| 2006-11-12 |
jm | Originally written for Symbian, it is no longer needed... |
commit | commitdiff | tree | snapshot |
| 2006-11-12 |
jm | Now compiles fine with gcc 3.4 (which doesn't know... |
commit | commitdiff | tree | snapshot |
| 2006-11-12 |
jm | Added some checks to detect if the echo canceller has... |
commit | commitdiff | tree | snapshot |
| 2006-11-12 |
jm | Documentation for the echo canceller and preprocessor... |
commit | commitdiff | tree | snapshot |
| 2006-11-12 |
jm | cleanup, comments, better handling of clipping in AEC. |
commit | commitdiff | tree | snapshot |
| 2006-11-12 |
jm | Implemented atan in fixed-point and used it for toBARK... |
commit | commitdiff | tree | snapshot |
| 2006-11-11 |
jm | converted most of the filterbank initialisation to... |
commit | commitdiff | tree | snapshot |
| 2006-11-11 |
jm | a few checks just to be on the same (>0) side. |
commit | commitdiff | tree | snapshot |
| 2006-11-10 |
jm | pseudo-float divide operators should be more robust... |
commit | commitdiff | tree | snapshot |
| 2006-11-10 |
jm | Defined VSHR32 and using it instead of "if (shift>0... |
commit | commitdiff | tree | snapshot |
| 2006-11-09 |
jm | Reduced (and renamed) input buffer |
commit | commitdiff | tree | snapshot |
| 2006-11-09 |
jm | shouldn't call exit(). |
commit | commitdiff | tree | snapshot |
| 2006-11-09 |
jm | a bit of cleaning up and comments |
commit | commitdiff | tree | snapshot |
| 2006-11-08 |
jm | Added a test for the second operand in MULT16_32_QX |
commit | commitdiff | tree | snapshot |
| 2006-11-08 |
jm | Fixed potential overflow in gradient computation. Need... |
commit | commitdiff | tree | snapshot |
| 2006-11-07 |
jm | kissfft initialisation no longer uses any float operati... |
commit | commitdiff | tree | snapshot |
| 2006-11-07 |
jm | Added a higher precision cosine approximation and used... |
commit | commitdiff | tree | snapshot |
| 2006-11-07 |
jm | mdf.c now has zero float operations left (again) |
commit | commitdiff | tree | snapshot |
| 2006-11-07 |
jm | disabling some unused float operations |
commit | commitdiff | tree | snapshot |
| 2006-11-06 |
jm | Re-wrote spx_sqrt in terms of spx_ilog4, a few MACs... |
commit | commitdiff | tree | snapshot |
| 2006-11-05 |
jm | Converted gain floor computation to fixed-point. Think... |
commit | commitdiff | tree | snapshot |
| 2006-11-05 |
jm | Fast Bark-scale gain computation is in fixed-point |
commit | commitdiff | tree | snapshot |
| 2006-11-05 |
jm | some VAD stuff converted to fixed-point. Also, changed... |
commit | commitdiff | tree | snapshot |
| 2006-11-05 |
jm | OK, so allocating 4 bytes per short did tend to increas... |
commit | commitdiff | tree | snapshot |
| 2006-11-05 |
jm | converted noise averaging to fixed-point, completed... |
commit | commitdiff | tree | snapshot |
| 2006-11-05 |
jm | Fixed excessive MCRA min offset (but kept a small ont... |
commit | commitdiff | tree | snapshot |
| 2006-11-04 |
jm | Think I'm done converting the MCRA part. |
commit | commitdiff | tree | snapshot |
| 2006-11-04 |
jm | Converted output gain and Zframe. Also disabled AGC... |
commit | commitdiff | tree | snapshot |
| 2006-11-04 |
jm | Converted conditional speech presence prob to fixed... |
commit | commitdiff | tree | snapshot |
| 2006-11-04 |
jm | Converted Pframe, P1 and q to fixed-point. |
commit | commitdiff | tree | snapshot |
| 2006-11-03 |
jm | almost done converting hypergeom_gain() to fixed-point |
commit | commitdiff | tree | snapshot |
| 2006-11-03 |
jm | Half-way done converting hypergeometric function approx... |
commit | commitdiff | tree | snapshot |
| 2006-11-03 |
