4
- psrldq xmm3, 2
- punpcklbw xmm3, xmm0 ;
- mm3 = p0..p3
- movdqa xmm1, xmm3 ;
- mm1 = p0..p3
- psllw xmm3, 2
-
-
- movdqa xmm5, xmm4
- psrldq xmm5, 3
- punpcklbw xmm5, xmm0 ;
- mm5 = p1..p4
- paddusw xmm3, xmm5 ;
- mm3 += mm6
-
- ;
- thresholding
- movdqa xmm7, xmm1 ;
- mm7 = p0..p3
- psubusw xmm7, xmm5 ;
- mm7 = p0..p3 - p1..p4
- psubusw xmm5, xmm1 ;
- mm5 = p1..p4 - p0..p3
- paddusw xmm7, xmm5 ;
- mm7 = abs(p0..p3 - p1..p4)
- pcmpgtw xmm7, xmm2
-
- movdqa xmm5, xmm4
- psrldq xmm5, 4
- punpcklbw xmm5, xmm0 ;
- mm5 = p2..p5
- paddusw xmm3, xmm5 ;
- mm3 += mm5
-
- ;
- thresholding
- movdqa xmm6, xmm1 ;
- mm6 = p0..p3
- psubusw xmm6, xmm5 ;
- mm6 = p0..p3 - p1..p4
- psubusw xmm5, xmm1 ;
- mm5 = p1..p4 - p0..p3
- paddusw xmm6, xmm5 ;
- mm6 = abs(p0..p3 - p1..p4)
- pcmpgtw xmm6, xmm2
- por xmm7, xmm6 ;
- accumulate thresholds
-
-
- movdqa xmm5, xmm4 ;
- mm5 = p-2..p5
- punpcklbw xmm5, xmm0 ;
- mm5 = p-2..p1
- paddusw xmm3, xmm5 ;
- mm3 += mm5
-
- ;
- thresholding
- movdqa xmm6, xmm1 ;
- mm6 = p0..p3
- psubusw xmm6, xmm5 ;
- mm6 = p0..p3 - p1..p4
- psubusw xmm5, xmm1 ;
- mm5 = p1..p4 - p0..p3
- paddusw xmm6, xmm5 ;
- mm6 = abs(p0..p3 - p1..p4)
- pcmpgtw xmm6, xmm2
- por xmm7, xmm6 ;
- accumulate thresholds
-
- psrldq xmm4, 1 ;
- mm4 = p-1..p5
- punpcklbw xmm4, xmm0 ;
- mm4 = p-1..p2
- paddusw xmm3, xmm4 ;
- mm3 += mm5
-
- ;
- thresholding
- movdqa xmm6, xmm1 ;
- mm6 = p0..p3
- psubusw xmm6, xmm4 ;
- mm6 = p0..p3 - p1..p4
- psubusw xmm4, xmm1 ;
- mm5 = p1..p4 - p0..p3
- paddusw xmm6, xmm4 ;
- mm6 = abs(p0..p3 - p1..p4)
- pcmpgtw xmm6, xmm2
- por xmm7, xmm6 ;
- accumulate thresholds
-
- paddusw xmm3, rd42 ;
- mm3 += round value
- psraw xmm3, 3 ;
- mm3 /= 8
-
- pand xmm1, xmm7 ;
- mm1 select vals > thresh from source
- pandn xmm7, xmm3 ;
- mm7 select vals < thresh from blurred result
- paddusw xmm1, xmm7 ;
- combination
-
- packuswb xmm1, xmm0 ;
- pack to bytes
- movq QWORD PTR [edi+edx-8], mm0 ;
- store previous four bytes
- movdq2q mm0, xmm1
-
- add edx, 8
- cmp edx, cols
- jl acrossnextcol;
-
- // last 8 pixels
- movq QWORD PTR [edi+edx-8], mm0
-
- // done with this rwo
- add esi, eax ;
- next line
- mov eax, dst_pixels_per_line ;
- destination pitch?
- add edi, eax ;
- next destination
- mov eax, src_pixels_per_line ;
- destination pitch?
