opus.git
11 years agomore restricted pointers in celt.c
Jean-Marc Valin [Sat, 19 Apr 2008 12:24:53 +0000 (22:24 +1000)]
more restricted pointers in celt.c

11 years agoadded a shortcut (SHORTCUT) to skip the spreading function in the pitch search
Jean-Marc Valin [Sat, 19 Apr 2008 11:39:26 +0000 (21:39 +1000)]
added a shortcut (SHORTCUT) to skip the spreading function in the pitch search

11 years agoencoder pre-emphasis now in 16-bits
Jean-Marc Valin [Sat, 19 Apr 2008 09:44:58 +0000 (19:44 +1000)]
encoder pre-emphasis now in 16-bits

11 years agobetter indexing in exp_rotation()
Jean-Marc Valin [Sat, 19 Apr 2008 02:58:29 +0000 (12:58 +1000)]
better indexing in exp_rotation()

11 years agobetter mdct_backward() indexing
Jean-Marc Valin [Sat, 19 Apr 2008 02:53:43 +0000 (12:53 +1000)]
better mdct_backward() indexing

11 years agooptimisation: giving more hints to the compiler about the sizes in
Jean-Marc Valin [Fri, 18 Apr 2008 13:13:14 +0000 (23:13 +1000)]
optimisation: giving more hints to the compiler about the sizes in
find_spectral_pitch()

11 years agoa few minor optimisations
Jean-Marc Valin [Fri, 18 Apr 2008 12:19:58 +0000 (22:19 +1000)]
a few minor optimisations

11 years agoremoved unneeded variable shifts from alg_quant()
Jean-Marc Valin [Fri, 18 Apr 2008 11:00:06 +0000 (21:00 +1000)]
removed unneeded variable shifts from alg_quant()

11 years agoseparated the two passes from interp_bits2pulses()
Jean-Marc Valin [Fri, 18 Apr 2008 09:35:34 +0000 (19:35 +1000)]
separated the two passes from interp_bits2pulses()

11 years agooptimisation: removed the shifts from the vq_index() inner loop
Jean-Marc Valin [Fri, 18 Apr 2008 09:34:58 +0000 (19:34 +1000)]
optimisation: removed the shifts from the vq_index() inner loop

11 years agooptimisation: better indexing/looping in vq_index()
Jean-Marc Valin [Fri, 18 Apr 2008 07:54:52 +0000 (17:54 +1000)]
optimisation: better indexing/looping in vq_index()

11 years agopseudo-stack no longer checks on every function entry whether it has been
Jean-Marc Valin [Fri, 18 Apr 2008 07:29:56 +0000 (17:29 +1000)]
pseudo-stack no longer checks on every function entry whether it has been
allocated

11 years agooptimisation: one less conditional branch in pulse2comb()
Jean-Marc Valin [Fri, 18 Apr 2008 06:58:19 +0000 (16:58 +1000)]
optimisation: one less conditional branch in pulse2comb()

11 years agooptimisation: Removed a bunch of conditional branches from comb2pulse()
Jean-Marc Valin [Fri, 18 Apr 2008 06:46:39 +0000 (16:46 +1000)]
optimisation: Removed a bunch of conditional branches from comb2pulse()

11 years agoSaturation in SIG2INT16 using MIN/MAX
Jean-Marc Valin [Fri, 18 Apr 2008 05:57:18 +0000 (15:57 +1000)]
Saturation in SIG2INT16 using MIN/MAX

11 years agoRemoved implicit 32=>16 conversion (changed to EXTRACT16)
Jean-Marc Valin [Fri, 18 Apr 2008 00:59:03 +0000 (10:59 +1000)]
Removed implicit 32=>16 conversion (changed to EXTRACT16)

11 years agooptimisation: Got rid of the 32-bit mul in find_spectral_pitch()
Jean-Marc Valin [Fri, 18 Apr 2008 00:33:06 +0000 (10:33 +1000)]
optimisation: Got rid of the 32-bit mul in find_spectral_pitch()

11 years agominor simplification in alg_quant()
Jean-Marc Valin [Wed, 16 Apr 2008 21:50:12 +0000 (07:50 +1000)]
minor simplification in alg_quant()

