1 Ann Hell Ex Machina Release Notes
2 =================================
7 * Uses `ffmpeg' as a fallback converter in case all others fail.
8 * New PulseAudio output driver.
9 * Very big internal changes; instead of having many different event
10 lists (raw song, softsynth or midi), all events are the same (still kept
11 in different lists, though). This simplify the internal storage greatly.
12 This is not a user-visible change.
13 * Other minor changes to the C API were done, as using const wherever
14 possible; as above, no user-visible changes.
15 * Random volumes are no longer architecture-dependent (Bug #1146).
20 * The volume (`v') command can now have two arguments; next volumes will
21 be set as a random value from that range.
22 * New softsynth directive `attack', to set the attack time for notes,
23 analogous to `sustain' (this directive was introduced in 1.0.9, but
24 was left undocumented here).
25 * The `wav' command now accepts a file name starting with a pipe
26 symbol (|), interpreting the string as a command to be executed, that
27 will end up creating the .wav file that will be used as a sound source.
28 The command must include a %s where the command wants the output file
29 (it will be filled with a unique, temporary name).
34 * The wobble effect now accepts a 'gain' argument, to control the amount
35 of effect applied to the signal. If not set, it uses a default of 0.8
36 (previous versions acted as if a 1.0 value was set).
37 * The way of ending a softsynth song has changed; instead of waiting for
38 the end of events and then abruptly ending (which led to adding fake
39 silences and notes until all sound is off), now the instruments are
40 monitored and only exits when no more sound is generated.
41 * Three new distortion effects, `atan', `distort' and `overdrive' have
43 * A new special mark is supported, START, that can be defined anywhere
44 and used as the starting point for the following new tracks.
45 * A new special mark is support, END, that will be automatically set
46 to the longest time seen on a track. So, combining the START and END
47 marks, it's possible to mix different songs to play one after the other,
53 /* move to the furthest point */
56 /* set start of new tracks there */
66 * Big rename; main executable has been changed from 'annhell' to 'ahxm',
67 the personal sample library changed from $HOME/annhell to $HOME/ahxmlib,
68 and the temporary cache moved to /tmp/ahxm-$PID.
69 * Paged wave loading now include an adaptive algorithm that sequentially
70 increases the memory block assigned to that wave if it pages frequently.
71 This way it's much less possible to be bitten by bug #1108.
72 * New command 'A' to set the key signature as a list of alterations
74 * New effect 'half_wobble'; same as 'wobble', but letting pass only the
75 first pass of a period.
76 * New directive 'portamento', allowing notes to slide up or down.
77 * New distortion effect 'foldback'.
82 * Number of channels no longer default to 2, but to all the channels used
83 in the song (this way, WAV files are automatically generated with all the
84 defined channels and not clipped to 2). Other output devices are still
85 silently clipped to 2 and the rest of channels dropped.
86 * The arpeggiator subcommand @ (track) has been changed to /, as it was
87 confusing the parser. Also, the track is relative instead of absolute
88 (now it means a track offset), so it can be reused.
89 * Midiin has been fixed, as it was severely broken (no initializations).
90 * OSS's fragment changed so that real-time applications (as midiin) work
91 again. This seems to be necessary when using ALSA's OSS emulation.
92 * A new midi2ahs.pl script, to crudely convert MIDI files to .ahs scripts.
97 * Minor internal changes.
98 * First public version.
103 * Channel spreading of wave files into instruments can now be partially
104 controlled with the optional "first channel" and "skip channels"
105 parameters. This allow using samples where the left and right channel
106 are in different wave files.
107 * New randomized blocks.
112 * A new directive, 'song_info', to add author and name metadata to a song.
113 * A new command line option, -C, to generate a .cue file.
114 * Fixed a critical bug when loading .wav files directly (instead of
115 converted from other formats as .flac).
120 * This release allows working with huge wave files, as they are
121 loaded page by page instead of full in memory (but beware of
127 * This release includes no new features, but just some internal
128 refactoring (towards supporting paged instruments).
133 * Tempo changes has been fixed.
134 * New '-P' command line switch to show time progress.
139 * The performance of the 'skip seconds' option has been dramatically
140 improved; but take note that output will sound differently, as notes
141 that started playing before the time boundary will no longer sound
142 (this will also happen with a real MIDI synthesizer, where note on
143 events will never be sent).
148 * First stable version.