opus.git
2008-04-10 Jean-Marc Valinfixed a few warnings, no real change
2008-04-10 Jean-Marc ValinReworked the static modes. Now, if all static modes...
2008-04-10 Jean-Marc Valinsome index work (simplifications for dumb compilers...
2008-04-10 Jean-Marc ValinAdded a missing RESTORE_STACK in intra_prediction()
2008-04-10 Jean-Marc ValinNot all compilers are equal -- making it clearer how...
2008-04-10 Jean-Marc ValinDefining IMUL32 for 32x32=>32 int multiplications and...
2008-04-10 Jean-Marc ValinSimplified indexing in intra_prediction()
2008-04-10 Jean-Marc Valinfixed ordering of the channels in the intra prediction.
2008-04-09 Jean-Marc ValinDefining DISABLE_STEREO now optimises for the mono...
2008-04-09 Jean-Marc ValinFixed a stereo regression introduced in e28f25f0d14959d...
2008-04-05 Jean-Marc ValinFixed rsqrt testcase for float
2008-04-05 Timothy B.... Rework CWRS code.
2008-03-28 Jean-Marc ValinChanged the rules for using the pulse spreading. It...
2008-03-27 Jean-Marc ValinRevert ABS16/32 on C55 -- ended up being slower
2008-03-27 Jean-Marc ValinABS16 and ABS32 for the C55
2008-03-27 Jean-Marc ValinMaking the pulsesAtOnce code 16-bit safe.
2008-03-27 Jean-Marc ValinJust commenting -- nothing to see.
2008-03-27 Jean-Marc ValinOptimisation: got rid of about 10% of the 32-bit divisi...
2008-03-27 Jean-Marc ValinRemoved a few int divisions from the intra prediction...
2008-03-27 Jean-Marc Valinremoved some int div
2008-03-26 Jean-Marc Valinfixed-point: using MULT16_16 instead of * in compute_ba...
2008-03-26 Jean-Marc ValinMaking a few functions static inline
2008-03-26 Jean-Marc ValinTrying to clean up celt_ilog2() vs. EC_ILOG a bit.
2008-03-26 Jean-Marc Valinmaking {next|prev}_cwrs* inline
2008-03-26 Jean-Marc Valinoptimisation: changed some for() loops to do-while...
2008-03-26 Jean-Marc ValinMaking it obvious to the compiler how to generate a...
2008-03-26 Jean-Marc Valinmix_pitch_and_residual() no longer computing Ryp twice
2008-03-26 Jean-Marc Valinoptimisation: defined a reciprocal square root (celt_rs...
2008-03-26 Jean-Marc ValinFixed the rcp() testcase for new assumptions (x is...
2008-03-26 Jean-Marc Valinoptimisation: shaving a few cycles off prev_cwrs* by...
2008-03-26 Jean-Marc Valinoptimisations: faster handling of the zero for compute_...
2008-03-26 Jean-Marc Valinchanged 1*rcp(x) to just rcp(x)
2008-03-26 Jean-Marc Valinoptimisation: intra_prediction() no longer needs to...
2008-03-25 Jean-Marc Valinoptimisation: The "simple" Rxy/sqrt(Ryy) case in alg_qu...
2008-03-25 Jean-Marc ValinDecision on whether to use pitch is now taken only...
2008-03-25 Jean-Marc Valinproperly defined EPSILON for the float case
2008-03-25 Jean-Marc ValinA bunch of pointers marked as "restrict" to ease the...
2008-03-25 Jean-Marc Valinoptimisation: spreading_func now in-place with no branc...
2008-03-25 Jean-Marc Valinoops. find_max32() now uses VERY_LARGE32 (instead of...
2008-03-25 Jean-Marc Valinoptimisations: Another bunch of simplifications to...
2008-03-25 Jean-Marc Valinoptimisation: Making use of restrict in find_spectral_p...
2008-03-25 Jean-Marc Valinoptimisations: caching sign of x in alg_quant(), change...
2008-03-25 Jean-Marc ValinOptimised intra prediction a bit -- removed a condition...
2008-03-25 Jean-Marc ValinRemoved support for more than one MDCT blocks per frame...
2008-03-25 Jean-Marc ValinRemoved the "pitch compression" in the residual quantis...
2008-03-24 Jean-Marc ValinUnrolled the inner loop in vq_index() so that the codeb...
2008-03-24 Jean-Marc ValinMaking bits2pulses() use a fixed number of iterations...
2008-03-23 Jean-Marc Valininclude "dsplib.h" in fixed_c5x.h
2008-03-23 Jean-Marc Valinreplaced divisions by recoprocals in intra prediction...
2008-03-22 Jean-Marc Valindefined find_max16 and overrode it for C55x
2008-03-22 Jean-Marc ValinMade a second version of ec_{en|de}code optimised for...
2008-03-22 Jean-Marc ValinNo longer trying to save bits when encoding integers...
2008-03-22 Jean-Marc Valinfixed-point: added cheap celt_div() division using...
2008-03-22 Jean-Marc Valinmore use of restrict
2008-03-22 Jean-Marc ValinUsing restrict to make it clear there's no aliasing...
2008-03-21 Jean-Marc ValinAdded a few "restrict" keywords and changed some divisi...
2008-03-21 Jean-Marc Valinfixed TI fft code -- again
2008-03-21 Jean-Marc ValinRemoved potentially unused var in MDCT init
2008-03-21 Jean-Marc Valinoops (TI FFT)
2008-03-21 Jean-Marc Valinlocal var name maxval was shadowing the TI function...
2008-03-21 Jean-Marc Valinmake sure TRIG_UPSCALE is properly defined
2008-03-21 Jean-Marc Valinfix for TI version of celt_maxabs16()
2008-03-21 Jean-Marc Valinfixed-point: fix for 32-bit TI FFT
2008-03-21 Jean-Marc Valinfixed-point: Wrapper for the 32-bit complex FFT used...
2008-03-21 Jean-Marc Valinfixed-point: defined celt_maxabs16() as basic operator
2008-03-20 Jean-Marc Valinfixed-point: MULT16_32_Q15 for TI DSP (not entirely...
2008-03-20 Jean-Marc Valins/ROUND/ROUND16/
2008-03-20 Jean-Marc Valinfixed-point: using TI intrinsic for celt_ilog2() if...
2008-03-20 Jean-Marc ValinWrapper for the TI dsplib FFT
2008-03-20 Jean-Marc ValinMaking the real/single FFT easier to replace
2008-03-18 Jean-Marc ValinRandom numbers should work on 16-bit archs.
2008-03-16 Jean-Marc Valinfixed-point: more TI macros. Comments on the existing...
2008-03-16 Jean-Marc ValinRemoved useless functions that caused warnings in skele...
2008-03-15 Jean-Marc ValinMaking sure not to use the C library calls directly
2008-03-15 Jean-Marc ValinSome work on assertions.
2008-03-14 Jean-Marc ValinNew C55 macro
2008-03-14 Jean-Marc Valinfixed-point: Fixed a "shift too large" warning by using...
2008-03-14 Jean-Marc Valinfixed-point: playing it safe. SHL32() now automatically...
2008-03-14 Jean-Marc Valinfixed-point: fixed a 16-bit issue in normalise_bands...
2008-03-14 Jean-Marc ValinRemoved unnecessary header inclusions
2008-03-13 Jean-Marc ValinAdjusting/fixing warnings
2008-03-13 Jean-Marc Valinfixed-point: Using reciproqual instead of DIV32_16...
2008-03-13 Jean-Marc ValinEnabling warnings and restricting symbol visibility
2008-03-13 Jean-Marc ValinAdded mathops-test
2008-03-13 Jean-Marc Valinfixed-point: simplification of the gain in mix_pitch_an...
2008-03-13 Jean-Marc Valinfixed-point: simplifying the arithmetic in alg_quant()
2008-03-12 Jean-Marc ValinMore accurate sqrt approximation using MULT16_16_Q15...
2008-03-12 Jean-Marc ValinTODO update
2008-03-12 Jean-Marc ValinIncreased accuracy of the reciprocal approximation...
2008-03-12 Jean-Marc ValinTestcases should now work even when symbols aren't...
2008-03-12 Jean-Marc ValinAdded info for properly exporting symbols
2008-03-12 Jean-Marc ValinChanged definition of VARDECL to make it more flexible...
2008-03-12 Jean-Marc ValinUsing reciprocal approximation instead of full 32-bit...
2008-03-12 Jean-Marc Valinconfigure update (MIXED_PRECISION and STATIC_MODES...
2008-03-12 Jean-Marc ValinDefined the celt_rcp() reciprocal approximation
2008-03-11 Jean-Marc Valin__GNUC_PREREQ should be causing problems anymore.
2008-03-11 Jean-Marc ValinAdded macro definitions for the TI C5x family (untested)
2008-03-11 Jean-Marc ValinBetter heuristic for deciding when to use the 32-bit...
2008-03-10 Jean-Marc Valinconst correctness. celt_mode_create() now makes a copy...
2008-03-10 Jean-Marc Valinoops, forgot to disable the main()
next