Welcome to mirror list, hosted at ThFree Co, Russian Federation.

gitlab.xiph.org/xiph/opus.git - Unnamed repository; edit this file 'description' to name the repository.
summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2016-07-04Adds some smoothing to the energy quantizationexp_lbr_tune9Jean-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.
2016-06-29Tune SILK rate for hybrid with FEC enabledJean-Marc Valin
2016-06-29Moving FEC decision to the Opus encoder level (from SILK level)Jean-Marc Valin
2016-06-29Taking loss rate into account in compute_equiv_rate()Jean-Marc Valin
2016-06-29Retuning the tf_analysis() lambdaJean-Marc Valin
Increasing the value at low rate seems to help a bit.
2016-06-29Fixes equiv_rate for CBRJean-Marc Valin
2016-06-29Minor TF cleanup (tf_sum was useless), plus commentsJean-Marc Valin
2016-06-29Use SPREAD_AGGRESSIVE on non-transient hybrid framesJean-Marc Valin
2016-06-29Don't use too much temporal resolution on hybrid frames at low rateJean-Marc Valin
Otherwise, we risk having "temporal holes" in the HF that anti-collapse can't always fill in.
2016-06-29Disable patch_transient_decision() on hybrid since it's causing artifactsJean-Marc Valin
It would trigger on the second frame of "S"s, causing holes in the spectrum
2016-06-29CBR: set gains to their previons values where we're busting the budgetJean-Marc Valin
2016-06-29Set pulses to zero if we can't meet rate targetJean-Marc Valin
2016-06-29Now that we're preventing huge gains, we no longer need to pretend there'sJean-Marc Valin
no dither
2016-06-29CBR: lock the gain on a subframe when the number of pulses stops going downJean-Marc Valin
2016-06-29merge back in the LTP gain safety limiterKoen Vos
2016-06-29CBR: Using the same gain increase for fixed-point as for floatJean-Marc Valin
2016-06-29Oops, edited the wrong comment in the previous commitJean-Marc Valin
2016-06-29Re-tuning compute_equiv_rate()Jean-Marc Valin
2016-06-29Getting rid of the SILK penalty for hybrid CBR.Jean-Marc Valin
That was likely a mistake from a misunderstanding on the SILK bit allocation
2016-06-29Improving CBR for hybrid modeJean-Marc Valin
Hybrid CBR now simply forces the SILK CBR rate to the "target" value. Also, we're getting rid of the -2 kb/s offset for CBR, which appeared to be harmful. In the case of hybrid constrained VBR, the cap is computed in the same way as the target.
2016-06-29Making the encoder more aggressive at using wider bandwidthsJean-Marc Valin
2016-06-29comment about reduced codelength component in LTP quantizer error metricKoen Vos
2016-06-29avoid double if statement for finding best quantized LTP coefsKoen Vos
2016-06-29increase NSQ decision delay from 32 to 40Koen Vos
2016-06-29reduced noise floor in noise shaping analysis by 2 dBKoen Vos
2016-06-29increase order of noise shaping filterKoen Vos
2016-06-29more conservative scaling of LTP corrs; remove assert that checks for ↵Koen Vos
negative residual energy; discard CB entries leading to such negative energies
2016-06-29Clean up: alignment of commentsKoen Vos
2016-06-29Clean up: replace tabs by spacesKoen Vos
2016-06-29removed prefilterKoen Vos
The NSQ SSE optimizations are disabled for now because they need to be updated
2016-06-29slight clean upKoen Vos
2016-06-29undo some loop unrollingKoen Vos
2016-06-29pitch analysis (in FIX) now scales the input down at the start, instead of ↵Koen Vos
at every stage
2016-06-29minor clean up of pointer offsetsKoen Vos
2016-06-29simplified computation of LTP coefsKoen Vos
2016-06-29Attenuate SILK PLC gain only for unvoiced speechFelicia Lim
2016-06-29We should now be able to be less conservative when using CBRJean-Marc Valin
2016-06-29Makes the encoder more aggressive about meeting the rate targetJean-Marc Valin
2016-06-29Using more fine-grained complexity settings for SILKJean-Marc Valin
Only complexity 0-1 now uses the plain non-delayed-decision NSQ since its quality is pretty bad. Previously 0-3 were using that plain NSQ.
2016-06-29Fixing bandwidth selection hysteresisJean-Marc Valin
Previously, the bandwidth detection could (e.g.) change the bandwidth from fullband to superwideband, and the hysteresis would then cause bandwidth to be stuck in superwideband.
2016-06-29Boosting the high-band bitrate on transients for hybrid modeJean-Marc Valin
2016-06-29fix potential overflowKoen Vos
2016-06-29Fixes signed integer overlof in silk_ADD_POS_SAT32()Jean-Marc Valin
Removes unused 64-bit version
2016-06-29Turn on DTX only when the noise is sufficiently quietFelicia Lim
Computes a pseudo SNR as the ratio between the running peak signal energy to instantaneous noise energy. DTX is turned on only if this is higher than a pre-determined threshold. Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>
2016-06-29New DTX that works in all modes (SILK/CELT/HYBRID)Felicia Lim
A frame is marked as valid for DTX if it contains noise or only digital silence. As before, there is an overhang period of 200 ms and a maximum consecutive DTX period of 400 ms. If the new DTX cannot be used because of the complexity setting and sampling frequency chosen, the SILK DTX will be used instead. Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>
2016-06-29Skip analysis if the frame contains only digital silenceFelicia Lim
Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>
2016-06-29NLSF_VQ_MAX_SURVIVORS is no longer usefulJean-Marc Valin
2016-06-29take advantage of more efficient NLSF quantization by reducing number of ↵Koen Vos
survivors in search JMV: edited to be a little more conservative
2016-06-29NLSF_del_dec_quant minor clean upKoen Vos
2016-06-29NLSF_encode now also uses NLSF weights from tableKoen Vos
generated by cgit v1.2.3 (git 2.25.1) at 2024-06-13 12:00:29 +0300