*2016-08-26* |
Tristan Matthews | pvq: use small LUT for integer sqrt((n+3)/2) sqrt(... |
blob | commitdiff |

*2016-08-26* |
Tristan Matthews | pvq: use od_beta_rcp in od_gain_compand |
blob | commitdiff | diff to current |

*2016-08-25* |
Tristan Matthews | pvq: use od_rcp_beta in od_compute_k |
blob | commitdiff | diff to current |

*2016-08-24* |
Tristan Matthews | pvq: use od_rcp_beta in od_compute_max_theta |
blob | commitdiff | diff to current |

*2016-08-23* |
Tristan Matthews | pvq: use int beta in od_pow |
blob | commitdiff | diff to current |

*2016-08-19* |
Tristan Matthews | pvq: keep beta int until used in computation |
blob | commitdiff | diff to current |

*2016-08-19* |
Tristan Matthews | pvq: fix documentation |
blob | commitdiff | diff to current |

*2016-08-19* |
Tristan Matthews | pvq: convert beta tables to int |
blob | commitdiff | diff to current |

*2016-08-18* |
Tristan Matthews | pvq: use od_log2 int approximation |
blob | commitdiff | diff to current |

*2016-08-17* |
Tristan Matthews | pvq: use approximation for od_rcp |
blob | commitdiff | diff to current |

*2016-08-11* |
Tristan Matthews | pvq: fix build for OD_FLOAT_PVQ |
blob | commitdiff | diff to current |

*2016-08-11* |
Tristan Matthews | pvq: fixed-point od_apply_householder |
blob | commitdiff | diff to current |

*2016-08-11* |
Tristan Matthews | pvq: convert compute_theta to fixed |
blob | commitdiff | diff to current |

*2016-08-04* |
Tristan Matthews | pvq: convert double constant to int |
blob | commitdiff | diff to current |

*2016-08-01* |
Tristan Matthews | pvq: convert od_log2 to int16_t input |
blob | commitdiff | diff to current |

*2016-08-01* |
Tristan Matthews | pvq: use integer approximation for exp2 |
blob | commitdiff | diff to current |

*2016-07-27* |
Tristan Matthews | pvq: change shift to multiply for clarity |
blob | commitdiff | diff to current |

*2016-07-27* |
Tristan Matthews | pvq: convert compute_max_theta to fixed (aside from... |
blob | commitdiff | diff to current |

*2016-07-21* |
Tristan Matthews | pvq: use od_sqrt in od_pvq_compute_gain |
blob | commitdiff | diff to current |

*2016-07-21* |
Tristan Matthews | pvq: convert !noref branch of od_compute_k to integer |
blob | commitdiff | diff to current |

*2016-07-20* |
Tristan Matthews | pvq: convert 1st part of compute_k to integer (aside... |
blob | commitdiff | diff to current |

*2016-07-20* |
Tristan Matthews | pvq: fix return type for od_sqrt |
blob | commitdiff | diff to current |

*2016-07-18* |
Tristan Matthews | pvq: avoid double division in od_gain_compand |
blob | commitdiff | diff to current |

*2016-07-15* |
Nathan E. Egge | Remove compiler warnings with --enable-float-pvq. |
blob | commitdiff | diff to current |

*2016-07-02* |
Tristan Matthews | pvq: use od_pow() when beta != 1 and beta != 1.5 |
blob | commitdiff | diff to current |

*2016-07-01* |
Tristan Matthews | pvq: use rounded exp2 and log2 when raising g to 1... |
blob | commitdiff | diff to current |

*2016-06-01* |
Tristan Matthews | pvq: use integer approximation of sqrt |
blob | commitdiff | diff to current |

*2016-06-01* |
Tristan Matthews | od_rsqrt: s only depends on log2(x) and inshift |
blob | commitdiff | diff to current |

*2016-06-01* |
Tristan Matthews | od_sqrt: s only depends on log2(x) and the input shift |
blob | commitdiff | diff to current |

*2016-04-07* |
Tristan Matthews | pvq: use int16_t for rsqrt calculations |
blob | commitdiff | diff to current |

*2016-04-02* |
Jean-Marc Valin | New k-tokenizer based on recursively splitting vectors |
blob | commitdiff | diff to current |

*2016-03-31* |
Tristan Matthews | pvq: use float rounded version of sqrt(cg) |
blob | commitdiff | diff to current |

*2016-03-31* |
Tristan Matthews | pvq: update parameter doc for od_pvq_compute_gain |
blob | commitdiff | diff to current |

*2016-03-25* |
Tristan Matthews | pvq: add --enable-float-pvq option to configure |
blob | commitdiff | diff to current |

*2016-03-14* |
Tristan Matthews | pvq: convert non activity masking case to integer |
blob | commitdiff | diff to current |

*2016-03-12* |
Tristan Matthews | pvq: use integer approximation of 1/sqrt(yy) |
blob | commitdiff | diff to current |

*2016-03-12* |
Tristan Matthews | pvq: use float rounded version of 1/sqrt(yy) |
blob | commitdiff | diff to current |

*2016-02-18* |
Tristan Matthews | pvq: use integer approximations of cos and sin |
blob | commitdiff | diff to current |

*2016-02-15* |
Tristan Matthews | pvq: change OD_TRIG_SCALE to 32768, saturate to 32767 |
blob | commitdiff | diff to current |

*2016-02-11* |
Jean-Marc Valin | PVQ: od_pvq_sin() and od_pvq_cos() now output an integer |
blob | commitdiff | diff to current |

*2016-02-11* |
Jean-Marc Valin | PVQ: Adds some missing casts |
blob | commitdiff | diff to current |

*2016-02-11* |
Jean-Marc Valin | PVQ: Change PVQ code to use int32_t as companded gain... |
blob | commitdiff | diff to current |

*2016-02-10* |
Jean-Marc Valin | PVQ: Minor type fix (no change in behaviour) |
blob | commitdiff | diff to current |

*2016-02-10* |
Jean-Marc Valin | PVQ: change the scaling of companded gain variables... |
blob | commitdiff | diff to current |

*2016-02-10* |
Jean-Marc Valin | PVQ: Quantize companded gain |
blob | commitdiff | diff to current |

*2016-02-10* |
Jean-Marc Valin | PVQ: Theta values converted to integer |
blob | commitdiff | diff to current |

*2016-02-10* |
Jean-Marc Valin | PVQ: Changing the scale of theta to that of a 16-bit... |
blob | commitdiff | diff to current |

*2016-02-10* |
Jean-Marc Valin | PVQ: Quantizing input and output of sin and cos |
blob | commitdiff | diff to current |

*2016-02-09* |
Jean-Marc Valin | PVQ: Extra comments about the fixed-point code |
blob | commitdiff | diff to current |

*2016-02-08* |
Jean-Marc Valin | PVQ: Changing accumulations to 32-bit int in od_apply_h... |
blob | commitdiff | diff to current |

*2016-02-08* |
Jean-Marc Valin | PVQ: Removes all double-precision calculations from... |
blob | commitdiff | diff to current |

*2016-02-08* |
Jean-Marc Valin | PVQ: Removes double-precision Householder reflection |
blob | commitdiff | diff to current |

*2016-02-08* |
Jean-Marc Valin | PVQ: Using 16-bit Householder in od_pvq_synthesis_partial() |
blob | commitdiff | diff to current |

*2016-02-08* |
Jean-Marc Valin | PVQ: QM scaling in od_pvq_synthesis_partial() in fixed... |
blob | commitdiff | diff to current |

*2016-02-08* |
Jean-Marc Valin | PVQ: Converting ypulse[] scaling to integer in od_pvq_s... |
blob | commitdiff | diff to current |

*2016-02-08* |
Jean-Marc Valin | PVQ: Linear gains changed to integer |
blob | commitdiff | diff to current |

*2016-02-08* |
Jean-Marc Valin | PVQ: Quantized gain |
blob | commitdiff | diff to current |

*2016-02-08* |
Jean-Marc Valin | PVQ: Perform PVQ search on 16-bit input |
blob | commitdiff | diff to current |

*2016-02-08* |
Jean-Marc Valin | PVQ: Use x16 as Householder input |
blob | commitdiff | diff to current |

*2016-02-08* |
Jean-Marc Valin | PVQ: Quantizing the Householder reflection vector |
blob | commitdiff | diff to current |

*2016-02-08* |
Jean-Marc Valin | PVQ: Quantizing the input and reference to 16 bits... |
blob | commitdiff | diff to current |

*2016-01-21* |
Tristan Matthews | pvq: remove logging |
blob | commitdiff | diff to current |

*2015-12-15* |
Tristan Matthews | pvq: improve packing |
blob | commitdiff | diff to current |

*2015-12-13* |
Tristan Matthews | pvq: fix integer overflow |
blob | commitdiff | diff to current |

*2015-12-11* |
Tristan Matthews | pvq: remove od_apply_qm |
blob | commitdiff | diff to current |

*2015-11-01* |
Jean-Marc Valin | move PVQ state to pvq.h |
blob | commitdiff | diff to current |

*2015-10-02* |
Tristan Matthews | pvq: drop unused define |
blob | commitdiff | diff to current |

*2015-08-20* |
Jean-Marc Valin | Adds the option of using the Thor deblocking filter... |
blob | commitdiff | diff to current |

*2015-07-28* |
Nathan E. Egge | Fix od_qm_get_index function. |
blob | commitdiff | diff to current |

*2015-07-27* |
Jean-Marc Valin | Using 4-point lapping for all block sizes |
blob | commitdiff | diff to current |

*2015-07-15* |
Tristan Matthews | encode: move QMs to where they are used |
blob | commitdiff | diff to current |

*2015-07-09* |
Ian Kronquist | Remove dependency on libogg for core daala |
blob | commitdiff | diff to current |

*2015-07-07* |
Tristan Matthews | pvq: improve some QM names |
blob | commitdiff | diff to current |

*2015-07-05* |
Tristan Matthews | pvq: add comment explaining 1 vs. 1.5 for beta |
blob | commitdiff | diff to current |

*2015-07-04* |
Nathan E. Egge | Add magnitude of 64-point basis functions. |
blob | commitdiff | diff to current |

*2015-07-01* |
Tristan Matthews | pvq: update doc to match code |
blob | commitdiff | diff to current |

*2015-06-02* |
Jean-Marc Valin | Changing block size variables from ln to bs |
blob | commitdiff | diff to current |

*2015-05-21* |
Tristan Matthews | Signal in bitstream which QM is in use |
blob | commitdiff | diff to current |

*2015-05-12* |
Tristan Matthews | Signal in bitstream if activity masking is in use |
blob | commitdiff | diff to current |

*2015-03-23* |
Jean-Marc Valin | More fixes to address Tim's comments |
blob | commitdiff | diff to current |

*2015-03-23* |
Jean-Marc Valin | Adds quantization matrices |
blob | commitdiff | diff to current |

*2015-03-23* |
Jean-Marc Valin | Updating quantization matrices to take into account... |
blob | commitdiff | diff to current |

*2015-03-12* |
Nathan E. Egge | Adding a 32-point lapping filter. |
blob | commitdiff | diff to current |

*2015-02-10* |
Tristan Matthews | pvq: use od_ prefix for non-static functions |
blob | commitdiff | diff to current |

*2015-02-02* |
Jean-Marc Valin | Interpolating quantization matrices across rates |
blob | commitdiff | diff to current |

*2014-12-23* |
Jean-Marc Valin | Use the same quantizers with and without activity maski... |
blob | commitdiff | diff to current |

*2014-12-17* |
Jean-Marc Valin | Signaling the quantization matrix in the bitstream |
blob | commitdiff | diff to current |

*2014-12-02* |
Jean-Marc Valin | fixes some comments |
blob | commitdiff | diff to current |

*2014-12-02* |
Jean-Marc Valin | Moves pvq.c functions to encoder/decoder-specific files |
blob | commitdiff | diff to current |

*2014-12-02* |
Jean-Marc Valin | More simplifications to pvq_synthesis_partial(). |
blob | commitdiff | diff to current |

*2014-12-02* |
Jean-Marc Valin | More PVQ cleanup. |
blob | commitdiff | diff to current |

*2014-11-23* |
Jean-Marc Valin | Reducing the use of pow() by optmizing the beta=1 and... |
blob | commitdiff | diff to current |

*2014-11-23* |
Jean-Marc Valin | Use special cases for "copy zero" and "copy reference" |
blob | commitdiff | diff to current |

*2014-11-18* |
Jean-Marc Valin | RDO on last two pulses |
blob | commitdiff | diff to current |

*2014-11-18* |
Jean-Marc Valin | Adding some RDO cost to using the H/V predictor |
blob | commitdiff | diff to current |

*2014-11-14* |
Jean-Marc Valin | Merge branch 'remove_intra2' |
blob | commitdiff | diff to current |

*2014-11-13* |
Jean-Marc Valin | fixes for issue 512 |
blob | commitdiff | diff to current |

*2014-11-13* |
Jean-Marc Valin | Change ordering of CfL flip flag so that it's encoder... |
blob | commitdiff | diff to current |

*2014-11-11* |
randomguy | Add casts to avoid precision loss warnings under MSVC. |
blob | commitdiff | diff to current |

*2014-11-10* |
Nathan E. Egge | Use OD_LOG2 macro instead of log2(). |
blob | commitdiff | diff to current |

next |