README
¶
audio
audio is a generic Go package designed to define a common interface to analyze
and/or process audio data.
At the heart of the package is the Buffer interface and its implementations:
FloatBufferFloat32BufferIntBuffer
Decoders, encoders, processors, analyzers and transformers can be written to accept or return these types and share a common interface.
The idea is that audio libraries can define this interface or its
implementations as input and return an audio.Buffer interface allowing all
audio libraries to be chainable.
Performance
The buffer implementations are designed so a buffer can be reused and mutated avoiding allocation penalties.
It is recommended to avoid using Float32Buffer unless performance is critical.
The major drawback of using float32s is that the Go stdlib was designed to work
with float64 and therefore the access to standard packages is limited.
Usage
Examples of how to use this interface is available under the go-audio organization.
Documentation
¶
Overview ¶
Package audio defines a common interface to analyze and/or process audio data.
At the heart of the package is the Buffer interface and its implementations: FloatBuffer and IntBuffer. Decoders, encoders, processors, analyzers and transformers can be written to accept or return these types and share a common interface.
Index ¶
- Variables
- func IEEEFloatToInt(b [10]byte) int
- func Int24BETo32(bytes []byte) int32
- func Int24LETo32(bytes []byte) int32
- func Int32toInt24BEBytes(n int32) []byte
- func Int32toInt24LEBytes(n int32) []byte
- func IntMaxSignedValue(b int) int
- func IntToIEEEFloat(i int) [10]byte
- func Uint24to32(bytes []byte) uint32
- func Uint32toUint24Bytes(n uint32) []byte
- type Buffer
- type Float32Buffer
- type FloatBuffer
- type Format
- type IntBuffer
- type PCMBuffer
- func (b *PCMBuffer) AsF32() (out []float32)
- func (b *PCMBuffer) AsF64() (out []float64)
- func (b *PCMBuffer) AsFloat32Buffer() *Float32Buffer
- func (b *PCMBuffer) AsFloatBuffer() *FloatBuffer
- func (b *PCMBuffer) AsI16() (out []int16)
- func (b *PCMBuffer) AsI32() (out []int32)
- func (b *PCMBuffer) AsI8() (out []int8)
- func (b *PCMBuffer) AsInt() (out []int)
- func (b *PCMBuffer) AsIntBuffer() *IntBuffer
- func (b *PCMBuffer) Clone() Buffer
- func (b *PCMBuffer) Len() int
- func (b *PCMBuffer) NumFrames() int
- func (b *PCMBuffer) PCMFormat() *Format
- func (b *PCMBuffer) SwitchPrimaryType(t PCMDataFormat)
- type PCMDataFormat
Constants ¶
This section is empty.
Variables ¶
var ( // FormatMono22500 is mono 22.5kHz format. FormatMono22500 = &Format{ NumChannels: 1, SampleRate: 22500, } // FormatMono44100 is mono 8bit 44.1kHz format. FormatMono44100 = &Format{ NumChannels: 1, SampleRate: 44100, } // FormatMono48000 is mono 48kHz format. FormatMono48000 = &Format{ NumChannels: 1, SampleRate: 48000, } // FormatMono96000 is mono 96kHz format. FormatMono96000 = &Format{ NumChannels: 1, SampleRate: 96000, } // FormatStereo22500 is stereo 22.5kHz format. FormatStereo22500 = &Format{ NumChannels: 2, SampleRate: 22500, } // FormatStereo44100 is stereo 8bit 44.1kHz format. FormatStereo44100 = &Format{ NumChannels: 2, SampleRate: 44100, } // FormatStereo48000 is stereo 48kHz format. FormatStereo48000 = &Format{ NumChannels: 2, SampleRate: 48000, } // FormatStereo96000 is stereo 96kHz format. FormatStereo96000 = &Format{ NumChannels: 2, SampleRate: 96000, } )
var ( // ErrInvalidBuffer is a generic error returned when trying to read/write to an invalid buffer. ErrInvalidBuffer = errors.New("invalid buffer") )
Functions ¶
func IEEEFloatToInt ¶
IEEEFloatToInt converts a 10 byte IEEE float into an int.
func Int24BETo32 ¶
Int24BETo32 converts an int24 value from 3 bytes into an int32 value
func Int24LETo32 ¶
Int24LETo32 converts an int24 value from 3 bytes into an int32 value
func Int32toInt24BEBytes ¶
Int32toInt24BEBytes converts an int32 into a big endian 3 byte int24 representation
func Int32toInt24LEBytes ¶
Int32toInt24LEBytes converts an int32 into a little endian 3 byte int24 representation
func IntMaxSignedValue ¶
IntMaxSignedValue returns the max value of an integer based on its memory size
func IntToIEEEFloat ¶
IntToIEEEFloat converts an int into a 10 byte IEEE float.
func Uint24to32 ¶
Uint24to32 converts a 3 byte uint23 into a uint32 BigEndian!
func Uint32toUint24Bytes ¶
Uint32toUint24Bytes converts a uint32 into a 3 byte uint24 representation
Types ¶
type Buffer ¶
type Buffer interface {
// PCMFormat is the format of buffer (describing the buffer content/format).
