vapylib 0.1.0__tar.gz

vapylib-0.1.0/LICENSE

@@ -0,0 +1 @@
+my vim doesnt have clipboard support so i cant copy gplv3 license text to here but it's a strong copyleft license. you can edit this program however you want, but you have to publish it with the same license.

vapylib-0.1.0/PKG-INFO

@@ -0,0 +1,13 @@
+Metadata-Version: 2.4
+Name: vapylib
+Version: 0.1.0
+Summary: a python voice synthesis lib
+License-Expression: GPL-3.0-only
+Requires-Python: >=3.10
+License-File: LICENSE
+Requires-Dist: numpy
+Requires-Dist: scipy
+Requires-Dist: sounddevice
+Requires-Dist: soundfile
+Requires-Dist: PyYAML
+Dynamic: license-file

vapylib-0.1.0/README.md

@@ -0,0 +1,16 @@
+# What is VAPyLib?
+VAPyLib (**V**oice **A**cting **Py**thon **Lib**rary) is a library that utilizes sine waves and filters to create singing voices using formant synthesis.
+# How to make songs for VAPyLib?
+To make a song, you need to create a `.yaml` or `.yml` file, and parse the song using the `parse_song(filepath)` function. Note that pitch can only be between `0.0` and `1.0`
+### Example `song.yaml`:
+```yaml
+version: 1
+syllables:
+    - "ma"
+    - "na"
+pitch_maps:
+    - 0.5,
+    - 0.7
+```
+## Note:
+VAPyLib is version 0.1.0, but put **1** instead.

vapylib-0.1.0/pyproject.toml

@@ -0,0 +1,12 @@
+[build-system]
+requires = ["setuptools"]
+build-backend = "setuptools.build_meta"
+[project]
+license = "GPL-3.0-only"
+name = "vapylib"
+version = "0.1.0"
+description = "a python voice synthesis lib"
+requires-python = ">=3.10"
+dependencies = ["numpy", "scipy", "sounddevice", "soundfile", "PyYAML"]
+[tool.setuptools]
+packages = ["vapylib"]

vapylib-0.1.0/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

vapylib-0.1.0/vapylib/__init__.py

@@ -0,0 +1 @@
+from .parser import parse_song

vapylib-0.1.0/vapylib/parser.py

@@ -0,0 +1,27 @@
+import yaml
+from . import synthesizer
+class Song:
+    def __init__(self, syllables, pitch_maps):
+        self.syllables = syllables
+        self.pitch_maps = pitch_maps
+    def play(self):
+        synthesizer.play(self)
+    def save(self, filename):
+        synthesizer.save_wav(self, filename)
+def parse_song(filepath):
+    with open(filepath, "r", encoding="utf8") as file:
+        data = yaml.safe_load(file)
+    if data.get("version") != 1:
+        raise ValueError("unsupported vapylib version")
+    syllables = data.get("syllables")
+    pitch_maps = data.get("pitch_maps")
+    if not isinstance(syllables, list):
+        raise ValueError("'syllables' must be a list")
+    if not isinstance(pitch_maps, list):
+        raise ValueError("'pitch_maps' must be a list")
+    if len(syllables) != len(pitch_maps):
+        raise ValueError("syllables and pitch_maps need to have equal amount of values")
+    for pitch in pitch_maps:
+        if not isinstance(pitch, (int, float)) or not 0 <= pitch <= 1:
+            raise ValueError(f"not good pitch ({pitch}), needs to be int or float and be between 0 and 1")
+    return Song(syllables, pitch_maps)