-
- dec ecx ;
- decrement count
- jnz nextrow ;
- next row
- }
-}
-
-
-void vp8_mbpost_proc_down_mmx(unsigned char *dst, int pitch, int rows, int cols, int flimit)
-{
- int c, i;
- __declspec(align(16))
- int flimit2[2];
- __declspec(align(16))
- unsigned char d[16][8];
-
- flimit = vp8_q2mbl(flimit);
-
- for (i = 0; i < 2; i++)
- flimit2[i] = flimit;
-
- rows += 8;
-
- for (c = 0; c < cols; c += 4)
- {
- unsigned char *s = &dst[c];
-
- __asm
- {
- mov esi, s ;
- pxor mm0, mm0 ;
-
- mov eax, pitch ;
- neg eax // eax = -pitch
-
- lea esi, [esi + eax*8]; // edi = s[-pitch*8]
- neg eax
-
-
- pxor mm5, mm5
- pxor mm6, mm6 ;
-
- pxor mm7, mm7 ;
- mov edi, esi
-
- mov ecx, 15 ;
-
- loop_initvar:
- movd mm1, DWORD PTR [edi];
- punpcklbw mm1, mm0 ;
-
- paddw mm5, mm1 ;
- pmullw mm1, mm1 ;
-
- movq mm2, mm1 ;
- punpcklwd mm1, mm0 ;
-
- punpckhwd mm2, mm0 ;
- paddd mm6, mm1 ;
-
- paddd mm7, mm2 ;
- lea edi, [edi+eax] ;
-
- dec ecx
- jne loop_initvar
- //save the var and sum
- xor edx, edx
- loop_row:
- movd mm1, DWORD PTR [esi] // [s-pitch*8]
- movd mm2, DWORD PTR [edi] // [s+pitch*7]
-
- punpcklbw mm1, mm0
- punpcklbw mm2, mm0
-
- paddw mm5, mm2
- psubw mm5, mm1
-
- pmullw mm2, mm2
- movq mm4, mm2
-
- punpcklwd mm2, mm0
- punpckhwd mm4, mm0
-
- paddd mm6, mm2
- paddd mm7, mm4
-
- pmullw mm1, mm1
- movq mm2, mm1
-
- punpcklwd mm1, mm0
- psubd mm6, mm1
-
- punpckhwd mm2, mm0
- psubd mm7, mm2
-
-
- movq mm3, mm6
- pslld mm3, 4
-
- psubd mm3, mm6
- movq mm1, mm5
-
- movq mm4, mm5
- pmullw mm1, mm1
-
- pmulhw mm4, mm4
- movq mm2, mm1
-
- punpcklwd mm1, mm4
- punpckhwd mm2, mm4
-
- movq mm4, mm7
- pslld mm4, 4
-
- psubd mm4, mm7
-
- psubd mm3, mm1
- psubd mm4, mm2
-
- psubd mm3, flimit2
- psubd mm4, flimit2
-
- psrad mm3, 31
- psrad mm4, 31
-
- packssdw mm3, mm4
- packsswb mm3, mm0
-
- movd mm1, DWORD PTR [esi+eax*8]
-
- movq mm2, mm1
- punpcklbw mm1, mm0
-
- paddw mm1, mm5
- mov ecx, edx
-
- and ecx, 127
- movq mm4, vp8_rv[ecx*2]
-
- paddw mm1, mm4
- //paddw xmm1, eight8s
- psraw mm1, 4
-
- packuswb mm1, mm0
- pand mm1, mm3
-
- pandn mm3, mm2
- por mm1, mm3
-
- and ecx, 15
- movd DWORD PTR d[ecx*4], mm1
-
- mov ecx, edx
- sub ecx, 8
-
- and ecx, 15
- movd mm1, DWORD PTR d[ecx*4]
-
- movd [esi], mm1
- lea esi, [esi+eax]
-
- lea edi, [edi+eax]
- add edx, 1
-
- cmp edx, rows
- jl loop_row
-
- }
-
- }
-}
-
-void vp8_mbpost_proc_down_xmm(unsigned char *dst, int pitch, int rows, int cols, int flimit)
-{
- int c, i;
- __declspec(align(16))
- int flimit4[4];
- __declspec(align(16))
- unsigned char d[16][8];
-
- flimit = vp8_q2mbl(flimit);
-
- for (i = 0; i < 4; i++)
- flimit4[i] = flimit;
-
- rows += 8;
-
- for (c = 0; c < cols; c += 8)
- {
- unsigned char *s = &dst[c];
-
- __asm
- {
- mov esi, s ;
- pxor xmm0, xmm0 ;
-
- mov eax, pitch ;
- neg eax // eax = -pitch
-
- lea esi, [esi + eax*8]; // edi = s[-pitch*8]
- neg eax
-
-
- pxor xmm5, xmm5
- pxor xmm6, xmm6 ;
-
- pxor xmm7, xmm7 ;
- mov edi, esi
-
- mov ecx, 15 ;
