eden-emu/opus
19
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
1375 commits 285 branches 59 tags 26 MiB
Commit graph

32 commits

Author SHA1 Message Date
Timothy B. Terriberry
ef2e650592 Add coarse energy entropy model tuning. 
This tunes the entropy model for coarse energy introduced in commit
 c1c40a76.
It uses a constant set of parameters, tuned from about an hour and a
 half of randomly selected test data encoded for each frame size,
 prediction type (inter/intra), and band number.
These will be slightly sub-optimal for different frame sizes, but
 should be better than what we were using.

For inter, this saves an average of 2.8, 5.2, 7.1, and 6.7 bits/frame
 for frame sizes of 120, 240, 480, and 960, respectively.
For intra, this saves an average of 1.5, 3.0, 4.5, and 5.3 bits/frame
 (for the same frame sizes, respectively).
 2010-11-09 17:54:41 +08:00
Jean-Marc Valin
1b36d6c8b6 Complexity setting now controls two-pass coarse energy quantization 2010-08-31 17:21:52 -04:00
Jean-Marc Valin
bb5288174d coarse probability model in static modes too 2010-08-25 22:12:18 -04:00
Jean-Marc Valin
5e7f02de2c Moving intra decision to quant_coarse_energy() 2010-08-08 09:48:22 -04:00
Jean-Marc Valin
504fb3c189 Using the real spectral means instead of the ones 
after inter-band prediction
 2010-08-06 15:56:22 -04:00
Jean-Marc Valin
bc272de74b Code simplifications for log->amplitude conversion 2010-08-02 09:41:31 -04:00
Jean-Marc Valin
617af25e5c Cleaning up intra_decision() 2010-07-23 16:54:45 -04:00
Jean-Marc Valin
ca6533cd88 Energy quantization tuning. 
Prediction coefficient now depends on the frame size. Also, the pdfs and means
have been adjusted and better quantized. This breaks compatibility of course.
 2010-07-23 16:10:36 -04:00
Jean-Marc Valin
6bf3b0a7a1 The coarse energy budget is no longer part of the bit-stream. 
It is now the encoder's responsability to take care of it to avoid
busting the budget.
 2010-07-19 14:32:40 -04:00
Jean-Marc Valin
525d7cfdc4 Support for adjusting the end band 2010-07-13 18:18:23 -04:00
Jean-Marc Valin
a01106587c Improved quality of small frame sizes at low bitrate. 
Adding a temporal energy floor to prevent extremely small values when there's
only one bin in the band.
 2010-06-23 15:22:54 -04:00
Jean-Marc Valin
bd0610d21b Increasing log energy resolution 2010-03-13 10:04:19 -05:00
Jean-Marc Valin
3a0bc3d95c Allowing CELT to skip the low frequencies 2010-02-25 22:46:55 -05:00
Jean-Marc Valin
234969c903 Removed the _t from all the celt*_t types to avoid clashing with POSIX 2009-10-17 22:12:42 -04:00
Jean-Marc Valin
8b2ff0da5a Updated copyright notices 2009-10-17 21:40:10 -04:00
Jean-Marc Valin
bf2398b049 first step for removing the number of channels from the mode 2009-10-15 07:28:19 -04:00
Jean-Marc Valin
7a56741b2e removing redundant calls to log2() 2009-06-13 23:02:54 -04:00
Jean-Marc Valin
52cb5fb3f6 Adding extra fine bits only when we have rounded down in the allocation 2009-06-10 08:08:55 -04:00
Jean-Marc Valin
24c25a23d1 Folding, coarse energy, source cleanup 2009-06-09 11:56:59 -04:00
Jean-Marc Valin
39710536fc This makes it possible to use the bits left after PVQ for fine energy. 2009-06-09 00:52:16 -04:00
Gregory Maxwell
888d8ce939 VBR support. VBR API and VBR support in celtenc. 2009-05-21 22:23:25 -04:00
Jean-Marc Valin
32ec58cc3e Dynamically selecting intra energy based on energy variations from the previous 
frame
 2009-05-01 21:28:58 -04:00
Jean-Marc Valin
18a3b79d24 Add support for intra-coding of the coarse energy. 2009-05-01 19:58:55 -04:00
Jean-Marc Valin
6775de3eae Unified allocation of fine energy and pulses. 2008-08-02 08:14:42 -04:00
Jean-Marc Valin
c890b58b69 Completed the separation of coarse and fine energy quantisation 2008-08-01 22:26:49 -04:00
Jean-Marc Valin
4ce9205f38 Optimisation: caching the divisions used in the Laplace encoder. 2008-04-23 13:42:10 +10:00
Jean-Marc Valin
5d56183404 fixed-point: log-energy for previous frame now a 16-bit value. This currently 
intruduces a bit of an encoder-decoder mismatch (Q8 in dB), but it'll be
reduced when the interals of quant_energy_mono() are properly converted to
fixed-point and oldEBands gets rounded instead of truncated.
 2008-02-27 11:59:05 +11:00
Jean-Marc Valin
e901fe35b8 fixed-point: added a celt_ener_t type for band energy. 2008-02-26 14:46:26 +11:00
Jean-Marc Valin
c9cc6d3e34 Introducing a (very) crude budget for the energy encoder. 2008-02-13 11:37:41 +11:00
Jean-Marc Valin
98d2a4917e No more cheating, everything fully quantised 2007-12-07 22:46:47 +11:00
Jean-Marc Valin
8143be3026 energy decoding partially done (cheating a bit) 2007-12-07 16:40:39 +11:00
Jean-Marc Valin
8b0137aa42 Quantisation of band energies (adding files) 2007-12-06 07:51:53 +11:00