11 years agofixed an issue (lacking parentheses) in the no-op version of BITREV
Jean-Marc Valin [Wed, 16 Apr 2008 21:41:39 +0000 (07:41 +1000)]
fixed an issue (lacking parentheses) in the no-op version of BITREV

11 years agoMade twiddle pointer in mdct more explicit
Jean-Marc Valin [Wed, 16 Apr 2008 21:29:53 +0000 (07:29 +1000)]
Made twiddle pointer in mdct more explicit

11 years agooptimisation: Making it clear to the compiler that many of the loops in cwrs
Jean-Marc Valin [Wed, 16 Apr 2008 04:15:07 +0000 (14:15 +1000)]
optimisation: Making it clear to the compiler that many of the loops in cwrs
need to iterate at least once.

11 years agoMore use of CELT_MEMSET
Jean-Marc Valin [Wed, 16 Apr 2008 01:58:48 +0000 (11:58 +1000)]
More use of CELT_MEMSET

11 years agoMake use of CELT_MEMSET() in find_spectral_pitch()
Jean-Marc Valin [Wed, 16 Apr 2008 01:47:24 +0000 (11:47 +1000)]
Make use of CELT_MEMSET() in find_spectral_pitch()

11 years agoFixed incorrect energy calculation in stereo intra prediction
Jean-Marc Valin [Wed, 16 Apr 2008 00:45:52 +0000 (10:45 +1000)]
Fixed incorrect energy calculation in stereo intra prediction

11 years agooptimisation: reworked intra_prediction() so that yy is computed fully only
Jean-Marc Valin [Wed, 16 Apr 2008 00:43:32 +0000 (10:43 +1000)]
optimisation: reworked intra_prediction() so that yy is computed fully only
once and then updated for each iteration

11 years agooptimisation: intra_prediction() uses a 16-bit numerator for the search
Jean-Marc Valin [Wed, 16 Apr 2008 00:29:01 +0000 (10:29 +1000)]
optimisation: intra_prediction() uses a 16-bit numerator for the search

11 years agoadded fixed_c6x.h
Jean-Marc Valin [Tue, 15 Apr 2008 11:31:42 +0000 (21:31 +1000)]
added fixed_c6x.h

11 years agooptimisation: managed to avoid dividing in the "full gain" case of alg_quant()
Jean-Marc Valin [Tue, 15 Apr 2008 11:14:18 +0000 (21:14 +1000)]
optimisation: managed to avoid dividing in the "full gain" case of alg_quant()

11 years agooops, forgot to make the gain a 16-bit var
Jean-Marc Valin [Tue, 15 Apr 2008 08:06:59 +0000 (18:06 +1000)]
oops, forgot to make the gain a 16-bit var

11 years agooptimisation: simplified the "full gain" case of alg_quant() to remove some
Jean-Marc Valin [Tue, 15 Apr 2008 08:04:33 +0000 (18:04 +1000)]
optimisation: simplified the "full gain" case of alg_quant() to remove some
32-bit muls.

11 years agooptimisation: another bunch of simplifications to the "simple case" of the
Jean-Marc Valin [Tue, 15 Apr 2008 07:31:23 +0000 (17:31 +1000)]
optimisation: another bunch of simplifications to the "simple case" of the
alg_quant() search.

11 years agoa few loop optimisations.
Jean-Marc Valin [Mon, 14 Apr 2008 22:01:33 +0000 (08:01 +1000)]
a few loop optimisations.

11 years agooptimisation: merged the init loop of vq_quant().
Jean-Marc Valin [Mon, 14 Apr 2008 21:46:48 +0000 (07:46 +1000)]
optimisation: merged the init loop of vq_quant().

11 years agofix minor compilation error/warning
Jean-Marc Valin [Fri, 11 Apr 2008 20:55:39 +0000 (06:55 +1000)]
fix minor compilation error/warning

11 years agofixed three declaration-after-statement issues
Jean-Marc Valin [Fri, 11 Apr 2008 07:23:01 +0000 (17:23 +1000)]
fixed three declaration-after-statement issues

11 years agofixed a few warnings, no real change
Jean-Marc Valin [Thu, 10 Apr 2008 21:23:53 +0000 (07:23 +1000)]
fixed a few warnings, no real change

11 years agoReworked the static modes. Now, if all static modes have the same frame size,
Jean-Marc Valin [Thu, 10 Apr 2008 18:07:52 +0000 (04:07 +1000)]
Reworked the static modes. Now, if all static modes have the same frame size,
overlap or number of channels, we can pass the hint to the compiler.