-
- loop_initvar:
- movq xmm1, QWORD PTR [edi];
- punpcklbw xmm1, xmm0 ;
-
- paddw xmm5, xmm1 ;
- pmullw xmm1, xmm1 ;
-
- movdqa xmm2, xmm1 ;
- punpcklwd xmm1, xmm0 ;
-
- punpckhwd xmm2, xmm0 ;
- paddd xmm6, xmm1 ;
-
- paddd xmm7, xmm2 ;
- lea edi, [edi+eax] ;
-
- dec ecx
- jne loop_initvar
- //save the var and sum
- xor edx, edx
- loop_row:
- movq xmm1, QWORD PTR [esi] // [s-pitch*8]
- movq xmm2, QWORD PTR [edi] // [s+pitch*7]
-
- punpcklbw xmm1, xmm0
- punpcklbw xmm2, xmm0
-
- paddw xmm5, xmm2
- psubw xmm5, xmm1
-
- pmullw xmm2, xmm2
- movdqa xmm4, xmm2
-
- punpcklwd xmm2, xmm0
- punpckhwd xmm4, xmm0
-
- paddd xmm6, xmm2
- paddd xmm7, xmm4
-
- pmullw xmm1, xmm1
- movdqa xmm2, xmm1
-
- punpcklwd xmm1, xmm0
- psubd xmm6, xmm1
-
- punpckhwd xmm2, xmm0
- psubd xmm7, xmm2
-
-
- movdqa xmm3, xmm6
- pslld xmm3, 4
-
- psubd xmm3, xmm6
- movdqa xmm1, xmm5
-
- movdqa xmm4, xmm5
- pmullw xmm1, xmm1
-
- pmulhw xmm4, xmm4
- movdqa xmm2, xmm1
-
- punpcklwd xmm1, xmm4
- punpckhwd xmm2, xmm4
-
- movdqa xmm4, xmm7
- pslld xmm4, 4
-
- psubd xmm4, xmm7
-
- psubd xmm3, xmm1
- psubd xmm4, xmm2
-
- psubd xmm3, flimit4
- psubd xmm4, flimit4
-
- psrad xmm3, 31
- psrad xmm4, 31
-
- packssdw xmm3, xmm4
- packsswb xmm3, xmm0
-
- movq xmm1, QWORD PTR [esi+eax*8]
-
- movq xmm2, xmm1
- punpcklbw xmm1, xmm0
-
- paddw xmm1, xmm5
- mov ecx, edx
-
- and ecx, 127
- movdqu xmm4, vp8_rv[ecx*2]
-
- paddw xmm1, xmm4
- //paddw xmm1, eight8s
- psraw xmm1, 4
-
- packuswb xmm1, xmm0
- pand xmm1, xmm3
-
- pandn xmm3, xmm2
- por xmm1, xmm3
-
- and ecx, 15
- movq QWORD PTR d[ecx*8], xmm1
-
- mov ecx, edx
- sub ecx, 8
-
- and ecx, 15
- movq mm0, d[ecx*8]
-
- movq [esi], mm0
- lea esi, [esi+eax]
-
- lea edi, [edi+eax]
- add edx, 1
-
- cmp edx, rows
- jl loop_row
-
- }
-
- }
-}
-#if 0
-/****************************************************************************
- *
- * ROUTINE : plane_add_noise_wmt
- *
- * INPUTS : unsigned char *Start starting address of buffer to add gaussian
- * noise to
- * unsigned int Width width of plane
- * unsigned int Height height of plane
- * int Pitch distance between subsequent lines of frame
- * int q quantizer used to determine amount of noise
- * to add
- *
- * OUTPUTS : None.
- *
- * RETURNS : void.
- *
- * FUNCTION : adds gaussian noise to a plane of pixels
- *
- * SPECIAL NOTES : None.
- *
- ****************************************************************************/
-void vp8_plane_add_noise_wmt(unsigned char *Start, unsigned int Width, unsigned int Height, int Pitch, int q, int a)
-{
- unsigned int i;
-
- __declspec(align(16)) unsigned char blackclamp[16];
- __declspec(align(16)) unsigned char whiteclamp[16];
- __declspec(align(16)) unsigned char bothclamp[16];
- char char_dist[300];
- char Rand[2048];
- double sigma;
-// return;
- __asm emms
- sigma = a + .5 + .6 * (63 - q) / 63.0;
-
- // set up a lookup table of 256 entries that matches
- // a gaussian distribution with sigma determined by q.