jm | Fixed the PSHR* functions for a shift of zero and fixed... |
commit | commitdiff | tree | snapshot |
| 2006-11-03 |
jm | Converted qcurve and probability of speech presence |
commit | commitdiff | tree | snapshot |
| 2006-11-03 |
jm | I think it's generally a better idea to include the... |
commit | commitdiff | tree | snapshot |
| 2006-11-02 |
jm | linear frequency loop now entirely in fixed-point. |
commit | commitdiff | tree | snapshot |
| 2006-11-02 |
jm | gain computation in linear frequency loop converted... |
commit | commitdiff | tree | snapshot |
| 2006-11-02 |
jm | EXTRACT16() for correctness |
commit | commitdiff | tree | snapshot |
| 2006-11-02 |
jm | Oops. Previous overflow fix had actually made things... |
commit | commitdiff | tree | snapshot |
| 2006-11-02 |
jm | fixed an overflow when the numerator is negative |
commit | commitdiff | tree | snapshot |
| 2006-11-02 |
jm | I guess using the right precision in the operators... |
commit | commitdiff | tree | snapshot |
| 2006-11-02 |
jm | Converted part of the filterbank and added MULT16_32_P1... |
commit | commitdiff | tree | snapshot |
| 2006-11-02 |
jm | fixed undefined condition in PSHR/PSHR16/PSHR32 when... |
commit | commitdiff | tree | snapshot |
| 2006-11-02 |
jm | Improved the playback buffering. |
commit | commitdiff | tree | snapshot |
| 2006-11-01 |
jm | converted computation of old_ps and prior_ratio |
commit | commitdiff | tree | snapshot |
| 2006-11-01 |
jm | converted theta and prior_ratio to fixed-point. |
commit | commitdiff | tree | snapshot |
| 2006-11-01 |
jm | all important variables in the preprocessor state now... |
commit | commitdiff | tree | snapshot |
| 2006-11-01 |
jm | gain and gain2 are now spx_word16_t, though computation... |
commit | commitdiff | tree | snapshot |
| 2006-11-01 |
jm | SNR loop now entirely in fixed-point |
commit | commitdiff | tree | snapshot |
| 2006-11-01 |
jm | a priori SNR in spx_word16_t |
commit | commitdiff | tree | snapshot |
| 2006-11-01 |
jm | a posteriori SNR now in spx_word16_t |
commit | commitdiff | tree | snapshot |
| 2006-11-01 |
jm | gamma calculation converted (not the value). |
commit | commitdiff | tree | snapshot |
| 2006-11-01 |
jm | Removed explicit call to exit() in error handling. |
commit | commitdiff | tree | snapshot |
| 2006-10-31 |
jm | Quantizing noise estimate (with higher accuracy). Also... |
commit | commitdiff | tree | snapshot |
| 2006-10-30 |
jm | Re-converted window to fixed-point |
commit | commitdiff | tree | snapshot |
| 2006-10-30 |
jm | A bit of backtracking to fix quality regressions. Need... |
commit | commitdiff | tree | snapshot |
| 2006-10-29 |
jm | a priori SNR and a posteriori SNR computed with fixed... |
commit | commitdiff | tree | snapshot |
| 2006-10-27 |
jm | another MS user complaining about C99 code. |
commit | commitdiff | tree | snapshot |
| 2006-10-27 |
jm | window is now spx_word16_t |
commit | commitdiff | tree | snapshot |
| 2006-10-27 |
jm | Last bits of the easy fixed-point conversions. |
commit | commitdiff | tree | snapshot |
| 2006-10-27 |
jm | another bunch of easy conversions to fixed-point (befor... |
commit | commitdiff | tree | snapshot |
| 2006-10-27 |
jm | ps and echo_noise in fixed-point. |
commit | commitdiff | tree | snapshot |
| 2006-10-27 |
jm | Now using fixed-point FFTs in the preprocessor. |
commit | commitdiff | tree | snapshot |
| 2006-10-26 |
giles | Remove relative path references to libspeex which cause... |
commit | commitdiff | tree | snapshot |
| 2006-10-26 |
giles | Merge updates to the autogen.sh script from other packa... |
commit | commitdiff | tree | snapshot |
| 2006-10-26 |