11 years agosome index work (simplifications for dumb compilers) on IMDCT
Jean-Marc Valin [Thu, 10 Apr 2008 08:54:02 +0000 (18:54 +1000)]
some index work (simplifications for dumb compilers) on IMDCT

11 years agoAdded a missing RESTORE_STACK in intra_prediction()
Jean-Marc Valin [Thu, 10 Apr 2008 04:43:59 +0000 (14:43 +1000)]
Added a missing RESTORE_STACK in intra_prediction()

11 years agoNot all compilers are equal -- making it clearer how the MDCT indexing is done
Jean-Marc Valin [Thu, 10 Apr 2008 04:38:14 +0000 (14:38 +1000)]
Not all compilers are equal -- making it clearer how the MDCT indexing is done

11 years agoDefining IMUL32 for 32x32=>32 int multiplications and using it in the range
Jean-Marc Valin [Thu, 10 Apr 2008 03:24:48 +0000 (13:24 +1000)]
Defining IMUL32 for 32x32=>32 int multiplications and using it in the range
coder

11 years agoSimplified indexing in intra_prediction()
Jean-Marc Valin [Thu, 10 Apr 2008 02:21:26 +0000 (12:21 +1000)]
Simplified indexing in intra_prediction()

11 years agofixed ordering of the channels in the intra prediction.
Jean-Marc Valin [Thu, 10 Apr 2008 01:00:51 +0000 (11:00 +1000)]
fixed ordering of the channels in the intra prediction.

11 years agoDefining DISABLE_STEREO now optimises for the mono case
Jean-Marc Valin [Wed, 9 Apr 2008 23:13:05 +0000 (09:13 +1000)]
Defining DISABLE_STEREO now optimises for the mono case

11 years agoFixed a stereo regression introduced in e28f25f0d14959d521fda0cdb8f1220995bc50e8
Jean-Marc Valin [Wed, 9 Apr 2008 23:00:12 +0000 (09:00 +1000)]
Fixed a stereo regression introduced in e28f25f0d14959d521fda0cdb8f1220995bc50e8

11 years agoFixed rsqrt testcase for float
Jean-Marc Valin [Sat, 5 Apr 2008 04:36:05 +0000 (14:36 +1000)]
Fixed rsqrt testcase for float

11 years agoRework CWRS code.
Timothy B. Terriberry [Fri, 4 Apr 2008 14:16:19 +0000 (10:16 -0400)]
Rework CWRS code.

This eliminates an extra O(nm) lookups on decode, and reduces the rate control
 from O(nm^2) to O(nm), in addition to eliminating O(m) lookups on both encode
 and decode.
Although the interface is slightly more complex, the internal code is also
 simpler.

11 years agoChanged the rules for using the pulse spreading. It should be used less often
Jean-Marc Valin [Fri, 28 Mar 2008 06:59:36 +0000 (17:59 +1100)]
Changed the rules for using the pulse spreading. It should be used less often
and mainly in the mid-range.

11 years agoRevert ABS16/32 on C55 -- ended up being slower
Jean-Marc Valin [Thu, 27 Mar 2008 21:45:46 +0000 (08:45 +1100)]
Revert ABS16/32 on C55 -- ended up being slower

11 years agoABS16 and ABS32 for the C55
Jean-Marc Valin [Thu, 27 Mar 2008 21:31:33 +0000 (08:31 +1100)]
ABS16 and ABS32 for the C55

11 years agoMaking the pulsesAtOnce code 16-bit safe.
Jean-Marc Valin [Thu, 27 Mar 2008 13:33:04 +0000 (00:33 +1100)]
Making the pulsesAtOnce code 16-bit safe.

11 years agoJust commenting -- nothing to see.
Jean-Marc Valin [Thu, 27 Mar 2008 07:05:01 +0000 (18:05 +1100)]
Just commenting -- nothing to see.

