Add a new frame to the buffer. This function is only required if a buffer is needed. Otherwise, extract_frame_features() will be suff

Get the features for a track within the dataset. Parameters ---------- data : dict or string if dict: Trac

Concatenate two collections of batched notes. Parameters ---------- batched_notes : ndarray (N x 3) Array of note intervals an

Pad the audio such that the first sample is located halfway through the first frame. Parameters ---------- a

Remove null (zero-frequency) observations from the pitch list. Parameters ---------- pitch_list : list of ndarray (N x [...])

Pad audio such that it is divisible by the specified divisor. Parameters ---------- audio : ndarray Mono-c

Download the dataset to disk if possible. This is to be extended by a child class. Parameters ---------- save_dir :

Extract the duration of audio associated with JAMS annotations. Parameters ---------- jam : JAMS object JAMS file data Re

Acquire the next frame from the stream.

Acquire the next frame of features from the stream. Returns ---------- features : ndarray Features for one

Extract all MIDI notes within a JAMS file. Parameters ---------- jam : JAMS object JAMS file data Returns ----------

Extract a pitch list from JAMS annotations. Parameters ---------- jam : JAMS object JAMS file data _times : ndarray or Non

Remove notes with nominal pitch outside the supported range of an instrument profile and notes with intervals entirely outside of specified b

Remove pitch observations (MIDI) outside the supported range of a profile. Parameters ---------- pitch_list : list of ndarray (N x [

Log the averaged results using TensorBoardX and reset the results tracking. Parameters ---------- writer : tensorboa

Perform the main processing steps for Onsets & Frames (V1). Parameters ---------- feats : Tensor (B x C x T x F)

Retrieve the currently buffered frames. Returns ---------- features : dict Dictionary containing a frame

Determine the current system time. Parameters ---------- decimals : int Number of digits to keep when rounding Returns

Default key describing estimates in the event no key was provided.

Determine the cumulative amount of time elapsed during the (potentially more than one) active phase of the stream. Parameter

Determine the number of frames the module will return for a piece of audio. This is the default behavior. It can be overridd

Determine the number of frames we expect from provided audio. Parameters ---------- audio: ndarray Mono-ch

Get the fret for a given MIDI pitch and string. Parameters ---------- midi_pitch : int MIDI pitch to query

Get the path to the annotations of a track. Parameters ---------- track : string GuitarSet track name

Get the path to the annotations of a track. Parameters ---------- track : string MAPS track name

Determine the range of audio samples which will produce features with a given number of frames. This is the default behavior

Determine the range of audio samples which will produce features with a given number of frames. Parameters ---------

Gather sample rates from inner modules, making sure they are equal. Returns ---------- sample_rate : int or float

Get the tracks associated with a dataset partition. Parameters ---------- split : string Name of the parti

Get the tracks associated with a dataset partition. Parameters ---------- split : string Name of the parti

Get the path to the audio of a track. Parameters ---------- track : string GuitarSet track name R

Get the path to the audio of a track. Parameters ---------- track : string MAPS track name Return

Increase the frame count by one.

Convert a multi pitch array into a pitch list. Array of corresponding times does not change and is assumed to be managed outside of the f

Normalizes an array of activations using infinity norm. Parameters ---------- activations : ndarray Provided activations

Convert note pitches from MIDI to Hertz. Array of corresponding intervals does not change and is assumed to be managed outside of the fun

Convert note pitches from Hertz to MIDI. Array of corresponding intervals does not change and is assumed to be managed outside of the fun

Obtain the offsets of loose MIDI note groups in multi pitch format. Parameters ---------- pitches : ndarray (N) Array of pitch

Obtain the onsets of loose MIDI note groups in multi pitch format. Parameters ---------- pitches : ndarray (N) Array of pitche

Convert pitch list from MIDI to Hertz. Array of corresponding times does not change and is assumed to be managed outside of the function.

Convert pitch list from Hertz to MIDI. Array of corresponding times does not change and is assumed to be managed outside of the function.