vapylib-0.1.0/vapylib/synthesizer.py

@@ -0,0 +1,107 @@
+import numpy as np
+from scipy import signal
+import sounddevice as sd
+import soundfile as sf
+def _pitch_to_frequency(pitch, min_hz=100, max_hz=800):
+    return min_hz + pitch * (max_hz - min_hz)
+_phonemes_list = [
+    "a", "e", "i", "o", "u",
+    "f", "s", "m", "n"
+]
+def _split_syllable(syllable):
+    phonemes = sorted(_phonemes_list, key=len, reverse=True)
+    result = []
+    position = 0
+    while position < len(syllable):
+        for phoneme in phonemes:
+            if syllable.startswith(phoneme, position):
+                result.append(phoneme)
+                position += len(phoneme)
+                break
+            else:
+                raise ValueError(f"not known phoneme at {syllable[position:]}")
+    return result
+#a = 800 1200 2300
+#e = 500 1500 2600
+#i = 250 2750 3195
+#o = 515 960 2695
+#u = 255 1200 1935
+_vowels = {
+    "a": [800, 1200, 2300],
+    "e": [500, 1500, 2600],
+    "i": [250, 2750, 3195],
+    "o": [515, 960, 2695],
+    "u": [255, 1200, 1935]
+}
+_fricatives = {
+    "f": {
+        "low": 1000,
+        "high": 8000
+    },
+    "s": {
+        "low": 4000,
+        "high": 12000
+    }
+}
+_nasals = {
+    "m": [250, 1000, 2200],
+    "n": [300, 1500, 2500]
+}
+_nasal_antiresonances = {
+    "m": [750, 1800],
+    "n": [1200, 2500]
+}
+def _apply_resonances(source, frequencies, sample_rate):
+    result = source
+    for frequency in frequencies:
+        if frequency >= sample_rate / 2:
+            continue
+        b, a = signal.iirpeak(frequency, Q=4, fs=sample_rate)
+        result = signal.lfilter(b, a, result)
+    return result
+def _apply_antiresonances(source, frequencies, sample_rate):
+    result = source
+    for frequency in frequencies:
+        if frequency >= sample_rate / 2:
+            continue
+        b, a = signal.iirnotch(frequency, Q=4, fs=sample_rate)
+        result = signal.lfilter(b, a, result)
+def _generate_source(pitch, sample_rate=44100):
+    frequency = _pitch_to_frequency(pitch)
+    sample_count = int(sample_rate * 0.5)
+    t = np.arange(sample_count) / sample_rate
+    return np.sin(2 * np.pi * frequency * t)
+def _generate_vowel(phoneme, pitch, sample_rate=44100):
+    source = _generate_source(pitch, 0.5, sample_rate)
+    return _apply_resonances(source, _vowels[phoneme], sample_rate)
+def _generate_fricative(phoneme, pitch, sample_rate=44100):
+    sample_count = int(sample_rate * 0.5)
+    source = np.random.normal(0, 1, sample_count)
+    return _apply_resonances(source, _fricatives[phoneme], sample_rate)
+def _generate_nasal(phoneme, pitch, sample_rate=44100):
+    source = _generate_source(pitch, 0.5, sample_rate)
+    return _apply_antiresonances(_apply_resonances(source, _nasals[phoneme], sample_rate), _nasal_antiresonances[phoneme], sample_rate)
+def _generate_audio(song, sample_rate=44100):
+    audio = []
+    for syllable, pitch in zip(song.syllables, song.pitch_maps):
+        syllable_audio = []
+        for phoneme in _split_syllable(syllable):
+            if phoneme in _vowels:
+                sound = _generate_vowel(phoneme, pitch, sample_rate=sample_rate)
+            elif phoneme in _fricatives:
+                sound = _generate_fricative(phoneme, pitch, sample_rate=sample_rate)
+            elif phoneme in _nasals:
+                sound = _generate_nasal(phoneme, pitch, sample_rate=sample_rate)
+            else:
+                continue
+            syllable_audio.append(sound)
+        if syllable_audio:
+            audio.extend(np.concatenate(syllable_audio))
+    return np.array(audio, dtype=np.float32)
+def play(song):
+    audio = _generate_audio(song)
+    sd.play(audio, samplerate=44100)
+    sd.wait()
+def save_wav(song, filename):
+    audio = _generate_audio(song)
+    sf.write(filename, audio, 44100)

vapylib-0.1.0/vapylib.egg-info/PKG-INFO

@@ -0,0 +1,13 @@
+Metadata-Version: 2.4
+Name: vapylib
+Version: 0.1.0
+Summary: a python voice synthesis lib
+License-Expression: GPL-3.0-only
+Requires-Python: >=3.10
+License-File: LICENSE
+Requires-Dist: numpy
+Requires-Dist: scipy
+Requires-Dist: sounddevice
+Requires-Dist: soundfile
+Requires-Dist: PyYAML
+Dynamic: license-file

vapylib-0.1.0/vapylib.egg-info/SOURCES.txt

@@ -0,0 +1,11 @@
+LICENSE
+README.md
+pyproject.toml
+vapylib/__init__.py
+vapylib/parser.py
+vapylib/synthesizer.py
+vapylib.egg-info/PKG-INFO
+vapylib.egg-info/SOURCES.txt
+vapylib.egg-info/dependency_links.txt
+vapylib.egg-info/requires.txt
+vapylib.egg-info/top_level.txt

vapylib-0.1.0/vapylib.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

vapylib-0.1.0/vapylib.egg-info/requires.txt

@@ -0,0 +1,5 @@
+numpy
+scipy
+sounddevice
+soundfile
+PyYAML

vapylib-0.1.0/vapylib.egg-info/top_level.txt

@@ -0,0 +1 @@
+vapylib