name: AudioFrontEnd : public FrontEndBase


g++ [flags ...] file ... -l /isip/tools/lib/$ISIP_BINARY/lib_sp.a

AudioFrontEnd(const AudioFrontEnd& arg);
boolean getVector(VectorFloat& coeffs, long channel, long frame_index);
AudioFile* getAudio();
long getNumChannels() const;
DMODE getOutputMode() const;
boolean run(const Filename& output, const Filename& input);
quick start:

AudioFrontEnd reg_fe0;
Filename reg_file(L"$ISIP_DEVEL/doc/examples/data/audio/little_endian.raw");
Filename output(L".new_features6.out");

Sof sof;"diagnose_recipe.sof");, 0);
sof.close();, reg_file);

This AudioFrontEnd class implements a general front end for audio signal data processing. The user specifies what processing should take place by creating a
Recipe of general Algorithm objects. The simplest way to create this is with the graphical tool we have developed specifically for this task: transform builder.

The first step of the class is to determine what algorithms from the user's Component graph to apply, and in what order, to attain the desired output from the given inputs. Once determined, the class handles reading inputs, marshaling data in and out of the algorithm apply methods, and (optionally) writing outputs.

The class has two levels of interface. The first is the getVector() method, which allows a higher level program (like the recognizer) to obtain feature vectors on a frame-by-frame basis. The second is the run interface, which is used by the isip_transform driver program to produce an output file for each input file.


public constants:

error codes:

protected data:

required public methods:

class-specific public methods:

private methods: