Age | Commit message (Collapse) | Author | |
---|---|---|---|
2009-07-01 | Fix compilation of test cases with fixed-debug enabled | Gregory Maxwell | |
2009-06-29 | Additional mathops.h tests. | Gregory Maxwell | |
2009-06-05 | Fixes for the tests in pseudo-stack mode. | Gregory Maxwell | |
2009-05-27 | Make ectest use random random numbers. Add a new test | Gregory Maxwell | |
"tandem-test" which tests the encoder and decoder in many modes. | |||
2009-05-27 | CWRS clean-ups and optimizations. | Timothy B. Terriberry | |
Adds specialized O(N*log(K)) versions of cwrsi() and O(N) versions of icwrs() for N={3,4,5}, which allows them to operate all the way up to the theoretical pulse limit without serious performance degredation. Also substantially reduces the computation time and stack usage of get_required_bits(). On x86-64, this gives a 2% speed-up for 256 sample frames, and almost a 16% speed-up for 64 sample frames. | |||
2008-12-23 | Unit tests now #include the relevant source files to prevent symbol visibility | Jean-Marc Valin | |
problems when building shared libraries only. | |||
2008-12-21 | Fix ectest to not check a case which isn't guaranteed to work, and which we ↵ | Timothy B. Terriberry | |
don't use. When I removed the special case for EC_ILOG(0) in commit 06390d082dcdfa8addb3dde337543bc0f0ebae44, it broke ec_dec_uint() with _ft=1 (which should encode the value 0 using 0 bits). This feature was tested by ectest.c, but not actually used by libcelt. An assert has been added to ec_dec_uint() to ensure that we don't try to use this feature by accident. ec_enc_uint() was actually correct, but support for this feature has been removed and the assert put in its place. | |||
2008-11-11 | fixed test for cases where M_PI isn't defined | Jean-Marc Valin | |
2008-10-05 | Removing the 64-bit part of the range coder. | Jean-Marc Valin | |
2008-09-22 | Merge branch 'cwrs_speedup' | Jean-Marc Valin | |
Conflicts: libcelt/cwrs.c tests/cwrs32-test.c | |||
2008-09-22 | Change cwrsi() to operate on rows of U instead of columns. | Timothy B.B Terriberry | |
It is no slower with a large number of pulses, and as much as 30% faster with a large number of dimensions. | |||
2008-09-19 | Fixing the manual stack handling code | Jean-Marc Valin | |
2008-09-19 | Merge branch 'cwrs_speedup' (derf's cwrs changes) | Jean-Marc Valin | |
Conflicts: libcelt/cwrs.c | |||
2008-09-19 | Change CWRS indexing to use Pyramid VQ's magnitude ordering. | Timothy B.B Terriberry | |
This lets us encode and decode directly from the pulse vector without an intermediate transformation. This makes old streams undecodable. Additionally, ncwrs_u32() has been sped up for large N by using the sliding recurrence from Mohorko et al. ncwrs_u64 could be sped up in a similar manner, but would require a larger table of multiplicative inverses (or several 32x32->64 bit multiplies). Note that U(N,M) is now everywhere 1/2 the value it used to be. | |||
2008-09-13 | Converted everything to 32-bit CWRS (using split after that) | Jean-Marc Valin | |
2008-05-27 | Making the Laplace test case pass again | Jean-Marc Valin | |
2008-05-05 | Disabled pulse spreading until I can show it actually helps | Jean-Marc Valin | |
2008-04-25 | Brought the mdct test-case with the recent changes to the mdct semantic | Jean-Marc Valin | |
2008-04-22 | Moved the windowing from compute_inv_mdcts() to mdct_backward() | Jean-Marc Valin | |
2008-04-21 | Moved the windowing operation from compute_mdcts() to mdct_forward() in an | Jean-Marc Valin | |
attempt to reduce copying. | |||
2008-04-21 | Further simplifications to comb2pulses() to remove all conditional branches. | Jean-Marc Valin | |
2008-04-18 | pseudo-stack no longer checks on every function entry whether it has been | Jean-Marc Valin | |
allocated | |||
2008-04-18 | optimisation: Removed a bunch of conditional branches from comb2pulse() | Jean-Marc Valin | |
2008-04-16 | optimisation: Making it clear to the compiler that many of the loops in cwrs | Jean-Marc Valin | |
need to iterate at least once. | |||
2008-04-11 | fixed a few warnings, no real change | Jean-Marc Valin | |
2008-04-05 | Fixed rsqrt testcase for float | Jean-Marc Valin | |
2008-04-05 | Rework CWRS code. | Timothy B. Terriberry | |
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. | |||
2008-03-26 | optimisation: defined a reciprocal square root (celt_rsqrt) for use in | Jean-Marc Valin | |
find_spectral_pitch instead of using celt_rcp(celt_sqrt(x)) | |||
2008-03-26 | Fixed the rcp() testcase for new assumptions (x is positive) | Jean-Marc Valin | |
2008-03-19 | Random numbers should work on 16-bit archs. | Jean-Marc Valin | |
2008-03-13 | Adjusting/fixing warnings | Jean-Marc Valin | |
2008-03-13 | Added mathops-test | Jean-Marc Valin | |
2008-03-12 | Testcases should now work even when symbols aren't visible in the dso | Jean-Marc Valin | |
2008-03-07 | Changed the pulse spreading rotations so that the number of iterations is | Jean-Marc Valin | |
variable instead of the angle. Should now be both less CPU-intensive and makes fixed-point code simpler. | |||
2008-03-05 | Defining RADIX_TWO_ONLY removes all butterflies, except for radix 2 and 4. | Jean-Marc Valin | |
2008-03-04 | fixed-point: exp_rotation() now fully converted, using an approximation of the | Jean-Marc Valin | |
cos() and sin() functions. | |||
2008-03-03 | cwrs32-test should be 16-bit clean now. | Jean-Marc Valin | |
2008-03-03 | Laplace encoder now works with 16-bit frequencies | Jean-Marc Valin | |
2008-03-02 | oops, fixed the Laplace testcase | Jean-Marc Valin | |
2008-03-01 | Making it easier to debug laplace testcase | Jean-Marc Valin | |
2008-02-27 | fixed-point: exp_rotation() mostly converted (still need to convert the cos/sin) | Jean-Marc Valin | |
2008-02-26 | C90-fying ectest | Jean-Marc Valin | |
2008-02-25 | MDCT is in fixed-point now | Jean-Marc Valin | |
2008-02-25 | MDCT conversion, part I. | Jean-Marc Valin | |
2008-02-24 | real fft snr comparison testcase includes DC | Jean-Marc Valin | |
2008-02-24 | Float FFT now does the same scaling as the fixed-point FFT | Jean-Marc Valin | |
2008-02-24 | minor tweak to FFT | Jean-Marc Valin | |
2008-02-24 | Added a mixed-precision version of the FFT with 32-bit data and 16-bit twiddles. | Jean-Marc Valin | |
2008-02-22 | MDCT now scales down by N/2 instead of N/4. The factor two is moved to the | Jean-Marc Valin | |
overlap-add during synthesis. | |||
2008-02-22 | testcase for the MDCT and IMDCT | Jean-Marc Valin | |