Skip to content
GitLab
Explore
Sign in
Register
Primary navigation
Search or go to…
Project
Opus
Manage
Activity
Members
Labels
Plan
Issues
Issue boards
Milestones
Wiki
Code
Merge requests
Repository
Branches
Commits
Tags
Repository graph
Compare revisions
Build
Pipelines
Jobs
Pipeline schedules
Artifacts
Deploy
Releases
Container Registry
Model registry
Operate
Environments
Monitor
Incidents
Analyze
Value stream analytics
Contributor analytics
CI/CD analytics
Repository analytics
Model experiments
Help
Help
Support
GitLab documentation
Compare GitLab plans
Community forum
Contribute to GitLab
Provide feedback
Terms and privacy
Keyboard shortcuts
?
Snippets
Groups
Projects
Show more breadcrumbs
Xiph.Org
Opus
Commits
825ead82
Commit
825ead82
authored
16 years ago
by
Jean-Marc Valin
Browse files
Options
Downloads
Patches
Plain Diff
Simplified allocation by doing it one band at a time, with memory
parent
d9b9565a
No related branches found
Branches containing commit
No related tags found
Tags containing commit
No related merge requests found
Changes
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
libcelt/rate.c
+49
-47
49 additions, 47 deletions
libcelt/rate.c
with
49 additions
and
47 deletions
libcelt/rate.c
+
49
−
47
View file @
825ead82
...
...
@@ -125,41 +125,38 @@ static int vec_bits2pulses(const CELTMode *m, const celt_int16_t * const *cache,
for
(
i
=
0
;
i
<
len
;
i
++
)
{
pulses
[
i
]
=
bits2pulses
(
m
,
cache
[
i
],
bits
[
i
]);
sum
+=
cache
[
i
][
pulses
[
i
]];
sum
+=
bits
[
i
];
}
/*printf ("sum = %d\n", sum);*/
return
sum
;
}
static
int
interp_bits2pulses
(
const
CELTMode
*
m
,
const
celt_int16_t
*
const
*
cache
,
int
*
bits1
,
int
*
bits2
,
int
*
ebits1
,
int
*
ebits2
,
int
total
,
int
*
pulses
,
int
*
ebits
,
int
len
)
static
int
interp_bits2pulses
(
const
CELTMode
*
m
,
const
celt_int16_t
*
const
*
cache
,
int
*
bits1
,
int
*
bits2
,
int
*
ebits1
,
int
*
ebits2
,
int
total
,
int
*
pulses
,
int
*
bits
,
int
*
ebits
,
int
len
)
{
int
esum
;
int
esum
,
psum
;
int
lo
,
hi
,
out
;
int
j
;
VARDECL
(
int
,
bits
);
const
int
C
=
CHANNELS
(
m
);
SAVE_STACK
;
ALLOC
(
bits
,
len
,
int
);
lo
=
0
;
hi
=
1
<<
BITRES
;
while
(
hi
-
lo
!=
1
)
{
int
mid
=
(
lo
+
hi
)
>>
1
;
psum
=
0
;
esum
=
0
;
for
(
j
=
0
;
j
<
len
;
j
++
)
{
e
bits
[
j
]
=
(((
1
<<
BITRES
)
-
mid
)
*
ebits1
[
j
]
+
mid
*
ebits2
[
j
]
+
(
1
<<
(
BITRES
-
1
)))
>>
BITRES
;
e
sum
+=
ebits
[
j
];
e
sum
+
=
(((
1
<<
BITRES
)
-
mid
)
*
ebits1
[
j
]
+
mid
*
ebits2
[
j
]
+
(
1
<<
(
BITRES
-
1
)))
>>
BITRES
;
p
sum
+=
((
1
<<
BITRES
)
-
mid
)
*
bits1
[
j
]
+
mid
*
bits2
[
j
];
}
for
(
j
=
0
;
j
<
len
;
j
++
)
bits
[
j
]
=
((
1
<<
BITRES
)
-
mid
)
*
bits1
[
j
]
+
mid
*
bits2
[
j
];
if
(
vec_bits2pulses
(
m
,
cache
,
bits
,
pulses
,
len
)
>
(
total
-
C
*
esum
)
<<
BITRES
)
if
(
psum
>
(
total
-
C
*
esum
)
<<
BITRES
)
hi
=
mid
;
else
lo
=
mid
;
}
esum
=
0
;
psum
=
0
;
/*printf ("interp bisection gave %d\n", lo);*/
for
(
j
=
0
;
j
<
len
;
j
++
)
{
...
...
@@ -167,40 +164,12 @@ static int interp_bits2pulses(const CELTMode *m, const celt_int16_t * const *cac
esum
+=
ebits
[
j
];
}
for
(
j
=
0
;
j
<
len
;
j
++
)
bits
[
j
]
=
((
1
<<
BITRES
)
-
lo
)
*
bits1
[
j
]
+
lo
*
bits2
[
j
];
out
=
vec_bits2pulses
(
m
,
cache
,
bits
,
pulses
,
len
);
/*printf ("left to allocate: %d\n", total-C*esum-(out>>BITRES));*/
/* Do some refinement to use up all bits. In the first pass, we can only add pulses to
bands that are under their allocated budget. In the second pass, anything goes */
for
(
j
=
0
;
j
<
len
;
j
++
)
{
if
(
cache
[
j
][
pulses
[
j
]]
<
bits
[
j
]
&&
pulses
[
j
]
<
MAX_PULSES
-
1
)
{
if
(
out
+
cache
[
j
][
pulses
[
j
]
+
1
]
-
cache
[
j
][
pulses
[
j
]]
<=
(
total
-
C
*
esum
)
<<
BITRES
)
{
out
=
out
+
cache
[
j
][
pulses
[
j
]
+
1
]
-
cache
[
j
][
pulses
[
j
]];
pulses
[
j
]
+=
1
;
}
}
}
while
(
1
)
{
int
incremented
=
0
;
for
(
j
=
0
;
j
<
len
;
j
++
)
{
if
(
pulses
[
j
]
<
MAX_PULSES
-
1
)
{
if
(
out
+
cache
[
j
][
pulses
[
j
]
+
1
]
-
cache
[
j
][
pulses
[
j
]]
<=
(
total
-
C
*
esum
)
<<
BITRES
)
{
out
=
out
+
cache
[
j
][
pulses
[
j
]
+
1
]
-
cache
[
j
][
pulses
[
j
]];
pulses
[
j
]
+=
1
;
incremented
=
1
;
}
}
}
if
(
!
