This commit adds some actions for creating and cycling through tracks.
set_octave_and_ensure_note_change() was refactored to allow switching
tracks to implement the same behaviour.
KnobsWidget is getting pretty bad.
This commit adds multi-track functionality without exposing it to the
user.
All I really did was rename AudioEngine to Track and allow more than one
Track in TrackManager. A lot of the changes are just changing widgets to
take a TrackManager and use current_track().
The TrackManager creates Tracks and gives them a read-only reference to
the global time value. When the TrackManager wants to fill a sample in
the buffer (in fill_buffer()), it calls fill_sample() on each Track.
The delay code is slightly different - a Track will fill its
m_delay_buffer with the sample it just created rather than the most
recent sample in the buffer (which used to be the same thing).
TrackManager manages the current octave.
Other than those few things, this is a pretty basic separation of
concerns.
This patch allows roll notes to be of different sizes. This necessitates
a new internal representation of time. BPM and time signatures are
mostly implemented but not exposed.
Roll notes are now sample-accurate and the grid is aligned to 60 BPM
4/4. The roll is divided by the time signature raised to some power of
2, giving the musical divisions of (in the case of 4/4) 16, 32, 64 etc.
Before, our timing was derived from the buffer size and we relied on
that to implement delay. Delay has been rewritten to be sample-granular.
It's now exposed as the proper "divisions of a beat".
Something to be wary of is that the last buffer in the loop is also used
for the start of the next loop. In other words, we loop mid-buffer. This
means we write WAVs with a tiny bit of silence due to breaking the loop
after filling half a buffer.
The data structure for the roll is an array of SinglyLinkedLists of
RollNotes. Separating by pitch (via the array layout) makes insertion
much simpler and faster. Using sorted lists (and thus
SinglyLinkedListIterators) to do lookups is very quick as you know the
sample of the next note and can just compare it to the current sample. I
implemented this with HashMaps and the cost of lookups was abysmal. I
also tried a single SinglyLinkedList and the insertion code got even
more complicated than it already is.
I started adding things to a Draw namespace, but it somehow felt really
wrong seeing Draw::Rect and Draw::Bitmap, etc. So instead, let's rename
the library to LibGfx. :^)
Notice that we are calculating release time according to the level when
the note is turned off rather than the sustain level. Naively using the
sustain level gives very long release times if you turn the note off
during attack, whereas this deterministically gives the same release
time.
1. Make decay sample-granular rather than buffer-granular
You only have ~43 buffers per second which can make a jagged signal.
2. Calculate decay in milliseconds
Decay is supposed to be a time value.
Now you can change the defaults in AudioEngine and not be totally
confused.
1. The default for m_octave_knob was actually upside-down.
2. The default for m_decay_knob wasn't respecting AudioEngine.
3. The default for m_delay_knob wasn't respecting AudioEngine.
Goals:
- Switch to a more typical LibGUI arrangement
- Separate GUI (MainWidget) and audio (AudioEngine)
- Improve on existing features while retaining the same feature set
Improvements:
- Each GUI element is a separate widget
- The wave (WaveWidget) scales with the window
- The piano roll (RollWidget) scales horizontally and scrolls vertically
- The piano (KeysWidget) fits as many notes as possible
- The knobs (KnobsWidget) are now sliders
- All mouse and key events are handled in constant time
- The octave can be changed while playing notes
- The same note can be played with the mouse, keyboard and roll at the
same time, and the volume of the resulting note is scaled accordingly
- Note frequency constants use the maximum precision available in a
double