PCMFormat() *Format
// NumFrames returns the number of frames contained in the buffer.
NumFrames() int
// AsFloatBuffer returns a float 64 buffer from this buffer.
AsFloatBuffer() *FloatBuffer
// AsFloat32Buffer returns a float 32 buffer from this buffer.
AsFloat32Buffer() *Float32Buffer
// AsIntBuffer returns an int buffer from this buffer.
AsIntBuffer() *IntBuffer
// Clone creates a clean clone that can be modified without
// changing the source buffer.
Clone() Buffer
}
Buffer is the representation of an audio buffer.
type Float32Buffer ¶
type Float32Buffer struct {
// Format is the representation of the underlying data format
Format *Format
// Data is the buffer PCM data as floats
Data []float32
// SourceBitDepth helps us know if the source was encoded on
// 8, 16, 24, 32, 64 bits.
SourceBitDepth int
}
Float32Buffer is an audio buffer with its PCM data formatted as float32.
func (*Float32Buffer) AsFloat32Buffer ¶
func (buf *Float32Buffer) AsFloat32Buffer() *Float32Buffer
AsFloat32Buffer implements the Buffer interface and returns itself.
func (*Float32Buffer) AsFloatBuffer ¶
func (buf *Float32Buffer) AsFloatBuffer() *FloatBuffer
AsFloatBuffer implements the Buffer interface and returns a float64 version of itself.
func (*Float32Buffer) AsIntBuffer ¶
func (buf *Float32Buffer) AsIntBuffer() *IntBuffer
AsIntBuffer returns a copy of this buffer but with data truncated to Ints. It is usually recommended to apply a transforms when going from a 24bit source to an int (16bit destination). Look at transforms.PCMScaleF32() for instance
func (*Float32Buffer) Clone ¶
func (buf *Float32Buffer) Clone() Buffer
Clone creates a clean clone that can be modified without changing the source buffer.
func (*Float32Buffer) NumFrames ¶
func (buf *Float32Buffer) NumFrames() int
NumFrames returns the number of frames contained in the buffer.
func (*Float32Buffer) PCMFormat ¶
func (buf *Float32Buffer) PCMFormat() *Format
PCMFormat returns the buffer format information.
type FloatBuffer ¶
type FloatBuffer struct {
// Format is the representation of the underlying data format
Format *Format
// Data is the buffer PCM data as floats
Data []float64
}
FloatBuffer is an audio buffer with its PCM data formatted as float64.
func (*FloatBuffer) AsFloat32Buffer ¶
func (buf *FloatBuffer) AsFloat32Buffer() *Float32Buffer
AsFloat32Buffer implements the Buffer interface and returns a float 32 version of itself.
func (*FloatBuffer) AsFloatBuffer ¶
func (buf *FloatBuffer) AsFloatBuffer() *FloatBuffer
AsFloatBuffer implements the Buffer interface and returns itself.
func (*FloatBuffer) AsIntBuffer ¶
func (buf *FloatBuffer) AsIntBuffer() *IntBuffer
AsIntBuffer returns a copy of this buffer but with data truncated to Ints.
func (*FloatBuffer) Clone ¶
func (buf *FloatBuffer) Clone() Buffer
Clone creates a clean clone that can be modified without changing the source buffer.
func (*FloatBuffer) NumFrames ¶
func (buf *FloatBuffer) NumFrames() int
NumFrames returns the number of frames contained in the buffer.
func (*FloatBuffer) PCMFormat ¶
func (buf *FloatBuffer) PCMFormat() *Format
PCMFormat returns the buffer format information.
type Format ¶
type Format struct {
// NumChannels is the number of channels contained in the data
NumChannels int
// SampleRate is the sampling rate in Hz
SampleRate int
}
Format is a high level representation of the underlying data.
type IntBuffer ¶
type IntBuffer struct {
// Format is the representation of the underlying data format
Format *Format
// Data is the buffer PCM data as ints
Data []int
// SourceBitDepth helps us know if the source was encoded on
// 8, 16, 24, 32, 64 bits.