- //
- {
- double i;
- int next, j;
-
- next = 0;
-
- for (i = -32; i < 32; i++)
- {
- double g = 256 * vp8_gaussian(sigma, 0, 1.0 * i);
- int a = (int)(g + .5);
-
- if (a)
- {
- for (j = 0; j < a; j++)
- {
- char_dist[next+j] = (char) i;
- }
-
- next = next + j;
- }
-
- }
-
- for (next = next; next < 256; next++)
- char_dist[next] = 0;
-
- }
-
- for (i = 0; i < 2048; i++)
- {
- Rand[i] = char_dist[rand() & 0xff];
- }
-
- for (i = 0; i < 16; i++)
- {
- blackclamp[i] = -char_dist[0];
- whiteclamp[i] = -char_dist[0];
- bothclamp[i] = -2 * char_dist[0];
- }
-
- for (i = 0; i < Height; i++)
- {
- unsigned char *Pos = Start + i * Pitch;
- char *Ref = Rand + (rand() & 0xff);
-
- __asm
- {
- mov ecx, [Width]
- mov esi, Pos
- mov edi, Ref
- xor eax, eax
-
- nextset:
- movdqu xmm1, [esi+eax] // get the source
-
- psubusb xmm1, blackclamp // clamp both sides so we don't outrange adding noise
- paddusb xmm1, bothclamp
- psubusb xmm1, whiteclamp
-
- movdqu xmm2, [edi+eax] // get the noise for this line
- paddb xmm1, xmm2 // add it in
- movdqu [esi+eax], xmm1 // store the result
-
- add eax, 16 // move to the next line
-
- cmp eax, ecx
- jl nextset
-
-
- }
-
- }
-}
-#endif
-__declspec(align(16))
-static const int four8s[4] = { 8, 8, 8, 8};
-void vp8_mbpost_proc_across_ip_xmm(unsigned char *src, int pitch, int rows, int cols, int flimit)
-{
- int r, i;
- __declspec(align(16))
- int flimit4[4];
- unsigned char *s = src;
- int sumsq;
- int sum;
-
-
- flimit = vp8_q2mbl(flimit);
- flimit4[0] =
- flimit4[1] =
- flimit4[2] =
- flimit4[3] = flimit;
-
- for (r = 0; r < rows; r++)
- {
-
-
- sumsq = 0;
- sum = 0;
-
- for (i = -8; i <= 6; i++)
- {
- sumsq += s[i] * s[i];
- sum += s[i];
- }
-
- __asm
- {
- mov eax, sumsq
- movd xmm7, eax
-
- mov eax, sum
- movd xmm6, eax
-
- mov esi, s
- xor ecx, ecx
-
- mov edx, cols
- add edx, 8
- pxor mm0, mm0
- pxor mm1, mm1
-
- pxor xmm0, xmm0
- nextcol4:
-
- movd xmm1, DWORD PTR [esi+ecx-8] // -8 -7 -6 -5
- movd xmm2, DWORD PTR [esi+ecx+7] // +7 +8 +9 +10
-
- punpcklbw xmm1, xmm0 // expanding
- punpcklbw xmm2, xmm0 // expanding
-
- punpcklwd xmm1, xmm0 // expanding to dwords
- punpcklwd xmm2, xmm0 // expanding to dwords
-
- psubd xmm2, xmm1 // 7--8 8--7 9--6 10--5
- paddd xmm1, xmm1 // -8*2 -7*2 -6*2 -5*2
-
- paddd xmm1, xmm2 // 7+-8 8+-7 9+-6 10+-5
- pmaddwd xmm1, xmm2 // squared of 7+-8 8+-7 9+-6 10+-5
-
- paddd xmm6, xmm2
- paddd xmm7, xmm1
-
- pshufd xmm6, xmm6, 0 // duplicate the last ones
- pshufd xmm7, xmm7, 0 // duplicate the last ones
-
- psrldq xmm1, 4 // 8--7 9--6 10--5 0000
- psrldq xmm2, 4 // 8--7 9--6 10--5 0000
-
- pshufd xmm3, xmm1, 3 // 0000 8--7 8--7 8--7 squared
- pshufd xmm4, xmm2, 3 // 0000 8--7 8--7 8--7 squared
-
- paddd xmm6, xmm4
- paddd xmm7, xmm3
-
- pshufd xmm3, xmm1, 01011111b // 0000 0000 9--6 9--6 squared
- pshufd xmm4, xmm2, 01011111b // 0000 0000 9--6 9--6 squared
-
- paddd xmm7, xmm3
- paddd xmm6, xmm4
-
- pshufd xmm3, xmm1, 10111111b // 0000 0000 8--7 8--7 squared