11 years agoOptimisation: got rid of about 10% of the 32-bit divisions by using ec_enc_uint
Jean-Marc Valin [Thu, 27 Mar 2008 04:41:38 +0000 (15:41 +1100)]
Optimisation: got rid of about 10% of the 32-bit divisions by using ec_enc_uint
in energy quantisation when possible.

11 years agoRemoved a few int divisions from the intra prediction code.
Jean-Marc Valin [Thu, 27 Mar 2008 03:18:28 +0000 (14:18 +1100)]
Removed a few int divisions from the intra prediction code.

11 years agoremoved some int div
Jean-Marc Valin [Thu, 27 Mar 2008 01:22:44 +0000 (12:22 +1100)]
removed some int div

11 years agofixed-point: using MULT16_16 instead of * in compute_band_energies()
Jean-Marc Valin [Wed, 26 Mar 2008 23:22:42 +0000 (10:22 +1100)]
fixed-point: using MULT16_16 instead of * in compute_band_energies()

11 years agoMaking a few functions static inline
Jean-Marc Valin [Wed, 26 Mar 2008 22:00:14 +0000 (09:00 +1100)]
Making a few functions static inline

11 years agoTrying to clean up celt_ilog2() vs. EC_ILOG a bit.
Jean-Marc Valin [Wed, 26 Mar 2008 21:40:45 +0000 (08:40 +1100)]
Trying to clean up celt_ilog2() vs. EC_ILOG a bit.

11 years agomaking {next|prev}_cwrs* inline
Jean-Marc Valin [Wed, 26 Mar 2008 12:33:06 +0000 (23:33 +1100)]
making {next|prev}_cwrs* inline

11 years agooptimisation: changed some for() loops to do-while() to give the compiler
Jean-Marc Valin [Wed, 26 Mar 2008 10:31:56 +0000 (21:31 +1100)]
optimisation: changed some for() loops to do-while() to give the compiler
a hint that there has to be at least one iteration.

11 years agoMaking it obvious to the compiler how to generate a dual-MAC in
Jean-Marc Valin [Wed, 26 Mar 2008 07:03:22 +0000 (18:03 +1100)]
Making it obvious to the compiler how to generate a dual-MAC in
mix_pitch_and_residual()

11 years agomix_pitch_and_residual() no longer computing Ryp twice
Jean-Marc Valin [Wed, 26 Mar 2008 06:57:49 +0000 (17:57 +1100)]
mix_pitch_and_residual() no longer computing Ryp twice

11 years agooptimisation: defined a reciprocal square root (celt_rsqrt) for use in
Jean-Marc Valin [Wed, 26 Mar 2008 05:42:42 +0000 (16:42 +1100)]
optimisation: defined a reciprocal square root (celt_rsqrt) for use in
find_spectral_pitch instead of using celt_rcp(celt_sqrt(x))

11 years agoFixed the rcp() testcase for new assumptions (x is positive)
Jean-Marc Valin [Wed, 26 Mar 2008 04:56:07 +0000 (15:56 +1100)]
Fixed the rcp() testcase for new assumptions (x is positive)

11 years agooptimisation: shaving a few cycles off prev_cwrs* by not computed the values
Jean-Marc Valin [Wed, 26 Mar 2008 04:46:51 +0000 (15:46 +1100)]
optimisation: shaving a few cycles off prev_cwrs* by not computed the values
we're not going to use.

11 years agooptimisations: faster handling of the zero for compute_band_energies() and
Jean-Marc Valin [Wed, 26 Mar 2008 03:36:07 +0000 (14:36 +1100)]
optimisations: faster handling of the zero for compute_band_energies() and
normalise_bands(). Also, another bunch of restrict qualifiers -- mainly
to tell the compiler there's no aliasing between the array bounds and the
data.

