Rubber Band Library 3.0.0
RubberBand::RubberBandStretcher Class Reference

#include <RubberBandStretcher.h>

Classes

struct  Logger
 Interface for log callbacks that may optionally be provided to the stretcher on construction. More...
 

Public Types

enum  Option {
  OptionProcessOffline = 0x00000000 , OptionProcessRealTime = 0x00000001 , OptionStretchElastic = 0x00000000 , OptionStretchPrecise = 0x00000010 ,
  OptionTransientsCrisp = 0x00000000 , OptionTransientsMixed = 0x00000100 , OptionTransientsSmooth = 0x00000200 , OptionDetectorCompound = 0x00000000 ,
  OptionDetectorPercussive = 0x00000400 , OptionDetectorSoft = 0x00000800 , OptionPhaseLaminar = 0x00000000 , OptionPhaseIndependent = 0x00002000 ,
  OptionThreadingAuto = 0x00000000 , OptionThreadingNever = 0x00010000 , OptionThreadingAlways = 0x00020000 , OptionWindowStandard = 0x00000000 ,
  OptionWindowShort = 0x00100000 , OptionWindowLong = 0x00200000 , OptionSmoothingOff = 0x00000000 , OptionSmoothingOn = 0x00800000 ,
  OptionFormantShifted = 0x00000000 , OptionFormantPreserved = 0x01000000 , OptionPitchHighSpeed = 0x00000000 , OptionPitchHighQuality = 0x02000000 ,
  OptionPitchHighConsistency = 0x04000000 , OptionChannelsApart = 0x00000000 , OptionChannelsTogether = 0x10000000 , OptionEngineFaster = 0x00000000 ,
  OptionEngineFiner = 0x20000000
}
 Processing options for the timestretcher. More...
 
enum  PresetOption { DefaultOptions = 0x00000000 , PercussiveOptions = 0x00102000 }
 
typedef int Options
 A bitwise OR of values from the RubberBandStretcher::Option enum. More...
 

Public Member Functions

 RubberBandStretcher (size_t sampleRate, size_t channels, Options options=DefaultOptions, double initialTimeRatio=1.0, double initialPitchScale=1.0)
 Construct a time and pitch stretcher object to run at the given sample rate, with the given number of channels. More...
 
 RubberBandStretcher (size_t sampleRate, size_t channels, std::shared_ptr< Logger > logger, Options options=DefaultOptions, double initialTimeRatio=1.0, double initialPitchScale=1.0)
 Construct a time and pitch stretcher object with a custom debug logger. More...
 
 ~RubberBandStretcher ()
 
void reset ()
 Reset the stretcher's internal buffers. More...
 
int getEngineVersion () const
 Return the active internal engine version, according to the OptionEngine flag supplied on construction. More...
 
void setTimeRatio (double ratio)
 Set the time ratio for the stretcher. More...
 
void setPitchScale (double scale)
 Set the pitch scaling ratio for the stretcher. More...
 
void setFormantScale (double scale)
 Set a pitch scale for the vocal formant envelope separately from the overall pitch scale. More...
 
double getTimeRatio () const
 Return the last time ratio value that was set (either on construction or with setTimeRatio()). More...
 
double getPitchScale () const
 Return the last pitch scaling ratio value that was set (either on construction or with setPitchScale()). More...
 
double getFormantScale () const
 Return the last formant scaling ratio that was set with setFormantScale, or 0.0 if the default automatic scaling is in effect. More...
 
size_t getPreferredStartPad () const
 In RealTime mode (unlike in Offline mode) the stretcher performs no automatic padding or delay/latency compensation at the start of the signal. More...
 
size_t getStartDelay () const
 Return the output delay of the stretcher. More...
 
size_t getLatency () const
 Return the start delay of the stretcher. More...
 
size_t getChannelCount () const
 Return the number of channels this stretcher was constructed with. More...
 
void setTransientsOption (Options options)
 Change an OptionTransients configuration setting. More...
 
void setDetectorOption (Options options)
 Change an OptionDetector configuration setting. More...
 
void setPhaseOption (Options options)
 Change an OptionPhase configuration setting. More...
 
void setFormantOption (Options options)
 Change an OptionFormant configuration setting. More...
 
void setPitchOption (Options options)
 Change an OptionPitch configuration setting. More...
 
void setExpectedInputDuration (size_t samples)
 Tell the stretcher exactly how many input sample frames it will receive. More...
 
void setMaxProcessSize (size_t samples)
 Tell the stretcher the maximum number of sample frames that you will ever be passing in to a single process() call. More...
 
