 Koen Vos
|
3938261b4c
slight clean up
|
9 年之前 |
|
Koen Vos
|
87b0592afe
undo some loop unrolling
|
9 年之前 |
|
Koen Vos
|
6313e8fc64
pitch analysis (in FIX) now scales the input down at the start, instead of at every stage
|
9 年之前 |
|
Koen Vos
|
4c2212db34
minor clean up of pointer offsets
|
9 年之前 |
|
Koen Vos
|
6954828bb0
simplified computation of LTP coefs
|
9 年之前 |
 Felicia Lim
|
52daddc00c
Attenuate SILK PLC gain only for unvoiced speech
|
8 年之前 |
 Jean-Marc Valin
|
25d34cbbfb
We should now be able to be less conservative when using CBR
|
8 年之前 |
|
Jean-Marc Valin
|
9e65fe1492
Makes the encoder more aggressive about meeting the rate target
|
8 年之前 |
|
Jean-Marc Valin
|
4ebf16b55c
Using more fine-grained complexity settings for SILK
|
8 年之前 |
|
Jean-Marc Valin
|
ae154850c0
Fixing bandwidth selection hysteresis
|
8 年之前 |
|
Jean-Marc Valin
|
b078eb2aee
Boosting the high-band bitrate on transients for hybrid mode
|
8 年之前 |
|
Koen Vos
|
58a6a2d6b2
fix potential overflow
|
8 年之前 |
|
Jean-Marc Valin
|
af8767ec73
Fixes signed integer overlof in silk_ADD_POS_SAT32()
|
8 年之前 |
|
Felicia Lim
|
2215595541
Turn on DTX only when the noise is sufficiently quiet
|
8 年之前 |
|
Felicia Lim
|
277c25aafe
New DTX that works in all modes (SILK/CELT/HYBRID)
|
8 年之前 |
|
Felicia Lim
|
ffe57eaaa0
Skip analysis if the frame contains only digital silence
|
8 年之前 |
|
Jean-Marc Valin
|
3bb0aafa6c
NLSF_VQ_MAX_SURVIVORS is no longer useful
|
8 年之前 |
|
Koen Vos
|
cc2e7f5d3d
take advantage of more efficient NLSF quantization by reducing number of survivors in search
|
9 年之前 |
|
Koen Vos
|
6cad0a2bf7
NLSF_del_dec_quant minor clean up
|
9 年之前 |
|
Koen Vos
|
8847a0661e
NLSF_encode now also uses NLSF weights from table
|
9 年之前 |
|
Koen Vos
|
868d1e66f0
NLSF VQ now uses absolute error, and prediction
|
9 年之前 |
|
Koen Vos
|
feaab8ae35
NLSF decoding now uses tables for NLSF weights
|
9 年之前 |
|
Koen Vos
|
98a92ac83f
sum_sqr_shift: reduced headroom (ie more preserved resolution); shift increments by one instead of two
|
9 年之前 |
|
Jean-Marc Valin
|
4023be5580
Simplifying fast_atan2f()
|
8 年之前 |
|
Jean-Marc Valin
|
f9c06c0eba
Quality: Increase CELT rate for voiced frames in hybrid mode
|
8 年之前 |
|
Jean-Marc Valin
|
6ac3b1bb19
Better modeling of the effect of CBR and complexity
|
8 年之前 |
|
Jean-Marc Valin
|
bf6b2d3624
Re-enabling high-band attenuation for hybrid mode
|
8 年之前 |
|
Jean-Marc Valin
|
093bd5c369
Quality: Retrained classifier MLP with better data
|
9 年之前 |
|
Jean-Marc Valin
|
34be34d69c
Using "hybrid" flag instead of "start!=0"
|
9 年之前 |
|
Jean-Marc Valin
|
a5e037fee4
Quality: Making the encoder more aggressive at using wider bandwidths
|
9 年之前 |