11 years agochanged 1*rcp(x) to just rcp(x)
Jean-Marc Valin [Wed, 26 Mar 2008 02:45:01 +0000 (13:45 +1100)]
changed 1*rcp(x) to just rcp(x)

11 years agooptimisation: intra_prediction() no longer needs to divide inside the search
Jean-Marc Valin [Wed, 26 Mar 2008 01:16:00 +0000 (12:16 +1100)]
optimisation: intra_prediction() no longer needs to divide inside the search

11 years agooptimisation: The "simple" Rxy/sqrt(Ryy) case in alg_quant no longer requires
Jean-Marc Valin [Tue, 25 Mar 2008 23:34:23 +0000 (10:34 +1100)]
optimisation: The "simple" Rxy/sqrt(Ryy) case in alg_quant no longer requires
a division

11 years agoDecision on whether to use pitch is now taken only based on energy in the
Jean-Marc Valin [Tue, 25 Mar 2008 21:10:27 +0000 (08:10 +1100)]
Decision on whether to use pitch is now taken only based on energy in the
first three bands (instead of total MDCT window energy)

11 years agoproperly defined EPSILON for the float case
Jean-Marc Valin [Tue, 25 Mar 2008 13:00:18 +0000 (00:00 +1100)]
properly defined EPSILON for the float case

11 years agoA bunch of pointers marked as "restrict" to ease the job of the compiler
Jean-Marc Valin [Tue, 25 Mar 2008 11:38:58 +0000 (22:38 +1100)]
A bunch of pointers marked as "restrict" to ease the job of the compiler

11 years agooptimisation: spreading_func now in-place with no branch in the loop and half
Jean-Marc Valin [Tue, 25 Mar 2008 11:09:32 +0000 (22:09 +1100)]
optimisation: spreading_func now in-place with no branch in the loop and half
the multiplications

11 years agooops. find_max32() now uses VERY_LARGE32 (instead of VERY_LARGE16)
Jean-Marc Valin [Tue, 25 Mar 2008 10:31:12 +0000 (21:31 +1100)]
oops. find_max32() now uses VERY_LARGE32 (instead of VERY_LARGE16)

11 years agooptimisations: Another bunch of simplifications to alg_quant(), mainly to
Jean-Marc Valin [Tue, 25 Mar 2008 10:28:40 +0000 (21:28 +1100)]
optimisations: Another bunch of simplifications to alg_quant(), mainly to
remove unnecessary copying and some conditional branches.

11 years agooptimisation: Making use of restrict in find_spectral_pitch() to disambiguate
Jean-Marc Valin [Tue, 25 Mar 2008 07:04:08 +0000 (18:04 +1100)]
optimisation: Making use of restrict in find_spectral_pitch() to disambiguate
X and Y.

11 years agooptimisations: caching sign of x in alg_quant(), changed celt_div()/celt_rcp()
Jean-Marc Valin [Tue, 25 Mar 2008 05:07:05 +0000 (16:07 +1100)]
optimisations: caching sign of x in alg_quant(), changed celt_div()/celt_rcp()
to assume denominator is positive.

11 years agoOptimised intra prediction a bit -- removed a conditional branch and replaced
Jean-Marc Valin [Tue, 25 Mar 2008 04:25:08 +0000 (15:25 +1100)]
Optimised intra prediction a bit -- removed a conditional branch and replaced
some calls to ec_enc_uint() by calls to ec_enc_bits() (saves divisions)

11 years agoRemoved support for more than one MDCT blocks per frame. I don't think there's
Jean-Marc Valin [Tue, 25 Mar 2008 03:38:55 +0000 (14:38 +1100)]
Removed support for more than one MDCT blocks per frame. I don't think there's
a use for that anymore now that the overlap can be less than one frame.

11 years agoRemoved the "pitch compression" in the residual quantisation. Also, removed
Jean-Marc Valin [Tue, 25 Mar 2008 03:15:41 +0000 (14:15 +1100)]
Removed the "pitch compression" in the residual quantisation. Also, removed
the more complex "n-best search" and replaced it with a greedy search

11 years agoUnrolled the inner loop in vq_index() so that the codebook unpacking doesn't
Jean-Marc Valin [Mon, 24 Mar 2008 10:39:49 +0000 (21:39 +1100)]
Unrolled the inner loop in vq_index() so that the codebook unpacking doesn't
require a branch.

11 years agoMaking bits2pulses() use a fixed number of iterations to allow further
Jean-Marc Valin [Mon, 24 Mar 2008 06:06:08 +0000 (17:06 +1100)]
Making bits2pulses() use a fixed number of iterations to allow further
optimisations.