- pshufd xmm4, xmm2, 10111111b // 0000 0000 8--7 8--7 squared
-
- paddd xmm7, xmm3
- paddd xmm6, xmm4
-
- movdqa xmm3, xmm6
- pmaddwd xmm3, xmm3
-
- movdqa xmm5, xmm7
- pslld xmm5, 4
-
- psubd xmm5, xmm7
- psubd xmm5, xmm3
-
- psubd xmm5, flimit4
- psrad xmm5, 31
-
- packssdw xmm5, xmm0
- packsswb xmm5, xmm0
-
- movd xmm1, DWORD PTR [esi+ecx]
- movq xmm2, xmm1
-
- punpcklbw xmm1, xmm0
- punpcklwd xmm1, xmm0
-
- paddd xmm1, xmm6
- paddd xmm1, four8s
-
- psrad xmm1, 4
- packssdw xmm1, xmm0
-
- packuswb xmm1, xmm0
- pand xmm1, xmm5
-
- pandn xmm5, xmm2
- por xmm5, xmm1
-
- movd [esi+ecx-8], mm0
- movq mm0, mm1
-
- movdq2q mm1, xmm5
- psrldq xmm7, 12
-
- psrldq xmm6, 12
- add ecx, 4
-
- cmp ecx, edx
- jl nextcol4
-
- }
- s += pitch;
- }
-}
-
-#if 0
-
-/****************************************************************************
- *
- * ROUTINE : plane_add_noise_mmx
- *
- * INPUTS : unsigned char *Start starting address of buffer to add gaussian
- * noise to
- * unsigned int Width width of plane
- * unsigned int Height height of plane
- * int Pitch distance between subsequent lines of frame
- * int q quantizer used to determine amount of noise
- * to add
- *
- * OUTPUTS : None.
- *
- * RETURNS : void.
- *
- * FUNCTION : adds gaussian noise to a plane of pixels
- *
- * SPECIAL NOTES : None.
- *
- ****************************************************************************/
-void vp8_plane_add_noise_mmx(unsigned char *Start, unsigned int Width, unsigned int Height, int Pitch, int q, int a)
-{
- unsigned int i;
- int Pitch4 = Pitch * 4;
- const int noise_amount = 2;
- const int noise_adder = 2 * noise_amount + 1;
-
- __declspec(align(16)) unsigned char blackclamp[16];
- __declspec(align(16)) unsigned char whiteclamp[16];
- __declspec(align(16)) unsigned char bothclamp[16];
-
- char char_dist[300];
- char Rand[2048];
-
- double sigma;
- __asm emms
- sigma = a + .5 + .6 * (63 - q) / 63.0;
-
- // set up a lookup table of 256 entries that matches
- // a gaussian distribution with sigma determined by q.
- //
- {
- double i, sum = 0;
- int next, j;
-
- next = 0;
-
- for (i = -32; i < 32; i++)
- {
- int a = (int)(.5 + 256 * vp8_gaussian(sigma, 0, i));
-
- if (a)
- {
- for (j = 0; j < a; j++)
- {
- char_dist[next+j] = (char) i;
- }
-
- next = next + j;
- }
-
- }
-
- for (next = next; next < 256; next++)
- char_dist[next] = 0;
-
- }
-
- for (i = 0; i < 2048; i++)
- {
- Rand[i] = char_dist[rand() & 0xff];
- }
-
- for (i = 0; i < 16; i++)
- {
- blackclamp[i] = -char_dist[0];
- whiteclamp[i] = -char_dist[0];
- bothclamp[i] = -2 * char_dist[0];
- }
-
- for (i = 0; i < Height; i++)
- {
- unsigned char *Pos = Start + i * Pitch;
- char *Ref = Rand + (rand() & 0xff);
-
- __asm
- {
- mov ecx, [Width]
- mov esi, Pos
- mov edi, Ref
- xor eax, eax
-
- nextset:
- movq mm1, [esi+eax] // get the source
-
- psubusb mm1, blackclamp // clamp both sides so we don't outrange adding noise
- paddusb mm1, bothclamp
- psubusb mm1, whiteclamp
-
- movq mm2, [edi+eax] // get the noise for this line
- paddb mm1, mm2 // add it in
- movq [esi+eax], mm1 // store the result
-
- add eax, 8 // move to the next line
-
- cmp eax, ecx
- jl nextset
-
-
- }
-
- }
-}
-#else
-extern char an[8][64][3072];
-extern int cd[8][64];
-
-void vp8_plane_add_noise_mmx(unsigned char *Start, unsigned int Width, unsigned int Height, int Pitch, int q, int a)
-{
- unsigned int i;
- __declspec(align(16)) unsigned char blackclamp[16];
- __declspec(align(16)) unsigned char whiteclamp[16];
- __declspec(align(16)) unsigned char bothclamp[16];
-
-
- __asm emms
-
- for (i = 0; i < 16; i++)
- {
- blackclamp[i] = -cd[a][q];
- whiteclamp[i] = -cd[a][q];
- bothclamp[i] = -2 * cd[a][q];
- }
-
- for (i = 0; i < Height; i++)
- {
- unsigned char *Pos = Start + i * Pitch;
- char *Ref = an[a][q] + (rand() & 0xff);
-
- __asm
- {
- mov ecx, [Width]
- mov esi, Pos
- mov edi, Ref
- xor eax, eax
-
- nextset:
- movq mm1, [esi+eax] // get the source
-
- psubusb mm1, blackclamp // clamp both sides so we don't outrange adding noise
- paddusb mm1, bothclamp
- psubusb mm1, whiteclamp
-
- movq mm2, [edi+eax] // get the noise for this line
- paddb mm1, mm2 // add it in
- movq [esi+eax], mm1 // store the result
-
- add eax, 8 // move to the next line
-
- cmp eax, ecx
- jl nextset
- }
- }
-}
-
-
-void vp8_plane_add_noise_wmt(unsigned char *Start, unsigned int Width, unsigned int Height, int Pitch, int q, int a)
-{
- unsigned int i;
-
- __declspec(align(16)) unsigned char blackclamp[16];
- __declspec(align(16)) unsigned char whiteclamp[16];
- __declspec(align(16)) unsigned char bothclamp[16];
-
- __asm emms
-
- for (i = 0; i < 16; i++)
- {
- blackclamp[i] = -cd[a][q];
- whiteclamp[i] = -cd[a][q];
- bothclamp[i] = -2 * cd[a][q];
- }
-
- for (i = 0; i < Height; i++)
- {
- unsigned char *Pos = Start + i * Pitch;
- char *Ref = an[a][q] + (rand() & 0xff);
-
- __asm
- {
- mov ecx, [Width]
- mov esi, Pos
- mov edi, Ref
- xor eax, eax
-
- nextset:
- movdqu xmm1, [esi+eax] // get the source
-
- psubusb xmm1, blackclamp // clamp both sides so we don't outrange adding noise
- paddusb xmm1, bothclamp
- psubusb xmm1, whiteclamp
-
- movdqu xmm2, [edi+eax] // get the noise for this line
- paddb xmm1, xmm2 // add it in
- movdqu [esi+eax], xmm1 // store the result
-
- add eax, 16 // move to the next line
-
- cmp eax, ecx
- jl nextset
- }
- }
-}
-
-#endif
diff --git a/vp8/decoder/x86/idct_blk_sse2.c b/vp8/decoder/x86/idct_blk_sse2.c
index 4c88db474b75a9843f681e09dfcdf5b5af04a48a..3a48068622af43ea99c6f5fd0f368d4462c4a9ba 100644
--- a/vp8/decoder/x86/idct_blk_sse2.c
+++ b/vp8/decoder/x86/idct_blk_sse2.c
@@ -12,17 +12,17 @@
#include "vp8/common/idct.h"
#include "vp8/decoder/dequantize.h"
-void idct_dequant_dc_0_2x_sse2
+void vp8_idct_dequant_dc_0_2x_sse2
(short *q, short *dq, unsigned char *pre,
unsigned char *dst, int dst_stride, short *dc);
-void idct_dequant_dc_full_2x_sse2
+void vp8_idct_dequant_dc_full_2x_sse2