jm | debugged spx_ilog2() and added spx_ilog4() |
commit | commitdiff | tree | snapshot |
| 2006-10-26 |
jm | Implemented exponential function (spx_exp) in fixed... |
commit | commitdiff | tree | snapshot |
| 2006-10-26 |
jm | The World Institute for Mental Health recommends switch... |
commit | commitdiff | tree | snapshot |
| 2006-10-25 |
jm | added speex_echo_cancellation() without the last param... |
commit | commitdiff | tree | snapshot |
| 2006-10-25 |
jm | Oops, re-enabling kissfft as default |
commit | commitdiff | tree | snapshot |
| 2006-10-25 |
jm | Quick and dirty VAD replacement for now. |
commit | commitdiff | tree | snapshot |
| 2006-10-24 |
jm | choose between linear and Bark scale computation of... |
commit | commitdiff | tree | snapshot |
| 2006-10-24 |
jm | Gain floor is now computed on the Bark scale. Some... |
commit | commitdiff | tree | snapshot |
| 2006-10-23 |
jm | some tuning |
commit | commitdiff | tree | snapshot |
| 2006-10-23 |
jm | cleanup again |
commit | commitdiff | tree | snapshot |
| 2006-10-23 |
jm | code cleanup |
commit | commitdiff | tree | snapshot |
| 2006-10-22 |
jm | Now possible to control the amount of noise and echo... |
commit | commitdiff | tree | snapshot |
| 2006-10-21 |
jm | tuning and cleanup |
commit | commitdiff | tree | snapshot |
| 2006-10-20 |
jm | Mixing linear scale and bark scale gains, plus some... |
commit | commitdiff | tree | snapshot |
| 2006-10-20 |
jm | making sure all global variables are const (for Symbian... |
commit | commitdiff | tree | snapshot |
| 2006-10-11 |
jm | Added Bark filter bank. Need to make the Bark/Mel choic... |
commit | commitdiff | tree | snapshot |
| 2006-10-11 |
jm | Some references for the black magic |
commit | commitdiff | tree | snapshot |
| 2006-10-11 |
jm | Can you believe I just commented the preprocessor code! |
commit | commitdiff | tree | snapshot |
| 2006-10-11 |
jm | adaptive update window |
commit | commitdiff | tree | snapshot |
| 2006-10-11 |
jm | replaced frame probability (Pframe) |
commit | commitdiff | tree | snapshot |
| 2006-10-11 |
jm | actually use the filter bank properly (well, sort of) |
commit | commitdiff | tree | snapshot |
| 2006-10-10 |
jm | Big preprocessor cleanup. Implemented mel filterbank... |
commit | commitdiff | tree | snapshot |
| 2006-10-10 |
jm | Fixed a bug on 16-bit platforms (thanks Jim) |
commit | commitdiff | tree | snapshot |
| 2006-10-04 |
jm | Ouch! Fixed a really bad uninitialised memory (well... |
commit | commitdiff | tree | snapshot |
| 2006-09-08 |
jm | Adding the "f" suffix to float constants (path by Micha... |
commit | commitdiff | tree | snapshot |
| 2006-09-04 |
jm | version string with no dash Speex-1.2beta1 |
commit | commitdiff | tree | snapshot |
| 2006-08-31 |
jm | Tuning for low-energy echo. |
commit | commitdiff | tree | snapshot |
| 2006-08-31 |
jm | Increase learning rate for some mis-adaptated condition... |
commit | commitdiff | tree | snapshot |
| 2006-08-31 |
jm | should recover better from no initial adaptation. |
commit | commitdiff | tree | snapshot |
| 2006-08-30 |
jm | more replacements of int with spx_int32_t |
commit | commitdiff | tree | snapshot |
| 2006-08-30 |
jm | minor tuning |
commit | commitdiff | tree | snapshot |
| 2006-08-28 |
jm | Changed part of the API to use spx_int32_t instead... |
commit | commitdiff | tree | snapshot |
| 2006-08-23 |
jm | forgot to free memory for low_innov. sb_celp is now... |
commit | commitdiff | tree | snapshot |
| 2006-08-22 |
jm | removed unused variable |
commit | commitdiff | tree | snapshot |
| 2006-08-17 |
jm | libtool version info |
commit | commitdiff | tree | snapshot |
| next |
xiph / speexdsp.git / shortlog