size_t getSamplesRequired () const
 Ask the stretcher how many audio sample frames should be provided as input in order to ensure that some more output becomes available. More...
 
void setKeyFrameMap (const std::map< size_t, size_t > &)
 Provide a set of mappings from "before" to "after" sample numbers so as to enforce a particular stretch profile. More...
 
void study (const float *const *input, size_t samples, bool final)
 Provide a block of "samples" sample frames for the stretcher to study and calculate a stretch profile from. More...
 
void process (const float *const *input, size_t samples, bool final)
 Provide a block of "samples" sample frames for processing. More...
 
int available () const
 Ask the stretcher how many audio sample frames of output data are available for reading (via retrieve()). More...
 
size_t retrieve (float *const *output, size_t samples) const
 Obtain some processed output data from the stretcher. More...
 
float getFrequencyCutoff (int n) const
 Return the value of internal frequency cutoff value n. More...
 
void setFrequencyCutoff (int n, float f)
 Set the value of internal frequency cutoff n to f Hz. More...
 
size_t getInputIncrement () const
 Retrieve the value of the internal input block increment value. More...
 
std::vector< int > getOutputIncrements () const
 In offline mode, retrieve the sequence of internal block increments for output, for the entire audio data, provided the stretch profile has been calculated. More...
 
std::vector< float > getPhaseResetCurve () const
 In offline mode, retrieve the sequence of internal phase reset detection function values, for the entire audio data, provided the stretch profile has been calculated. More...
 
std::vector< int > getExactTimePoints () const
 In offline mode, retrieve the sequence of internal frames for which exact timing has been sought, for the entire audio data, provided the stretch profile has been calculated. More...
 
void calculateStretch ()
 Force the stretcher to calculate a stretch profile. More...
 
void setDebugLevel (int level)
 Set the level of debug output. More...
 

Static Public Member Functions

static void setDefaultDebugLevel (int level)
 Set the default level of debug output for subsequently constructed stretchers. More...
 

Protected Member Functions

 RubberBandStretcher (const RubberBandStretcher &)=delete
 
RubberBandStretcheroperator= (const RubberBandStretcher &)=delete
 

Protected Attributes

Impl * m_d
 

Detailed Description

Definition at line 124 of file RubberBandStretcher.h.

Member Typedef Documentation

◆ Options

A bitwise OR of values from the RubberBandStretcher::Option enum.

Definition at line 417 of file RubberBandStretcher.h.

Member Enumeration Documentation

◆ Option

Processing options for the timestretcher.

The preferred options should normally be set in the constructor, as a bitwise OR of the option flags. The default value (DefaultOptions) is intended to give good results in most situations.

  1. Flags prefixed OptionProcess determine how the timestretcher will be invoked. These options may not be changed after construction.

    • OptionProcessOffline - Run the stretcher in offline mode. In this mode the input data needs to be provided twice, once to study(), which calculates a stretch profile for the audio, and once to process(), which stretches it.
    • OptionProcessRealTime - Run the stretcher in real-time mode. In this mode only process() should be called, and the stretcher adjusts dynamically in response to the input audio.

    The Process setting is likely to depend on your architecture: non-real-time operation on seekable files: Offline; real-time or streaming operation: RealTime.

  2. Flags prefixed OptionEngine select the core Rubber Band processing engine to be used. These options may not be changed after construction.

    • OptionEngineFaster - Use the Rubber Band Library R2 (Faster) engine. This is the engine implemented in Rubber Band Library v1.x and v2.x, and it remains the default in newer versions. It uses substantially less CPU than the R3 engine and there are still many situations in which it is likely to be the more appropriate choice.
    • OptionEngineFiner - Use the Rubber Band Library R3 (Finer) engine. This engine was introduced in Rubber Band Library v3.0. It produces higher-quality results than the R2 engine for most material, especially complex mixes, vocals and other sounds that have soft onsets and smooth pitch changes, and music with substantial bass content. However, it uses much more CPU power than the R2 engine.

    Important note: Consider calling getEngineVersion() after construction to make sure the engine you requested is active. That's not because engine selection can fail, but because Rubber Band Library ignores any unknown options supplied on construction - so a program that requests the R3 engine but ends up linked against an older version of the library (prior to v3.0) will silently use the R2 engine instead. Calling the v3.0 function getEngineVersion() will ensure a link failure in this situation instead, and supply a reassuring run-time check.

  3. Flags prefixed OptionTransients control the component frequency phase-reset mechanism in the R2 engine, that may be used at transient points to provide clarity and realism to percussion and other significant transient sounds. These options have no effect when using the R3 engine. These options may be changed after construction when running in real-time mode, but not when running in offline mode.

