 Jean-Marc Valin
|
2ad6a746cb
Adds some smoothing to the energy quantization
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
afd7cb77c7
Tune SILK rate for hybrid with FEC enabled
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
7ca418797c
Moving FEC decision to the Opus encoder level (from SILK level)
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
168b21f11e
Taking loss rate into account in compute_equiv_rate()
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
913f09e786
Retuning the tf_analysis() lambda
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
64739a6c06
Fixes equiv_rate for CBR
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
3b1df2a484
Minor TF cleanup (tf_sum was useless), plus comments
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
43e038d67f
Use SPREAD_AGGRESSIVE on non-transient hybrid frames
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
f4e2ca3b6f
Don't use too much temporal resolution on hybrid frames at low rate
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
879482d9b2
Disable patch_transient_decision() on hybrid since it's causing artifacts
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
173b997db2
CBR: set gains to their previons values where we're busting the budget
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
1a32f1bfaa
Set pulses to zero if we can't meet rate target
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
f77ad1f16d
Now that we're preventing huge gains, we no longer need to pretend there's
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
b3efe8646e
CBR: lock the gain on a subframe when the number of pulses stops going down
|
8 gadi atpakaļ |
 Koen Vos
|
eee09eca90
merge back in the LTP gain safety limiter
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
f3ad2473e9
CBR: Using the same gain increase for fixed-point as for float
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
3bc720ff08
Oops, edited the wrong comment in the previous commit
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
230e5f4e62
Re-tuning compute_equiv_rate()
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
65dcfd33b5
Getting rid of the SILK penalty for hybrid CBR.
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
7c8be0c68e
Improving CBR for hybrid mode
|
8 gadi atpakaļ |
|
Jean-Marc Valin
|
bf6dc52806
Making the encoder more aggressive at using wider bandwidths
|
8 gadi atpakaļ |
|
Koen Vos
|
0b4a8b64d7
comment about reduced codelength component in LTP quantizer error metric
|
8 gadi atpakaļ |
|
Koen Vos
|
35144f4d09
avoid double if statement for finding best quantized LTP coefs
|
8 gadi atpakaļ |
|
Koen Vos
|
2b29ef8b86
increase NSQ decision delay from 32 to 40
|
8 gadi atpakaļ |
|
Koen Vos
|
63e7173764
reduced noise floor in noise shaping analysis by 2 dB
|
8 gadi atpakaļ |
|
Koen Vos
|
91a2e30466
increase order of noise shaping filter
|
8 gadi atpakaļ |
|
Koen Vos
|
8f6098383e
more conservative scaling of LTP corrs; remove assert that checks for negative residual energy; discard CB entries leading to such negative energies
|
8 gadi atpakaļ |
|
Koen Vos
|
55b7934a67
Clean up: alignment of comments
|
9 gadi atpakaļ |
|
Koen Vos
|
d3f094b768
Clean up: replace tabs by spaces
|
9 gadi atpakaļ |
|
Koen Vos
|
303c10ecee
removed prefilter
|
9 gadi atpakaļ |
