AudioChunk | Stores PCM data of a sound sequence and computes some transformations on it |
Corpus | Stores a set of training examples for a neural net |
Editor | The main class for the graphical network trainer |
FeatureExtractor | An abstract class, used as an interface between the processing unit and the features algorithms |
Filter | Abstract class standing for a feature filter (counts the high / low values, computes the averageā¦) |
AverageFilter | A filter for computing the temporal average of a value |
CentroidFilter | A filter for computing the centroid of a feature (i.e. the average indice). See code |
CompareFilter | Compares the values of a feature to a set of known patterns |
DistanceFilter | Computes the absolute value of the difference between a feature and a known pattern |
FluxFilter | A filter for computing the amount of variations of a given feature |
HighLowRatioFilter | Counts the number of high or low values on a given feature It computes the average and compares each value to the average multiplied by "bound", the float can be set in the XML file. The user can also decide if the number of values higher or lower should be counted |
MelFilter | A filter displaying a feature on a logarithmic scale |
MemoryFilter | A class for keeping track of the last n values of features |
RangeFilter | A class for computing the average between a set of features (usually for the spectrum) |
HZCRRExtr | Extracts the High ZCR Ratio (ZCR stands for Zero Crossing Rate) |
LSTERExtr | Extracts the Low STE Ratio (STE stands for Short Time Energy) |
SpectrumExtr | Retrives the spectrum using FFT |
STEExtr | Computes the Short Time Energy |
ZCRExtr | Computes the Zero Crossing Rate |
Graphique | Draws waveform profile or frenquency spectrum of the stream |
Hasher | Generic audio hash algorithm interface Abstract class for interfacing with SoundAnalyser |
PeakHasher | A hasher based on spectrum peaks |
MainWindow | Defines GUI and event connections |
NeuralNetwork | A slightly better neural network |
PCATransformer | Computes the PCA and transform a vector |
Plage | Stores informations (metadata) about a sound sequence |
Profile | Stores informations about both the audio source and the metadata source of a radio stream |
Program | Lists sound sequences extracted from the metadatas |
QuadNode | A node for the quadtree |
QuadTree | |
QuadTree::rect | A rectangle |
SettingsDialog | A settings dialog |
StreamPlayer | Manages stream decoding and sends it to the computing part of the application This class is intended to be rewritten for more specific usages (spectrum analysis, feature extraction, aso.). Its the interface between libVLC (which reads the media, decodes, resamples, and does all the hard DSP) and the using of the data |
SoundAnalyser | Main class that sets up and runs a pipeline |
Classifier | A class reading a file and classifying its content |
CorpusBuilder | Creates a corpus from a set of audio files |
FeatureDrawer | Draws features generated by a pipeline |
LivePlayer | Main class for live stream classification and visualisation |
Fingerprinter | A class reading a file and classifying its content |
SpectrumRecorder | A class which creates an image of the spectrum of a given file This is not used to compute features, that's just a tool |
View2D | A renderer for a 2D neural network |
Viewport | Internal struture used by View2D to keep in memory old views of a same network |