    • OptionTransientsCrisp - Reset component phases at the peak of each transient (the start of a significant note or percussive event). This, the default setting, usually results in a clear-sounding output; but it is not always consistent, and may cause interruptions in stable sounds present at the same time as transient events. The OptionDetector flags (below) can be used to tune this to some extent.
    • OptionTransientsMixed - Reset component phases at the peak of each transient, outside a frequency range typical of musical fundamental frequencies. The results may be more regular for mixed stable and percussive notes than OptionTransientsCrisp, but with a "phasier" sound. The balance may sound very good for certain types of music and fairly bad for others.
    • OptionTransientsSmooth - Do not reset component phases at any point. The results will be smoother and more regular but may be less clear than with either of the other transients flags.
  4. Flags prefixed OptionDetector control the type of transient detector used in the R2 engine. These options have no effect when using the R3 engine. These options may be changed after construction when running in real-time mode, but not when running in offline mode.

    • OptionDetectorCompound - Use a general-purpose transient detector which is likely to be good for most situations. This is the default.
    • OptionDetectorPercussive - Detect percussive transients. Note that this was the default and only option in Rubber Band versions prior to 1.5.
    • OptionDetectorSoft - Use an onset detector with less of a bias toward percussive transients. This may give better results with certain material (e.g. relatively monophonic piano music).
  5. Flags prefixed OptionPhase control the adjustment of component frequency phases in the R2 engine from one analysis window to the next during non-transient segments. These options have no effect when using the R3 engine. These options may be changed at any time.

    • OptionPhaseLaminar - Adjust phases when stretching in such a way as to try to retain the continuity of phase relationships between adjacent frequency bins whose phases are behaving in similar ways. This, the default setting, should give good results in most situations.
    • OptionPhaseIndependent - Adjust the phase in each frequency bin independently from its neighbours. This usually results in a slightly softer, phasier sound.
  6. Flags prefixed OptionThreading control the threading model of the stretcher. These options may not be changed after construction.

    • OptionThreadingAuto - Permit the stretcher to determine its own threading model. In the R2 engine this means using one processing thread per audio channel in offline mode if the stretcher is able to determine that more than one CPU is available, and one thread only in realtime mode. The R3 engine does not currently have a multi-threaded mode, but if one is introduced in future, this option may use it. This is the default.
    • OptionThreadingNever - Never use more than one thread.
    • OptionThreadingAlways - Use multiple threads in any situation where OptionThreadingAuto would do so, except omit the check for multiple CPUs and instead assume it to be true.
  7. Flags prefixed OptionWindow control the window size for FFT processing in the R2 engine. (The window size actually used will depend on many factors, but it can be influenced.) These options have no effect when using the R3 engine. These options may not be changed after construction.

    • OptionWindowStandard - Use the default window size. The actual size will vary depending on other parameters. This option is expected to produce better results than the other window options in most situations.
    • OptionWindowShort - Use a shorter window. This may result in crisper sound for audio that depends strongly on its timing qualities.
    • OptionWindowLong - Use a longer window. This is likely to result in a smoother sound at the expense of clarity and timing.
  8. Flags prefixed OptionSmoothing control the use of window-presum FFT and time-domain smoothing in the R2 engine. These options have no effect when using the R3 engine. These options may not be changed after construction.

    • OptionSmoothingOff - Do not use time-domain smoothing. This is the default.
    • OptionSmoothingOn - Use time-domain smoothing. This will result in a softer sound with some audible artifacts around sharp transients, but it may be appropriate for longer stretches of some instruments and can mix well with OptionWindowShort.
  9. Flags prefixed OptionFormant control the handling of formant shape (spectral envelope) when pitch-shifting. These options affect both the R2 and R3 engines. These options may be changed at any time.

    • OptionFormantShifted - Apply no special formant processing. The spectral envelope will be pitch shifted as normal. This is the default.
    • OptionFormantPreserved - Preserve the spectral envelope of the unshifted signal. This permits shifting the note frequency without so substantially affecting the perceived pitch profile of the voice or instrument.
  10. Flags prefixed OptionPitch control the method used for pitch shifting. In the R2 engine they may be changed at any time but affect only realtime mode (in offline mode the method cannot be changed). In the R3 engine they affect both realtime and offline modes but are fixed on construction.

    • OptionPitchHighSpeed - Favour CPU cost over sound quality. This is the default. Use this when time-stretching only, or for fixed pitch shifts where CPU usage is of concern. Do not use this for arbitrarily time-varying pitch shifts (see OptionPitchHighConsistency below).
    • OptionPitchHighQuality - Favour sound quality over CPU cost. Use this for fixed pitch shifts where sound quality is of most concern. Do not use this for arbitrarily time-varying pitch shifts (see OptionPitchHighConsistency below).
    • OptionPitchHighConsistency - Use a method that supports dynamic pitch changes without discontinuities, including when crossing the 1.0 pitch scale. This may cost more in CPU than the other two options, especially when the pitch scale is exactly 1.0. You should use this option whenever you wish to support dynamically changing pitch shifts during processing.
  11. Flags prefixed OptionChannels control the method used for processing two-channel stereo audio. These have different, but related, effects in the R2 and R3 engines. These options may not be changed after construction.

    • OptionChannelsApart - Channels are handled for maximum individual fidelity, at the expense of synchronisation. In the R3 engine, this means frequency-bin synchronisation is maintained more closely for lower-frequency content than higher. In R2, it means the stereo channels are processed individually and only synchronised at transients. In both engines this gives the highest quality for the individual channels but a more diffuse stereo image and an unnatural increase in "width". This option is the default.
    • OptionChannelsTogether - Channels are handled for higher synchronisation at the expense of individual fidelity. In the R3 engine, this means stereo synchronisation is maintained more closely for the full frequency range. In R2, it means the first two channels are considered to be a stereo pair and are processed in mid-side format, with mid and side processed as if they were separate channels before being recombined. This usually leads to better focus in the centre but relatively less stereo space and width and lower fidelity for individual channel content.

    Finally, flags prefixed OptionStretch are obsolete flags provided for backward compatibility only. They are ignored by the stretcher.

Enumerator
OptionProcessOffline 
OptionProcessRealTime 
OptionStretchElastic 
OptionStretchPrecise 
OptionTransientsCrisp 
OptionTransientsMixed 
OptionTransientsSmooth 
OptionDetectorCompound 
OptionDetectorPercussive 
OptionDetectorSoft 
OptionPhaseLaminar 
OptionPhaseIndependent 
OptionThreadingAuto 
OptionThreadingNever 
OptionThreadingAlways 
OptionWindowStandard 
OptionWindowShort 
OptionWindowLong 
OptionSmoothingOff 
OptionSmoothingOn 
OptionFormantShifted 
OptionFormantPreserved 
OptionPitchHighSpeed 
OptionPitchHighQuality 
OptionPitchHighConsistency 
OptionChannelsApart 
OptionChannelsTogether 
OptionEngineFaster 
OptionEngineFiner 

Definition at line 367 of file RubberBandStretcher.h.

◆ PresetOption

Enumerator
DefaultOptions 
PercussiveOptions 

Definition at line 419 of file RubberBandStretcher.h.

Constructor & Destructor Documentation

◆ RubberBandStretcher() [1/3]

RubberBand::RubberBandStretcher::RubberBandStretcher ( size_t  sampleRate,
size_t  channels,
Options  options = DefaultOptions,
double  initialTimeRatio = 1.0,
double  initialPitchScale = 1.0 
)

Construct a time and pitch stretcher object to run at the given sample rate, with the given number of channels.

Initial time and pitch scaling ratios and other processing options may be provided. In particular, the behaviour of the stretcher depends strongly on whether offline or real-time mode is selected on construction (via OptionProcessOffline or OptionProcessRealTime option - offline is the default).

In offline mode, you must provide the audio block-by-block in two passes: in the first pass calling study(), in the second pass calling process() and receiving the output via retrieve(). In real-time mode, there is no study pass, just a single streaming pass in which the audio is passed to process() and output received via retrieve().

In real-time mode you can change the time and pitch ratios at any time, but in offline mode they are fixed and cannot be changed after the study pass has begun. (However, see setKeyFrameMap() for a way to do pre-planned variable time stretching in offline mode.)

See the option documentation above for more details.

◆ RubberBandStretcher() [2/3]

RubberBand::RubberBandStretcher::RubberBandStretcher ( size_t  sampleRate,
size_t  channels,
std::shared_ptr< Logger logger,
Options  options = DefaultOptions,
double  initialTimeRatio = 1.0,
double  initialPitchScale = 1.0 
)

Construct a time and pitch stretcher object with a custom debug logger.

This may be useful for debugging if the default logger output (which simply goes to cerr) is not visible in the runtime environment, or if the application has a standard or more realtime-appropriate logging mechanism.

See the documentation for the other constructor above for details of the arguments other than the logger.

Note that although the supplied logger gets to decide what to do with log messages, the separately-set debug level (see setDebugLevel() and setDefaultDebugLevel()) still determines whether any given debug message is sent to the logger in the first place.

◆ ~RubberBandStretcher()

RubberBand::RubberBandStretcher::~RubberBandStretcher ( )

◆ RubberBandStretcher() [3/3]

RubberBand::RubberBandStretcher::RubberBandStretcher ( const RubberBandStretcher )
protecteddelete

Member Function Documentation

◆ reset()

void RubberBand::RubberBandStretcher::reset ( )

Reset the stretcher's internal buffers.

The stretcher should subsequently behave as if it had just been constructed (although retaining the current time and pitch ratio).

◆ getEngineVersion()

int RubberBand::RubberBandStretcher::getEngineVersion ( ) const

Return the active internal engine version, according to the OptionEngine flag supplied on construction.

This will return 2 for the R2 (Faster) engine or 3 for the R3 (Finer) engine.

This function was added in Rubber Band Library v3.0.

◆ setTimeRatio()

void RubberBand::RubberBandStretcher::setTimeRatio ( double  ratio)

Set the time ratio for the stretcher.

This is the ratio of stretched to unstretched duration – not tempo. For example, a ratio of 2.0 would make the audio twice as long (i.e. halve the tempo); 0.5 would make it half as long (i.e. double the tempo); 1.0 would leave the duration unaffected.

If the stretcher was constructed in Offline mode, the time ratio is fixed throughout operation; this function may be called any number of times between construction (or a call to reset()) and the first call to study() or process(), but may not be called after study() or process() has been called.

If the stretcher was constructed in RealTime mode, the time ratio may be varied during operation; this function may be called at any time, so long as it is not called concurrently with process(). You should either call this function from the same thread as process(), or provide your own mutex or similar mechanism to ensure that setTimeRatio and process() cannot be run at once (there is no internal mutex for this purpose).

◆ setPitchScale()

void RubberBand::RubberBandStretcher::setPitchScale ( double  scale)

Set the pitch scaling ratio for the stretcher.

This is the ratio of target frequency to source frequency. For example, a ratio of 2.0 would shift up by one octave; 0.5 down by one octave; or 1.0 leave the pitch unaffected.

To put this in musical terms, a pitch scaling ratio corresponding to a shift of S equal-tempered semitones (where S is positive for an upwards shift and negative for downwards) is pow(2.0, S / 12.0).

If the stretcher was constructed in Offline mode, the pitch scaling ratio is fixed throughout operation; this function may be called any number of times between construction (or a call to reset()) and the first call to study() or process(), but may not be called after study() or process() has been called.

If the stretcher was constructed in RealTime mode, the pitch scaling ratio may be varied during operation; this function may be called at any time, so long as it is not called concurrently with process(). You should either call this function from the same thread as process(), or provide your own mutex or similar mechanism to ensure that setPitchScale and process() cannot be run at once (there is no internal mutex for this purpose).

◆ setFormantScale()

void RubberBand::RubberBandStretcher::setFormantScale ( double  scale)

Set a pitch scale for the vocal formant envelope separately from the overall pitch scale.

This is a ratio of target frequency to source frequency. For example, a ratio of 2.0 would shift the formant envelope up by one octave; 0.5 down by one octave; or 1.0 leave the formant unaffected.

By default this is set to the special value of 0.0, which causes the scale to be calculated automatically. It will be treated as 1.0 / the pitch scale if OptionFormantPreserved is specified, or 1.0 for OptionFormantShifted.

Conversely, if this is set to a value other than the default 0.0, formant shifting will happen regardless of the state of the OptionFormantPreserved/OptionFormantShifted option.

This function is provided for special effects only. You do not need to call it for ordinary pitch shifting, with or without formant preservation - just specify or omit the OptionFormantPreserved option as appropriate. Use this function only if you want to shift formants by a distance other than that of the overall pitch shift.

This function is supported only in the R3 (OptionEngineFiner) engine. It has no effect in R2 (OptionEngineFaster).

This function was added in Rubber Band Library v3.0.

◆ getTimeRatio()

double RubberBand::RubberBandStretcher::getTimeRatio ( ) const

Return the last time ratio value that was set (either on construction or with setTimeRatio()).

◆ getPitchScale()

double RubberBand::RubberBandStretcher::getPitchScale ( ) const

Return the last pitch scaling ratio value that was set (either on construction or with setPitchScale()).

◆ getFormantScale()

double RubberBand::RubberBandStretcher::getFormantScale ( ) const

Return the last formant scaling ratio that was set with setFormantScale, or 0.0 if the default automatic scaling is in effect.

This function is supported only in the R3 (OptionEngineFiner) engine. It always returns 0.0 in R2 (OptionEngineFaster).

This function was added in Rubber Band Library v3.0.

◆ getPreferredStartPad()

size_t RubberBand::RubberBandStretcher::getPreferredStartPad ( ) const

In RealTime mode (unlike in Offline mode) the stretcher performs no automatic padding or delay/latency compensation at the start of the signal.

This permits applications to have their own custom requirements, but it also means that by default some samples will be lost or attenuated at the start of the output and the correct linear relationship between input and output sample counts may be lost.

Most applications using RealTime mode should solve this by calling getPreferredStartPad() and supplying the returned number of (silent) samples at the start of their input, before their first "true" process() call; and then also calling getStartDelay() and trimming the returned number of samples from the start of their stretcher's output.

Ensure you have set the time and pitch scale factors to their proper starting values before calling getRequiredStartPad() or getStartDelay().

In Offline mode, padding and delay compensation are handled internally and both functions always return zero.

This function was added in Rubber Band Library v3.0.

See also
getStartDelay

◆ getStartDelay()

size_t RubberBand::RubberBandStretcher::getStartDelay ( ) const

Return the output delay of the stretcher.

This is the number of audio samples that one should discard at the start of the output, after padding the start of the input with getPreferredStartPad(), in order to ensure that the resulting audio has the expected time alignment with the input.

Ensure you have set the time and pitch scale factors to their proper starting values before calling getPreferredStartPad() or getStartDelay().

In Offline mode, padding and delay compensation are handled internally and both functions always return zero.

This function was added in Rubber Band Library v3.0. Previously it was called getLatency(). It was renamed to avoid confusion with the number of samples needed at input to cause a block of processing to handle (returned by getSamplesRequired()) which is also sometimes referred to as latency.

See also
getPreferredStartPad

◆ getLatency()

size_t RubberBand::RubberBandStretcher::getLatency ( ) const

Return the start delay of the stretcher.

This is a deprecated alias for getStartDelay().

Deprecated:

◆ getChannelCount()

size_t RubberBand::RubberBandStretcher::getChannelCount ( ) const

Return the number of channels this stretcher was constructed with.

◆ setTransientsOption()

void RubberBand::RubberBandStretcher::setTransientsOption ( Options  options)

Change an OptionTransients configuration setting.

This may be called at any time in RealTime mode. It may not be called in Offline mode (for which the transients option is fixed on construction). This has no effect when using the R3 engine.

◆ setDetectorOption()

void RubberBand::RubberBandStretcher::setDetectorOption ( Options  options)

Change an OptionDetector configuration setting.

This may be called at any time in RealTime mode. It may not be called in Offline mode (for which the detector option is fixed on construction). This has no effect when using the R3 engine.

◆ setPhaseOption()

void RubberBand::RubberBandStretcher::setPhaseOption ( Options  options)

Change an OptionPhase configuration setting.

This may be called at any time in any mode. This has no effect when using the R3 engine.

Note that if running multi-threaded in Offline mode, the change may not take effect immediately if processing is already under way when this function is called.

◆ setFormantOption()

void RubberBand::RubberBandStretcher::setFormantOption ( Options  options)

Change an OptionFormant configuration setting.

This may be called at any time in any mode.

Note that if running multi-threaded in Offline mode, the change may not take effect immediately if processing is already under way when this function is called.

◆ setPitchOption()

void RubberBand::RubberBandStretcher::setPitchOption ( Options  options)

Change an OptionPitch configuration setting.

This may be called at any time in RealTime mode. It may not be called in Offline mode (for which the pitch option is fixed on construction). This has no effect when using the R3 engine.

◆ setExpectedInputDuration()

void RubberBand::RubberBandStretcher::setExpectedInputDuration ( size_t  samples)

Tell the stretcher exactly how many input sample frames it will receive.

This is only useful in Offline mode, when it allows the stretcher to ensure that the number of output samples is exactly correct. In RealTime mode no such guarantee is possible and this value is ignored.

Note that the value of "samples" refers to the number of audio sample frames, which may be multi-channel, not the number of individual samples. (For example, one second of stereo audio sampled at 44100Hz yields a value of 44100 sample frames, not 88200.) This rule applies throughout the Rubber Band API.

◆ setMaxProcessSize()

void RubberBand::RubberBandStretcher::setMaxProcessSize ( size_t  samples)

Tell the stretcher the maximum number of sample frames that you will ever be passing in to a single process() call.

If you don't call this, the stretcher will assume that you are calling getSamplesRequired() at each cycle and are never passing more samples than are suggested by that function.

