README
¶
GO-WORLD
GO-WORLD is a Go port to WORLD - a high-quality speech analysis, modification and synthesis system. WORLD provides a way to decompose a speech signal into:
- Fundamental frequency (F0)
- spectral envelope
- aperiodicity
and re-synthesize a speech signal from these paramters. See here for the original WORLD.
Supported Platforms
- Linux
- Mac OS X
Note that the original WORLD works in windows as well. In order to use WORLD in windows, you have to build WORLD yourself since currently we don't have a installation script.
Installation
Binary dependency
First you need to install WORLD as a shared library:
git clone https://github.com/r9y9/WORLD.git && cd world
git checkout v0.2.1-3
./waf configure && ./waf
sudo ./waf install
GO-WORLD
go get github.com/r9y9/go-world
complete!
Usage
Import the package
import "github.com/r9y9/go-world"
and create a world instance with sample rate [hz] and frame period [ms].
w := world.New(sampleRate, framePeriod) // e.g. (44100, 5)
and then you can do whatever you want with WORLD.
F0 estimation and refinement
Dio
timeAxis, f0 := w.Dio(input, w.NewDioOption()) // default option is used
StoneMask
refinedF0 := w.StoneMask(input, timeAxis, f0)
Spectral envelope estimation by CheapTrick
spectrogram := w.CheapTrick(input, timeAxis, f0)
Aperiodicity ratio estimation by D4C
apiriodicity := w.D4C(input, timeAxis, f0)
Synthesis
w.Synthesis(f0, spectrogram, apiriodicity, len(input))
Example
License
Modified-BSD
Documentation
¶
Overview ¶
Package world provide a Go interface to WORLD - a high-quality speech analysis, modification and synthesis system.
Index ¶
- func CheapTrick(x []float64, fs int, timeAxis, f0 []float64) [][]float64
- func D4C(x []float64, fs int, timeAxis, f0 []float64) [][]float64
- func Dio(x []float64, fs int, option DioOption) ([]float64, []float64)
- func GetFFTSizeForCheapTrick(fs int) int
- func GetFFTSizeForStar(fs int) int
- func GetSamplesForDIO(fs, x_length int, framePeriod float64) int
- func StoneMask(x []float64, fs int, timeAxis, f0 []float64) []float64
- func Synthesis(f0 []float64, spectrogram, aperiodicity [][]float64, framePeriod float64, ...) []float64
- type DioOption
- type World
- func (w *World) CheapTrick(x []float64, timeAxis, f0 []float64) [][]float64
- func (w *World) D4C(x, timeAxis, f0 []float64) [][]float64
- func (w *World) Dio(x []float64, option DioOption) ([]float64, []float64)
- func (w *World) NewDioOption() DioOption
- func (w *World) Star(x []float64, timeAxis, f0 []float64) [][]float64
- func (w *World) StoneMask(x []float64, timeAxis, f0 []float64) []float64
- func (w *World) Synthesis(f0 []float64, spectrogram, aperiodicity [][]float64, length int) []float64
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func GetFFTSizeForCheapTrick ¶
func GetFFTSizeForStar ¶
GetFFTSizeForStar is deprecated. Use GetFFTSizeForCheapTrick instead.
func GetSamplesForDIO ¶
Types ¶
type DioOption ¶
type DioOption struct {
F0Floor float64
F0Ceil float64
ChannelsInOctave float64
FramePeriod float64 // [ms]
Speed int // (1,2, ..., 12)
AllowedRange float64 // recommended: 0.02 * FramePeriod
}
func NewDioOption ¶
func NewDioOption() DioOption
type World ¶
World represents WORLD - high-quality speech analysis, modification and synthesis system.
func (*World) CheapTrick ¶
func (*World) NewDioOption ¶
NewDioOption returns a default DioOpton.
Source Files
Directories