SourceBitDepth int
}
IntBuffer is an audio buffer with its PCM data formatted as int.
func (*IntBuffer) AsFloat32Buffer ¶
func (buf *IntBuffer) AsFloat32Buffer() *Float32Buffer
AsFloat32Buffer returns a copy of this buffer but with data converted to float 32.
func (*IntBuffer) AsFloatBuffer ¶
func (buf *IntBuffer) AsFloatBuffer() *FloatBuffer
AsFloatBuffer returns a copy of this buffer but with data converted to floats.
func (*IntBuffer) AsIntBuffer ¶
AsIntBuffer implements the Buffer interface and returns itself.
func (*IntBuffer) Clone ¶
Clone creates a clean clone that can be modified without changing the source buffer.
type PCMBuffer ¶
type PCMBuffer struct {
// Format describes the format of the buffer data.
Format *Format
// I8 is a store for audio sample data as integers.
I8 []int8
// I16 is a store for audio sample data as integers.
I16 []int16
// I32 is a store for audio sample data as integers.
I32 []int32
// F32 is a store for audio samples data as float64.
F32 []float32
// F64 is a store for audio samples data as float64.
F64 []float64
// DataType indicates the primary format used for the underlying data.
// The consumer of the buffer might want to look at this value to know what store
// to use to optimaly retrieve data.
DataType PCMDataFormat
// SourceBitDepth helps us know if the source was encoded on
// 1 (int8), 2 (int16), 3(int24), 4(int32), 8(int64) bytes.
SourceBitDepth uint8
}
PCMBuffer encapsulates uncompressed audio data and provides useful methods to read/manipulate this PCM data. It's a more flexible buffer type allowing the developer to handle different kind of buffer data formats and convert between underlying types.
func (*PCMBuffer) AsF32 ¶
AsF32 returns the buffer's samples as float32 sample values. If the buffer isn't in this format, a copy is created and converted. Note that converting might result in unexpected truncations.
func (*PCMBuffer) AsF64 ¶
AsF64 returns the buffer's samples as float64 sample values. If the buffer isn't in this format, a copy is created and converted. Note that converting might result in unexpected truncations.
func (*PCMBuffer) AsFloat32Buffer ¶
func (b *PCMBuffer) AsFloat32Buffer() *Float32Buffer
AsFloat32Buffer implements the Buffer interface and returns a float 32 version of itself.
func (*PCMBuffer) AsFloatBuffer ¶
func (b *PCMBuffer) AsFloatBuffer() *FloatBuffer
AsFloatBuffer returns a copy of this buffer but with data converted to floats.
func (*PCMBuffer) AsI16 ¶
AsI16 returns the buffer's samples as int16 sample values. If the buffer isn't in this format, a copy is created and converted. Note that converting might result in loss of resolution.
func (*PCMBuffer) AsI32 ¶
AsI32 returns the buffer's samples as int32 sample values. If the buffer isn't in this format, a copy is created and converted. Note that converting a float to an int might result in unexpected truncations.
func (*PCMBuffer) AsI8 ¶
AsI8 returns the buffer's samples as int8 sample values. If the buffer isn't in this format, a copy is created and converted. Note that converting might result in loss of resolution.
func (*PCMBuffer) AsInt ¶
AsInt returns the buffer content as integers (int32s). It's recommended to avoid this method since it creates an extra copy of the buffer content.
func (*PCMBuffer) AsIntBuffer ¶
AsIntBuffer returns a copy of this buffer but with data truncated to Ints.
func (*PCMBuffer) Clone ¶
Clone creates a clean clone that can be modified without changing the source buffer.
func (*PCMBuffer) SwitchPrimaryType ¶
func (b *PCMBuffer) SwitchPrimaryType(t PCMDataFormat)
SwitchPrimaryType is a convenience method to switch the primary data type. Use this if you process/swap a different type than the original type. Notes that conversion might be lossy if you switch to a lower resolution format.
type PCMDataFormat ¶
type PCMDataFormat uint8
PCMDataFormat is an enum type to indicate the underlying data format used.
const ( // DataTypeUnknown refers to an unknown format DataTypeUnknown PCMDataFormat = iota // DataTypeI8 indicates that the content of the audio buffer made of 8-bit integers. DataTypeI8 // DataTypeI16 indicates that the content of the audio buffer made of 16-bit integers. DataTypeI16 // DataTypeI32 indicates that the content of the audio buffer made of 32-bit integers. DataTypeI32 // DataTypeF32 indicates that the content of the audio buffer made of 32-bit floats. DataTypeF32 // DataTypeF64 indicates that the content of the audio buffer made of 64-bit floats. DataTypeF64 )
Source Files