incremented
)
break
;
bits
[
j
]
=
((
1
<<
BITRES
)
-
lo
)
*
bits1
[
j
]
+
lo
*
bits2
[
j
];
psum
+=
bits
[
j
];
}
out
=
psum
;
/*printf ("left to allocate: %d\n", total-C*esum-(out>>BITRES));*/
RESTORE_STACK
;
return
(
out
+
BITROUND
)
>>
BITRES
;
}
...
...
@@ -208,6 +177,7 @@ static int interp_bits2pulses(const CELTMode *m, const celt_int16_t * const *cac
int
compute_allocation
(
const
CELTMode
*
m
,
int
*
offsets
,
const
int
*
stereo_mode
,
int
total
,
int
*
pulses
,
int
*
ebits
)
{
int
lo
,
hi
,
len
,
ret
,
i
;
VARDECL
(
int
,
bits
);
VARDECL
(
int
,
bits1
);
VARDECL
(
int
,
bits2
);
VARDECL
(
int
,
ebits1
);
...
...
@@ -217,6 +187,7 @@ int compute_allocation(const CELTMode *m, int *offsets, const int *stereo_mode,
SAVE_STACK
;
len
=
m
->
nbEBands
;
ALLOC
(
bits
,
len
,
int
);
ALLOC
(
bits1
,
len
,
int
);
ALLOC
(
bits2
,
len
,
int
);
ALLOC
(
ebits1
,
len
,
int
);
...
...
@@ -241,6 +212,7 @@ int compute_allocation(const CELTMode *m, int *offsets, const int *stereo_mode,
hi
=
m
->
nbAllocVectors
-
1
;
while
(
hi
-
lo
!=
1
)
{
int
psum
=
0
;
int
j
;
int
mid
=
(
lo
+
hi
)
>>
1
;
for
(
j
=
0
;
j
<
len
;
j
++
)
...
...
@@ -248,10 +220,11 @@ int compute_allocation(const CELTMode *m, int *offsets, const int *stereo_mode,
bits1
[
j
]
=
(
m
->
allocVectors
[
mid
*
len
+
j
]
+
offsets
[
j
])
<<
BITRES
;
if
(
bits1
[
j
]
<
0
)
bits1
[
j
]
=
0
;
psum
+=
bits1
[
j
];
/*printf ("%d ", bits[j]);*/
}
/*printf ("\n");*/
if
(
vec_bits2pulses
(
m
,
cache
,
bits1
,
pulses
,
len
)
>
(
total
-
C
*
m
->
energy_alloc
[
mid
*
(
len
+
1
)
+
len
])
<<
BITRES
)
if
(
psum
>
(
total
-
C
*
m
->
energy_alloc
[
mid
*
(
len
+
1
)
+
len
])
<<
BITRES
)
hi
=
mid
;
else
lo
=
mid
;
...
...
@@ -271,9 +244,38 @@ int compute_allocation(const CELTMode *m, int *offsets, const int *stereo_mode,
if
(
bits2
[
j
]
<
0
)
bits2
[
j
]
=
0
;
}
ret
=
interp_bits2pulses
(
m
,
cache
,
bits1
,
bits2
,
ebits1
,
ebits2
,
total
,
pulses
,
ebits
,
len
);
RESTORE_STACK
;
return
ret
;
ret
=
interp_bits2pulses
(
m
,
cache
,
bits1
,
bits2
,
ebits1
,
ebits2
,
total
,
pulses
,
bits
,
ebits
,
len
);
}
{
int
balance
=
0
;
for
(
i
=
0
;
i
<
len
;
i
++
)
{
int
P
,
curr_balance
;
curr_balance
=
(
len
-
i
);
if
(
curr_balance
>
3
)
curr_balance
=
3
;
curr_balance
=
balance
/
curr_balance
;
//balance -= curr_balance;
P
=
bits2pulses
(
m
,
cache
[
i
],
bits
[
i
]
+
curr_balance
);
balance
+=
bits
[
i
]
-
cache
[
i
][
P
];
pulses
[
i
]
=
P
;
//printf ("(%d %d) ", bits[i], cache[i][outBits]);
}
if
(
balance
<
0
)
{
for
(
i
=
len
-
1
;
i
>=
0
;
i
--
)
{
if
(
pulses
[
i
])
{
pulses
[
i
]
--
;
break
;
}
}
}
//printf ("\n");
}
RESTORE_STACK
;
return
ret
;
}
This diff is collapsed.
Click to expand it.
Preview
0%
Loading
Try again
or
attach a new file
.
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Save comment
Cancel
Please
register
or
sign in
to comment