If your application has some external constraint that means you prefer a fixed block size, then your normal mode of operation would be to provide that block size to this function; to loop calling process() with that size of block; after each call to process(), test whether output has been generated by calling available(); and, if so, call retrieve() to obtain it. See getSamplesRequired() for a more suitable operating mode for applications without such external constraints.

This function may not be called after the first call to study() or process().

Note that this value is only relevant to process(), not to study() (to which you may pass any number of samples at a time, and from which there is no output).

Note that the value of "samples" refers to the number of audio sample frames, which may be multi-channel, not the number of individual samples. (For example, one second of stereo audio sampled at 44100Hz yields a value of 44100 sample frames, not 88200.) This rule applies throughout the Rubber Band API.

◆ getSamplesRequired()

size_t RubberBand::RubberBandStretcher::getSamplesRequired ( ) const

Ask the stretcher how many audio sample frames should be provided as input in order to ensure that some more output becomes available.

If your application has no particular constraint on processing block size and you are able to provide any block size as input for each cycle, then your normal mode of operation would be to loop querying this function; providing that number of samples to process(); and reading the output using available() and retrieve(). See setMaxProcessSize() for a more suitable operating mode for applications that do have external block size constraints.

Note that this value is only relevant to process(), not to study() (to which you may pass any number of samples at a time, and from which there is no output).

Note that the return value refers to the number of audio sample frames, which may be multi-channel, not the number of individual samples. (For example, one second of stereo audio sampled at 44100Hz yields a value of 44100 sample frames, not 88200.) This rule applies throughout the Rubber Band API.

See also
getStartDelay

◆ setKeyFrameMap()

void RubberBand::RubberBandStretcher::setKeyFrameMap ( const std::map< size_t, size_t > &  )

Provide a set of mappings from "before" to "after" sample numbers so as to enforce a particular stretch profile.

The argument is a map from audio sample frame number in the source material, to the corresponding sample frame number in the stretched output. The mapping should be for key frames only, with a "reasonable" gap between mapped samples.

This function cannot be used in RealTime mode.

This function may not be called after the first call to process(). It should be called after the time and pitch ratios have been set; the results of changing the time and pitch ratios after calling this function are undefined. Calling reset() will clear this mapping.

The key frame map only affects points within the material; it does not determine the overall stretch ratio (that is, the ratio between the output material's duration and the source material's duration). You need to provide this ratio separately to setTimeRatio(), otherwise the results may be truncated or extended in unexpected ways regardless of the extent of the frame numbers found in the key frame map.

◆ study()

void RubberBand::RubberBandStretcher::study ( const float *const *  input,
size_t  samples,
bool  final 
)

Provide a block of "samples" sample frames for the stretcher to study and calculate a stretch profile from.

This is only meaningful in Offline mode, and is required if running in that mode. You should pass the entire input through study() before any process() calls are made, as a sequence of blocks in individual study() calls, or as a single large block.

"input" should point to de-interleaved audio data with one float array per channel. Sample values are conventionally expected to be in the range -1.0f to +1.0f. "samples" supplies the number of audio sample frames available in "input". If "samples" is zero, "input" may be NULL.

Note that the value of "samples" refers to the number of audio sample frames, which may be multi-channel, not the number of individual samples. (For example, one second of stereo audio sampled at 44100Hz yields a value of 44100 sample frames, not 88200.) This rule applies throughout the Rubber Band API.

Set "final" to true if this is the last block of data that will be provided to study() before the first process() call.

◆ process()

void RubberBand::RubberBandStretcher::process ( const float *const *  input,
size_t  samples,
bool  final 
)

Provide a block of "samples" sample frames for processing.

See also getSamplesRequired() and setMaxProcessSize().

"input" should point to de-interleaved audio data with one float array per channel. Sample values are conventionally expected to be in the range -1.0f to +1.0f. "samples" supplies the number of audio sample frames available in "input".

Note that the value of "samples" refers to the number of audio sample frames, which may be multi-channel, not the number of individual samples. (For example, one second of stereo audio sampled at 44100Hz yields a value of 44100 sample frames, not 88200.) This rule applies throughout the Rubber Band API.

Set "final" to true if this is the last block of input data.

◆ available()

int RubberBand::RubberBandStretcher::available ( ) const

Ask the stretcher how many audio sample frames of output data are available for reading (via retrieve()).

This function returns 0 if no frames are available: this usually means more input data needs to be provided, but if the stretcher is running in threaded mode it may just mean that not enough data has yet been processed. Call getSamplesRequired() to discover whether more input is needed.

