summaryrefslogtreecommitdiffstats
path: root/libtoolame-dab/HISTORY
blob: ba53e4ad73268909a484e8b400881439ce5e6978 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
v0.1a	- initial release (7 Nov 1998)

v0.1b  	- released 1 Oct 99
	- filter subband and window subband from Lame3.x added
	- cleaned up some memory stuff in psy.c
	- a fast mode switch (-f) which turns off psychoacoustics
		-instead of having the psymodel calculate SNR 
		 just use a bunch of preset values.  I just made up 
		 some SNR values based upon my test file.  These 
		 could probably be tuned so that there's a set of SNR
		 values for rock, pop, classical, voice etc.

v0.1d	- released 14 Oct 99
	- cleaned up some code
		- removed decoding source code (use mpg123 instead :)
		- renamed subs.c to fft.c (which is mostly what it does :)
		- moved the filter/window_subband into subband.c
		- split encode.c into generic code (encode), LayerI code (encodeI)
			and layerII code (encodeII)
	- no longer use obtain_parameters(). interactive input is a pain

v0.1e	- released 15 Oct 99
	- cleaned up some code
		- defaults to 192kbps, psy model 1 with extension mp2 (see encoder.h)
	- added (-a) downmix to mono option (code in encode.c/get_audio() )
	- transplanted FHT-based fft() from LAME.
	- II_f_f_t() now calls fht(). heaps faster. (in tonal.c used by psymodel 1)
		This could be neater though.
	- PDS's II_scale_factor_calc() put in encodeII.c

v0.1f	- released 16 Oct 99
	- PDS's updated scale_factor_calc
	- integrated a few functions in tonal.c/II_psycho_one() which gave about a 15%
		speedup when using psymodel 1

v0.1g	- released 
	- tweaked tonal.c/threshold() 10-15% improvement in this routine.

v0.2a	- released 24 october 1999
	- changed base source code from mpegaudio.tar to the dist10/lsf/encoder.  This new code includes lower sampling frequency (LSF) extensions.
	- added in most of the changes already done.
	- LSF mode is faulty.  Seg faults after about 190 frames here.  LSF mode works if you turn off psy models though (-f)

v0.2b	- released 30 October 1999
	- added a "-q n" mode. will only calculate the psy model every n frames.  
		"-q 0" is equivalent to "-f". 
	- reimplemented "downmix to mono" -a switch
	- WAV file support from Nick Burch <gagravarr@SoftHome.net>
	- some code cleanup - removed more LayerI & III stuff.

v0.2c	- skipped this release

v0.2d 	- released 6 November 1999
	- used Mathew Hendry's enhanced fht() with table lookup (20% faster)
	- made add_db an inline function and tweaked it (25% faster)
	- code cleanup - renamed psy.c->psycho_II.c, tonal.c->psycho_I.c, other misc stuff
	- added a "no padding" (-r) switch. Needed by the rfa.org (thanks to Federico Grau)
	- tweaked threshold 33% speedup.

v0.2e	- released 14 December 1999
	- Some minor speedups from PDS in encode.c
	- Changed // comments to /* */ style
	- Merged some functions in subband.c
	- Cleaned up file input parsing a little. 
	- Now reads PCM from /dev/stdin if this is specified as input filename (kludgy but works for me)
	- Started putting tables into exe. (only enwindow done so far)

v0.2f	- released 15 April 2000
	- mainly cosmetic changes in the source and code layout
	- removed a lot of unused functions, and split common.c into more logical subumits.
	- about 15-20% faster when using psymodel I. (merged some functions in psycho_i.c)

v0.2g	- released 22 April 2000
	- added a VBR mode. See README.VBR
	- cleaned up more code.

v0.2h   - stdio/stdout now parse a little better using '-'
	- tables now included in the exe thanks to Oliver Lietz <lietz@nanocosmos.de>
		(I haven't had a chance to test it much though)
	- including the tables in the exe seems to have boosted speed quite a bit (25%?)
	- re-implemented byte-swapping in get_audio.c (command line switch -x )

v0.2i   - Released 13 July 2001
	- DAB extensions from Phillipe Jouguet <philippe.jouguet@vdldiffusion.com>
        - Added a switch to swap the input channels (can someone confirm that toolame
             swaps them incorrectly by default)

v0.2j	- Released 13 February 2003
	- Definitely LGPL
	- VBR mode is stable for all sampling frequencies. (See README.VBR for details)
	- speedups and cleanups from
		Phillipe Jouguet - philippe.jouguet at vdl.fr
			(spelling, LSF using psyII, WAVE reading)
		Henrik Herranen - leopold at vlsi.fi
			(WAVE reading)
		Andreas Neukoetter - anti at webhome.de
			(verbosity patch '-t' switch for transcode plugin)
		Sami Sallinen - sami.sallinen at g-cluster.com
			filter_subband loop unroll
			 psycho_i fix for "% 1408" calcs)

v0.2k	- Released 16 February 2003
	- Nicolas Croiset
		unlimited file read size for stdin
		fixed bitstream ending so that mp2 files can be concatenated
		DAB options now comes with a size argument
		(finally?) fixed the core dump in psycho_1
	- Ricardo Schelp ricardoschelp at arnet.com.ar
		combined window/filter subband routine. nice speed increase.

v0.2l	- Released 3 March 2003
	- Added new psychomodels
          -1  use precomputed values
           0  very low complexity
           1  same old psy model 1
           2  same old psy model 2
           3  psy model 1 rewritten from scratch
           4  rewritten psy model 2
	- Rewrote the bitstream encoding routines
        - Sped up the sin/cos/atan stuff in psymodel 4 (using tables now)
         
v0.2l with ZMQ output - February 2014
    - added outputzmq.{h, c}
      Matthias P. Braendli
    - integrate into FDK-AAC-DABplus in February 2016

mike (mikecheng at NOT planckenergy.com) [remove the NOT]