11 years agoinclude "dsplib.h" in fixed_c5x.h
Jean-Marc Valin [Sun, 23 Mar 2008 21:32:49 +0000 (08:32 +1100)]
include "dsplib.h" in fixed_c5x.h

11 years agoreplaced divisions by recoprocals in intra prediction and folding
Jean-Marc Valin [Sun, 23 Mar 2008 21:15:40 +0000 (08:15 +1100)]
replaced divisions by recoprocals in intra prediction and folding

11 years agodefined find_max16 and overrode it for C55x
Jean-Marc Valin [Sat, 22 Mar 2008 21:06:29 +0000 (08:06 +1100)]
defined find_max16 and overrode it for C55x

11 years agoMade a second version of ec_{en|de}code optimised for encoding bits (no div
Jean-Marc Valin [Sat, 22 Mar 2008 11:58:45 +0000 (22:58 +1100)]
Made a second version of ec_{en|de}code optimised for encoding bits (no div
required) and using it in ec_{en|de}c_bits()

11 years agoNo longer trying to save bits when encoding integers near the upper limit
Jean-Marc Valin [Sat, 22 Mar 2008 11:23:58 +0000 (22:23 +1100)]
No longer trying to save bits when encoding integers near the upper limit
(and fix for celt_div with 16-bit numerator on a 16-bit CPU)

11 years agofixed-point: added cheap celt_div() division using a reciprocal
Jean-Marc Valin [Sat, 22 Mar 2008 10:17:45 +0000 (21:17 +1100)]
fixed-point: added cheap celt_div() division using a reciprocal

11 years agomore use of restrict
Jean-Marc Valin [Sat, 22 Mar 2008 06:53:12 +0000 (17:53 +1100)]
more use of restrict

11 years agoUsing restrict to make it clear there's no aliasing issues in the mdct.
Jean-Marc Valin [Sat, 22 Mar 2008 00:23:57 +0000 (11:23 +1100)]
Using restrict to make it clear there's no aliasing issues in the mdct.

11 years agoAdded a few "restrict" keywords and changed some divisions to shifts
Jean-Marc Valin [Fri, 21 Mar 2008 22:01:50 +0000 (09:01 +1100)]
Added a few "restrict" keywords and changed some divisions to shifts

11 years agofixed TI fft code -- again
Jean-Marc Valin [Fri, 21 Mar 2008 11:54:41 +0000 (22:54 +1100)]
fixed TI fft code -- again

11 years agoRemoved potentially unused var in MDCT init
Jean-Marc Valin [Fri, 21 Mar 2008 02:11:16 +0000 (13:11 +1100)]
Removed potentially unused var in MDCT init

11 years agooops (TI FFT)
Jean-Marc Valin [Fri, 21 Mar 2008 02:02:53 +0000 (13:02 +1100)]
oops (TI FFT)

11 years agolocal var name maxval was shadowing the TI function used to compute it
Jean-Marc Valin [Fri, 21 Mar 2008 01:53:37 +0000 (12:53 +1100)]
local var name maxval was shadowing the TI function used to compute it

11 years agomake sure TRIG_UPSCALE is properly defined
Jean-Marc Valin [Fri, 21 Mar 2008 01:51:05 +0000 (12:51 +1100)]
make sure TRIG_UPSCALE is properly defined

11 years agofix for TI version of celt_maxabs16()
Jean-Marc Valin [Fri, 21 Mar 2008 01:49:36 +0000 (12:49 +1100)]
fix for TI version of celt_maxabs16()

11 years agofixed-point: fix for 32-bit TI FFT
Jean-Marc Valin [Fri, 21 Mar 2008 01:47:41 +0000 (12:47 +1100)]
fixed-point: fix for 32-bit TI FFT

11 years agofixed-point: Wrapper for the 32-bit complex FFT used in the MDCT so we can use
Jean-Marc Valin [Fri, 21 Mar 2008 01:09:27 +0000 (12:09 +1100)]
fixed-point: Wrapper for the 32-bit complex FFT used in the MDCT so we can use
the TI dsplib FFT.

11 years agofixed-point: defined celt_maxabs16() as basic operator
Jean-Marc Valin [Fri, 21 Mar 2008 00:13:51 +0000 (11:13 +1100)]
fixed-point: defined celt_maxabs16() as basic operator