(short *q, short *dq, unsigned char *pre,
unsigned char *dst, int dst_stride, short *dc);
-void idct_dequant_0_2x_sse2
+void vp8_idct_dequant_0_2x_sse2
(short *q, short *dq ,unsigned char *pre,
unsigned char *dst, int dst_stride, int blk_stride);
-void idct_dequant_full_2x_sse2
+void vp8_idct_dequant_full_2x_sse2
(short *q, short *dq ,unsigned char *pre,
unsigned char *dst, int dst_stride, int blk_stride);
@@ -35,14 +35,14 @@ void vp8_dequant_dc_idct_add_y_block_sse2
for (i = 0; i < 4; i++)
{
if (((short *)(eobs))[0] & 0xfefe)
- idct_dequant_dc_full_2x_sse2 (q, dq, pre, dst, stride, dc);
+ vp8_idct_dequant_dc_full_2x_sse2 (q, dq, pre, dst, stride, dc);
else
- idct_dequant_dc_0_2x_sse2 (q, dq, pre, dst, stride, dc);
+ vp8_idct_dequant_dc_0_2x_sse2 (q, dq, pre, dst, stride, dc);
if (((short *)(eobs))[1] & 0xfefe)
- idct_dequant_dc_full_2x_sse2 (q+32, dq, pre+8, dst+8, stride, dc+2);
+ vp8_idct_dequant_dc_full_2x_sse2 (q+32, dq, pre+8, dst+8, stride, dc+2);
else
- idct_dequant_dc_0_2x_sse2 (q+32, dq, pre+8, dst+8, stride, dc+2);
+ vp8_idct_dequant_dc_0_2x_sse2 (q+32, dq, pre+8, dst+8, stride, dc+2);
q += 64;
dc += 4;
@@ -61,14 +61,14 @@ void vp8_dequant_idct_add_y_block_sse2
for (i = 0; i < 4; i++)
{
if (((short *)(eobs))[0] & 0xfefe)
- idct_dequant_full_2x_sse2 (q, dq, pre, dst, stride, 16);
+ vp8_idct_dequant_full_2x_sse2 (q, dq, pre, dst, stride, 16);
else
- idct_dequant_0_2x_sse2 (q, dq, pre, dst, stride, 16);
+ vp8_idct_dequant_0_2x_sse2 (q, dq, pre, dst, stride, 16);
if (((short *)(eobs))[1] & 0xfefe)
- idct_dequant_full_2x_sse2 (q+32, dq, pre+8, dst+8, stride, 16);
+ vp8_idct_dequant_full_2x_sse2 (q+32, dq, pre+8, dst+8, stride, 16);
else
- idct_dequant_0_2x_sse2 (q+32, dq, pre+8, dst+8, stride, 16);
+ vp8_idct_dequant_0_2x_sse2 (q+32, dq, pre+8, dst+8, stride, 16);
q += 64;
pre += 64;
@@ -82,33 +82,33 @@ void vp8_dequant_idct_add_uv_block_sse2
unsigned char *dstu, unsigned char *dstv, int stride, char *eobs)
{
if (((short *)(eobs))[0] & 0xfefe)
- idct_dequant_full_2x_sse2 (q, dq, pre, dstu, stride, 8);
+ vp8_idct_dequant_full_2x_sse2 (q, dq, pre, dstu, stride, 8);
else
- idct_dequant_0_2x_sse2 (q, dq, pre, dstu, stride, 8);
+ vp8_idct_dequant_0_2x_sse2 (q, dq, pre, dstu, stride, 8);
q += 32;
pre += 32;
dstu += stride*4;
if (((short *)(eobs))[1] & 0xfefe)
- idct_dequant_full_2x_sse2 (q, dq, pre, dstu, stride, 8);
+ vp8_idct_dequant_full_2x_sse2 (q, dq, pre, dstu, stride, 8);
else
- idct_dequant_0_2x_sse2 (q, dq, pre, dstu, stride, 8);
+ vp8_idct_dequant_0_2x_sse2 (q, dq, pre, dstu, stride, 8);
q += 32;
pre += 32;
if (((short *)(eobs))[2] & 0xfefe)
- idct_dequant_full_2x_sse2 (q, dq, pre, dstv, stride, 8);
+ vp8_idct_dequant_full_2x_sse2 (q, dq, pre, dstv, stride, 8);
else
- idct_dequant_0_2x_sse2 (q, dq, pre, dstv, stride, 8);
+ vp8_idct_dequant_0_2x_sse2 (q, dq, pre, dstv, stride, 8);
q += 32;
pre += 32;
dstv += stride*4;
if (((short *)(eobs))[3] & 0xfefe)
- idct_dequant_full_2x_sse2 (q, dq, pre, dstv, stride, 8);
