Historique des commits

Auteur SHA1 Message Date
  Koen Vos 0d342f1efc fix bug in previous commit il y a 8 ans
  Koen Vos 6450837c44 merge back in the LTP gain safety limiter il y a 8 ans
  Koen Vos 93d946b55d comment about reduced codelength component in LTP quantizer error metric il y a 8 ans
  Koen Vos 5d05bbc4ea avoid double if statement for finding best quantized LTP coefs il y a 8 ans
  Koen Vos cce3d89421 increase NSQ decision delay from 32 to 40 il y a 8 ans
  Koen Vos 6b712cefed reduced noise floor in noise shaping analysis by 2 dB il y a 8 ans
  Koen Vos 5ee334e507 increase order of noise shaping filter il y a 8 ans
  Koen Vos 3b8a9b183c more conservative scaling of LTP corrs; remove assert that checks for negative residual energy; discard CB entries leading to such negative energies il y a 8 ans
  Koen Vos 6600e72f47 fix potential overflow il y a 8 ans
  Koen Vos b4778904c8 Clean up: alignment of comments il y a 9 ans
  Koen Vos 4d126dbc71 Clean up: replace tabs by spaces il y a 9 ans
  Koen Vos 1de1868d01 removed prefilter il y a 9 ans
  Koen Vos 952cb66b21 slight clean up il y a 9 ans
  Koen Vos 532f8613f5 undo some loop unrolling il y a 9 ans
  Koen Vos bb60799967 pitch analysis (in FIX) now scales the input down at the start, instead of at every stage il y a 9 ans
  Koen Vos 8e199ac225 minor clean up of pointer offsets il y a 9 ans
  Koen Vos 6a9ef699d2 simplified computation of LTP coefs il y a 9 ans
  Koen Vos eae137c695 faster Burg implementation il y a 9 ans
  Jean-Marc Valin da995f7d9f NLSF_VQ_MAX_SURVIVORS is no longer useful il y a 8 ans
  Koen Vos e047cce17f take advantage of more efficient NLSF quantization by reducing number of survivors in search il y a 9 ans
  Koen Vos 5d6a027b01 NLSF_del_dec_quant minor clean up il y a 9 ans
  Koen Vos 91a76cabf1 NLSF_encode now also uses NLSF weights from table il y a 9 ans
  Koen Vos 9c2866112e NLSF VQ now uses absolute error, and prediction il y a 9 ans
  Koen Vos aed679cd7b NLSF decoding now uses tables for NLSF weights il y a 9 ans
  Koen Vos 55023d471c sum_sqr_shift: reduced headroom (ie more preserved resolution); shift increments by one instead of two il y a 9 ans
  Jean-Marc Valin 8fc157a4be Simplifying fast_atan2f() il y a 8 ans
  Jean-Marc Valin 4ed368e68d Quality: Increase CELT rate for voiced frames in hybrid mode il y a 8 ans
  Jean-Marc Valin f587785a0d Better modeling of the effect of CBR and complexity il y a 8 ans
  Jean-Marc Valin 282a72a613 Re-enabling high-band attenuation for hybrid mode il y a 8 ans
  Jean-Marc Valin b5faa6d361 Quality: Retrained classifier MLP with better data il y a 9 ans