Rewrite.

CHANGES.md

@@ -1,5 +1,25 @@
 ## Changes
 
+### Version 0.18 (Next Version)
+
+- This release involves a major rewrite and many changes.
+- 64-bit sample support is gone. All samples are now `f32`.
+- The 64-bit prelude presents a 32-bit interface now, with 64-bit internal state.
+- All types and traits with `32/64` suffix are replaced with the 32-bit version with the suffix removed.
+- Explicit SIMD support in block processing via the `wide` crate.
+- `no_std` support can be enabled by disabling the `std` feature.
+- Buffers for the `process` method have been rewritten. Both stack and heap allocation is supported.
+- Settings have been rewritten, to make them compatible with the `AudioUnit` system.
+- Waveshaping is done using a trait now. Remove `Shape::` prefix from shapes to migrate,
+  except for `AdaptiveTanh`, which is created using `AdaptiveTanh::new`.
+- The `Atan` shape was tweaked to return values in -1...1 while retaining a slope of 1 at the origin.
+- Asymmetric follow filters are now explicitly declared: `follow((a, r))` is now `afollow(a, r)`.
+- Functions corresponding to operators are now available as an option, for example, `A >> B` can now be written `pipe(A, B)`.
+  What used to be `pipe` etc. now has an `i` suffix, for example, `pipei`.
+- `rnd` function is now `rnd1`.
+- `hash` function is now `hash1`. Added new hash function `hash2`.
+- Wavetable oscillator now accepts an `Arc<Wavetable>` in the constructor.
+
 ### Version 0.17
 
 - `Wave32/64`: `silence` is now `zero`.

Cargo.toml

@@ -2,6 +2,7 @@
 name = "fundsp"
 description = "Audio processing and synthesis library."
 keywords = ["dsp", "audio", "synthesizer", "sound", "wave"]
+categories = ["multimedia::audio"]
 license = "MIT OR Apache-2.0"
 version = "0.17.1"
 authors = ["SamiPerttu <[email protected]>"]
@@ -12,83 +13,40 @@ edition = "2021"
 
 [dependencies]
 generic-array = "0.14.7"
-numeric-array = "0.5.2"
-num-complex = "0.4.5"
-rustfft = "6.2.0"
-realfft = "3.3.0"
-lazy_static = "1.4.0"
-tinyvec = { version = "1.6.0", features = ["alloc"] }
-rsor = "0.1.4"
-duplicate = "1.0.0"
+numeric-array = { version = "0.5.2", default-features = false }
 dyn-clone = "1.0.17"
+libm = "0.2.8"
+wide = "0.7.23"
+num-complex = { version = "0.4.6", default-features = false, features = ["libm"] }
+tinyvec = { version = "1.6.0", features = ["alloc"] }
+hashbrown = "0.14.5"
+microfft = { version = "0.6.0", features = ["size-32768"] }
+funutd = { version = "0.15.0", default-features = false }
+thingbuf = { version = "0.1.6", default-features = false, features = ["alloc"] }
+once_cell = { version = "1.19.0", default-features = false, features = ["race", "alloc"] }
 symphonia = { version = "0.5.4", optional = true, features = ["all"] }
-thingbuf = "0.1.5"
-funutd = "0.14.0"
-
-[features]
-default = ["files"]
-files = ["dep:symphonia"]
 
 [dev-dependencies]
-cpal = "0.15.3"
-anyhow = "1.0.82"
-plotters = "0.3.5"
+anyhow = "1.0.86"
 criterion = "0.5.1"
+cpal = "0.15.3"
+assert_no_alloc = "1.1.2"
+eframe = "0.27.2"
+plotters = "0.3.6"
 midi-msg = "0.5.0"
 midir = "0.9.1"
 read_input = "0.8.6"
-assert_no_alloc = "1.1.2"
-eframe = "0.27.2"
 rayon = "1.10.0"
 
+[features]
+default = ["std", "files"]
+std = []
+files = ["dep:symphonia"]
+
 [[bench]]
 name = "benchmark"
 harness = false
 
-[[example]]
-name = "beep"
-path = "examples/beep.rs"
-
-[[example]]
-name = "peek"
-path = "examples/peek.rs"
-
-[[example]]
-name = "sequence"
-path = "examples/sequence.rs"
-
-[[example]]
-name = "type"
-path = "examples/type.rs"
-
-[[example]]
-name = "plot"
-path = "examples/plot.rs"
-
-[[example]]
-name = "file"
-path = "examples/file.rs"
-
-[[example]]
-name = "grain"
-path = "examples/grain.rs"
-
-[[example]]
-name = "grain2"
-path = "examples/grain2.rs"
-
-[[example]]
-name = "network"
-path = "examples/network.rs"
-
-[[example]]
-name = "keys"
-path = "examples/keys.rs"
-
-[[example]]
-name = "optimize"
-path = "examples/optimize.rs"
-
 [package.metadata.docs.rs]
 all-features = true
 rustc-args = ["--cfg", "docsrs"]

FUTURE.md

@@ -10,7 +10,6 @@ This is a list of feature ideas for the future.
 - More physical models. Karplus-Strong exists already; figure out if it could be improved somehow.
 - Dynamic bypass wrapper that bypasses a node when input and output levels drop low enough.
 - Improve basic effects implemented in graph notation such as `reverb` (e.g., early reflections), `chorus`, `flanger` and `phaser`.
-- Some kind of parameter system that works with both `AudioNode` and `AudioUnit` systems. At its simplest, it could be a key-value system with a fixed set of keys.
 - More sound generators in the `gen` module.
 - Improve or replace the drum sounds in the library.
 - Real-time safe sound server that uses `cpal`. It could have a static set of read/write channels for rendering audio, including hardware channels.
@@ -19,3 +18,4 @@ This is a list of feature ideas for the future.
 - Expand `README.md` into a book.
 - Time stretching / pitch shifting algorithm.
 - FFT convolution engine and HRTF support.
+- Fading nodes in and out when replacing a node in `Net`.