+ vp8_idct_dequant_full_2x_sse2 (q, dq, pre, dstv, stride, 8);
else
- idct_dequant_0_2x_sse2 (q, dq, pre, dstv, stride, 8);
+ vp8_idct_dequant_0_2x_sse2 (q, dq, pre, dstv, stride, 8);
}
diff --git a/vp8/encoder/bitstream.c b/vp8/encoder/bitstream.c
index 64d1c93042c18734afcf4df04ad28868925bc433..e65a7f9d4dda64184400004940d8b2a1fb5a6f96 100644
--- a/vp8/encoder/bitstream.c
+++ b/vp8/encoder/bitstream.c
@@ -22,7 +22,8 @@
#include "vpx/vpx_encoder.h"
#include "vpx_mem/vpx_mem.h"
#include "bitstream.h"
-#include "vp8/common/defaultcoefcounts.h"
+
+#include "defaultcoefcounts.h"
#if CONFIG_SEGMENTATION
static int segment_cost = 0;
#endif
@@ -1295,7 +1296,7 @@ static int independent_coef_context_savings(VP8_COMP *cpi)
if (cpi->common.frame_type == KEY_FRAME)
{
/* Reset to default probabilities at key frames */
- sum_probs_over_prev_coef_context(vp8_default_coef_counts[i][j],
+ sum_probs_over_prev_coef_context(default_coef_counts[i][j],
prev_coef_count_sum);
}
else
diff --git a/vp8/common/defaultcoefcounts.c b/vp8/encoder/defaultcoefcounts.h
similarity index 98%
rename from vp8/common/defaultcoefcounts.c
rename to vp8/encoder/defaultcoefcounts.h
index 34d1fb1d510b16dd8c558f5e19cf3d90343f153d..3b54c823cb29970765e1168c832fa2a0485c048f 100644
--- a/vp8/common/defaultcoefcounts.c
+++ b/vp8/encoder/defaultcoefcounts.h
@@ -8,14 +8,12 @@
* be found in the AUTHORS file in the root of the source tree.
*/
-#include "defaultcoefcounts.h"
-
/* Generated file, included by entropy.c */
-const unsigned int vp8_default_coef_counts[BLOCK_TYPES]
- [COEF_BANDS]
- [PREV_COEF_CONTEXTS]
- [MAX_ENTROPY_TOKENS] =
+static const unsigned int default_coef_counts[BLOCK_TYPES]
+ [COEF_BANDS]
+ [PREV_COEF_CONTEXTS]
+ [MAX_ENTROPY_TOKENS] =
{
{
@@ -401,4 +399,4 @@ const unsigned int vp8_default_coef_counts_8x8[BLOCK_TYPES]
}
}
};
-#endif
\ No newline at end of file
+#endif
diff --git a/vp8/vp8_common.mk b/vp8/vp8_common.mk
index 1a883b8ab2eaf8e2cf90da7556234427921bd134..b6e5a41c2b1cb3b3ed9f275b88a7ffd739167aa6 100644
--- a/vp8/vp8_common.mk
+++ b/vp8/vp8_common.mk
@@ -19,8 +19,6 @@ VP8_COMMON_SRCS-yes += common/asm_com_offsets.c
VP8_COMMON_SRCS-yes += common/blockd.c
VP8_COMMON_SRCS-yes += common/coefupdateprobs.h
VP8_COMMON_SRCS-yes += common/debugmodes.c
-VP8_COMMON_SRCS-yes += common/defaultcoefcounts.h
-VP8_COMMON_SRCS-yes += common/defaultcoefcounts.c
VP8_COMMON_SRCS-yes += common/entropy.c
VP8_COMMON_SRCS-yes += common/entropymode.c
VP8_COMMON_SRCS-yes += common/entropymv.c
diff --git a/vp8/vp8cx.mk b/vp8/vp8cx.mk
index d46d99df6626188f2ff2956d52a8f48232db6915..b71a54aea938d906839e1aef1b99aa50c6366a07 100644
--- a/vp8/vp8cx.mk
+++ b/vp8/vp8cx.mk
@@ -34,6 +34,7 @@ VP8_CX_SRCS-yes += vp8_cx_iface.c
#INCLUDES += encoder
VP8_CX_SRCS-yes += encoder/asm_enc_offsets.c
+VP8_CX_SRCS-yes += encoder/defaultcoefcounts.h
VP8_CX_SRCS-yes += encoder/bitstream.c
VP8_CX_SRCS-yes += encoder/boolhuff.c
VP8_CX_SRCS-yes += encoder/dct.c