Note that the return value refers to the number of audio sample frames, which may be multi-channel, not the number of individual samples. (For example, one second of stereo audio sampled at 44100Hz yields a value of 44100 sample frames, not 88200.) This rule applies throughout the Rubber Band API.

This function returns -1 if all data has been fully processed and all output read, and the stretch process is now finished.

◆ retrieve()

size_t RubberBand::RubberBandStretcher::retrieve ( float *const *  output,
size_t  samples 
) const

Obtain some processed output data from the stretcher.

Up to "samples" samples will be stored in each of the output arrays (one per channel for de-interleaved audio data) pointed to by "output". The number of sample frames available to be retrieved can be queried beforehand with a call to available(). The return value is the actual number of sample frames retrieved.

Note that the value of "samples" and the return value refer to the number of audio sample frames, which may be multi-channel, not the number of individual samples. (For example, one second of stereo audio sampled at 44100Hz yields a value of 44100 sample frames, not 88200.) This rule applies throughout the Rubber Band API.

◆ getFrequencyCutoff()

float RubberBand::RubberBandStretcher::getFrequencyCutoff ( int  n) const

Return the value of internal frequency cutoff value n.

This function is not for general use.

◆ setFrequencyCutoff()

void RubberBand::RubberBandStretcher::setFrequencyCutoff ( int  n,
float  f 
)

Set the value of internal frequency cutoff n to f Hz.

This function is not for general use.

◆ getInputIncrement()

size_t RubberBand::RubberBandStretcher::getInputIncrement ( ) const

Retrieve the value of the internal input block increment value.

This function is provided for diagnostic purposes only and supported only with the R2 engine.

◆ getOutputIncrements()

std::vector< int > RubberBand::RubberBandStretcher::getOutputIncrements ( ) const

In offline mode, retrieve the sequence of internal block increments for output, for the entire audio data, provided the stretch profile has been calculated.

In realtime mode, retrieve any output increments that have accumulated since the last call to getOutputIncrements, to a limit of 16.

This function is provided for diagnostic purposes only and supported only with the R2 engine.

◆ getPhaseResetCurve()

std::vector< float > RubberBand::RubberBandStretcher::getPhaseResetCurve ( ) const

In offline mode, retrieve the sequence of internal phase reset detection function values, for the entire audio data, provided the stretch profile has been calculated.

In realtime mode, retrieve any phase reset points that have accumulated since the last call to getPhaseResetCurve, to a limit of 16.

This function is provided for diagnostic purposes only and supported only with the R2 engine.

◆ getExactTimePoints()

std::vector< int > RubberBand::RubberBandStretcher::getExactTimePoints ( ) const

In offline mode, retrieve the sequence of internal frames for which exact timing has been sought, for the entire audio data, provided the stretch profile has been calculated.

In realtime mode, return an empty sequence.

This function is provided for diagnostic purposes only and supported only with the R2 engine.

◆ calculateStretch()

void RubberBand::RubberBandStretcher::calculateStretch ( )

Force the stretcher to calculate a stretch profile.

Normally this happens automatically for the first process() call in offline mode.

This function is provided for diagnostic purposes only and supported only with the R2 engine.

◆ setDebugLevel()

void RubberBand::RubberBandStretcher::setDebugLevel ( int  level)

Set the level of debug output.

The supported values are:

0. Report errors only.

  1. Report some information on construction and ratio change. Nothing is reported during normal processing unless something changes.
  2. Report a significant amount of information about ongoing stretch calculations during normal processing.
  3. Report a large amount of information and also (in the R2 engine) add audible ticks to the output at phase reset points. This is seldom useful.

The default is whatever has been set using setDefaultDebugLevel(), or 0 if that function has not been called.

All output goes to cerr unless a custom RubberBandStretcher::Logger has been provided on construction. Because writing to cerr is not RT-safe, only debug level 0 is RT-safe in normal use by default. Debug levels 0 and 1 use only C-string constants as debug messages, so they are RT-safe if your custom logger is RT-safe. Levels 2 and 3 are not guaranteed to be RT-safe in any conditions as they may construct messages by allocation.

See also
Logger
setDefaultDebugLevel

◆ setDefaultDebugLevel()

static void RubberBand::RubberBandStretcher::setDefaultDebugLevel ( int  level)
static

Set the default level of debug output for subsequently constructed stretchers.

See also
setDebugLevel

◆ operator=()

RubberBandStretcher & RubberBand::RubberBandStretcher::operator= ( const RubberBandStretcher )
protecteddelete

Member Data Documentation

◆ m_d

Impl* RubberBand::RubberBandStretcher::m_d
protected

Definition at line 1038 of file RubberBandStretcher.h.


The documentation for this class was generated from the following file: