Age | Commit message (Collapse) | Author | |
---|---|---|---|
2016-07-25 | Fixes cap on gain in denormalise_bands()exp_overflow6 | Jean-Marc Valin | |
2016-07-25 | Fixes a shift<0 issue in transient_analysis() | Jean-Marc Valin | |
2016-07-25 | Revert "Fixes shift<-2 case for denormalise_bands()" | Jean-Marc Valin | |
This reverts commit b290bca91aeb047d03841bd1897c1cd0ca6d5344. | |||
2016-07-25 | Fix potential overflow in high-passed signal for transient detection | Jean-Marc Valin | |
2016-07-25 | Tighter bound in filter control in PLC | Jean-Marc Valin | |
We can't use the attenuation since it doesn't apply to the memory | |||
2016-07-25 | More headroom in PLC code | Jean-Marc Valin | |
extrapolation_len can be up to 960+120, so we we need a shift of 10 if the values are very large. | |||
2016-07-25 | Saturate MDCT output post-TDAC rather than pre- | Jean-Marc Valin | |
Gives us a tighter bound on the pitch postfilter input to avoid overflows | |||
2016-07-24 | Avoids reading beyond the current buffer in comb_filter() | Jean-Marc Valin | |
This could cause overflows when processing non-saturated TDAC values. | |||
2016-07-24 | Directly saturate the comb filter output | Jean-Marc Valin | |
Otherwise, the output can grow as large as SIG_SAT/(1-0.75), which can cause an overflow when adding two values (before multiplying by the filter tap). | |||
2016-07-22 | Avoiding an overflow in the LTP quantization | Jean-Marc Valin | |
Reduces the XX values to avoid an overflow in silk_VQ_WMat_EC(), even in floating-point. | |||
2016-07-22 | Disables MIPS silk_noise_shape_analysis_FIX() until it's updated | Jean-Marc Valin | |
2016-07-22 | More headroom in fixed-point silk_burg_modified() to avoid an overflow | Jean-Marc Valin | |
2016-07-22 | Increase headroom in silk_warped_autocorrelation_FIX() | Jean-Marc Valin | |
2016-07-22 | Increase headroom to avoid overflow in silk_pitch_analysis_core() | Jean-Marc Valin | |
2016-07-22 | PLC overflow fix | Jean-Marc Valin | |
2016-07-22 | Prevent overflows in PLC celt_iir() | Jean-Marc Valin | |
2016-07-22 | fixes overflows in celt_iir() by doing proper saturation | Jean-Marc Valin | |
2016-07-22 | getting rid of a PLC overflow (I think) | Jean-Marc Valin | |
2016-07-22 | post-filter saturation | Jean-Marc Valin | |
2016-07-22 | saturate MDCT output | Jean-Marc Valin | |
2016-07-22 | silencing overflows in MDCT and FFT | Jean-Marc Valin | |
2016-07-22 | Reduce risk of overflowing in mono downsampling | Jean-Marc Valin | |
Not observed yet (except when caused by a different bug), but "just in case" | |||
2016-07-22 | Avoid overflowing on ridiculously large energy | Jean-Marc Valin | |
2016-07-22 | Fixes shift<-2 case for denormalise_bands() | Jean-Marc Valin | |
2016-07-22 | Silence fixed point unused parameter warning | Mark Harris | |
2016-07-22 | test_opus_decode: Check all soft clip samples | Mark Harris | |
2016-07-22 | tests: Avoid shift into sign undefined behavior | Mark Harris | |
2016-07-21 | Avoids undefined behaviour from left-shifting negative values | Jean-Marc Valin | |
2016-07-20 | Fix use_dtx for DISABLE_FLOAT_API | Felicia Lim | |
Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org> | |||
2016-07-20 | Fix include warning | Tristan Matthews | |
2016-07-20 | Fix include warning | Tristan Matthews | |
2016-07-19 | Adds some smoothing to the energy quantizationexp_lbr_tune | Jean-Marc Valin | |
When the energy is stable, we slightly bias energy quantization towards the previous error to make the gain more stable (a constant offset is better than fluctuations). We reduce the bitrate by about 0.2% to 1% at low bitrate for the same quality. | |||
2016-07-19 | Tune SILK rate for hybrid with FEC enabled | Jean-Marc Valin | |
2016-07-19 | Moving FEC decision to the Opus encoder level (from SILK level) | Jean-Marc Valin | |
2016-07-19 | Taking loss rate into account in compute_equiv_rate() | Jean-Marc Valin | |
2016-07-19 | Retuning the tf_analysis() lambda | Jean-Marc Valin | |
Increasing the value at low rate seems to help a bit. | |||
2016-07-19 | Fixes equiv_rate for CBR | Jean-Marc Valin | |
2016-07-19 | Minor TF cleanup (tf_sum was useless), plus comments | Jean-Marc Valin | |
2016-07-19 | Use SPREAD_AGGRESSIVE on non-transient hybrid frames | Jean-Marc Valin | |
2016-07-19 | Don't use too much temporal resolution on hybrid frames at low rate | Jean-Marc Valin | |
Otherwise, we risk having "temporal holes" in the HF that anti-collapse can't always fill in. | |||
2016-07-19 | Disable patch_transient_decision() on hybrid since it's causing artifacts | Jean-Marc Valin | |
It would trigger on the second frame of "S"s, causing holes in the spectrum | |||
2016-07-19 | CBR: set gains to their previons values where we're busting the budget | Jean-Marc Valin | |
2016-07-19 | Set pulses to zero if we can't meet rate target | Jean-Marc Valin | |
2016-07-19 | Now that we're preventing huge gains, we no longer need to pretend there's | Jean-Marc Valin | |
no dither | |||
2016-07-19 | CBR: lock the gain on a subframe when the number of pulses stops going down | Jean-Marc Valin | |
2016-07-19 | merge back in the LTP gain safety limiter | Koen Vos | |
2016-07-19 | CBR: Using the same gain increase for fixed-point as for float | Jean-Marc Valin | |
2016-07-19 | Oops, edited the wrong comment in the previous commit | Jean-Marc Valin | |
2016-07-19 | Re-tuning compute_equiv_rate() | Jean-Marc Valin | |
2016-07-19 | Getting rid of the SILK penalty for hybrid CBR. | Jean-Marc Valin | |
That was likely a mistake from a misunderstanding on the SILK bit allocation |