Convert a MIDI pitch list into a multi pitch array. Parameters ---------- pitch_list : list of ndarray (N x [...]) Array of MI

Convert a pitch list into a dictionary of stacked pitch lists. Parameters ---------- times : ndarray (N) Time in seconds of be

Static function for plotting guitar tablature organized as notes by string. TODO - this does not scale well for big note groups Paramete

Static function for plotting a 2D pitch activation map or pianoroll. Parameters ---------- multi_pitch : ndarray (F x T) Discr

Static function for plotting pitch contours (pitch_list). Parameters ---------- times : ndarray (N) Time in seconds of beginni

Static function for plotting pitch contours (pitch_list). Parameters ---------- stacked_pitch_list : dict Dictionary containin

Static function for plotting any time-frequency representation. Parameters ---------- tfr : ndarray (F x T) Time-frequency rep

Static function for plotting a 1D waveform. Parameters ---------- samples : ndarray Waveform samples to plot times : ndarr

Perform necessary post-processing steps for the transcription model. Parameters ---------- batch : dict Di

Calculate loss and finalize model output. Parameters ---------- batch : dict Dictionary including model ou

This method can be overridden in order to insert extra steps. Parameters ---------- raw_output : dict Dict

Perform necessary pre-processing steps for the transcription model. Parameters ---------- batch : dict Dic

Print a time to the console. Parameters ---------- t : float Arbitrary time label : string or None (Optional) Label

Combines pre_proc(), estimate(), and write(), and returns output in a dictionary. Parameters ---------- raw_output :

Determine if the stream has finished. (Default behavior) Returns ---------- active : bool Flag indicating

Determine if the stream is currently active. Returns ---------- active : bool Flag indicating the stream h

Determine if a dictionary has an entry for a specified key. TODO - can use this many places (e.g. datasets) to be safe and neat Paramet

Determine if the stream has finished. Returns ---------- finished : bool Flag indicating the stream was te

Determine if the buffer is full. Returns ---------- frame_buffer_full : bool Flag indicating the buffer is

Remove any tracks contained in the given splits from the initial dataset partition, should they exist. Parameters --

Reset the waveform visualizer.

Reset the TFR visualizer.

Reset the stacked pitch list visualizer.

Reset the guitar tablature visualizer.

Zero-out the tracked state.

Perform inference in an offline fashion. Parameters ---------- track_data : dict Dictionary containing relevant features for a

TODO - no loss entry in predictions causes whole dict to be interpreted as loss Perform inference in an mock-online fashion. Parameters

Perform inference on a single frame. TODO - this is all repeat code of run_offline - only difference is lack of dict_unsqueeze Parameter

Choose the device to use when opening the microphone stream. Parameters ---------- idx : int or None (Optional)

Simple helper function to set a new verbose flag. Parameters ---------- verbose : bool Whether to print an

Update the class weighting. Parameters ---------- weights : ndarray (N) or None (optional) Class weights f

Retain pitch observations within a time window. Parameters ---------- times : ndarray (N) Time in seconds of beginning of each

View stacked multi pitch arrays as a set of individual activations. Parameters ---------- stacked_multi_pitch : ndarray (S x F x T)

Collapse stacked multi pitch arrays into a single class representation. Parameters ---------- stacked_multi_pitch : ndarray (S x F x

Convert a stacked MIDI pitch list into a stack of multi pitch arrays. This function assumes that all pitch lists are relative to the same

Stop streaming the audio.

Stop streaming audio from the microphone. Parameters ---------- pause : bool Whether to pause the stream i

Stop streaming the audio.

Indicate that the thread should stop running. Only call this if you are completely finished with the thread, since it cannot be start

Convert a tablature representation into a stacked multi pitch array. Array of corresponding times does not change and is assumed to be ma

Simple helper function to convert a PyTorch tensor into a NumPy array in order to keep code readable. Parameters ---------- data

Performs binary thresholding on an array of activations. Parameters ---------- activations : ndarray Provided activations

Convert features to decibels (dB) units. This is the default behavior. It can be overridden. Parameters ----------

Switch the axes of batched notes. Parameters ---------- batched_notes : ndarray (3 x N) or (N x 3) Array of note intervals and

Write result dictionary to a text file. Parameters ---------- results : dictionary Dictionary containing results of tracks arr

Retrieve the (potentially randomly sliced) item associated with the selected index. Parameters ---------- index : in

Initialize estimators and instantiate. Parameters ---------- estimators : list of Estimator Estimators to

Initialize parameters for the estimator. Parameters ---------- See Estimator class for others... inhibition

Initialize parameters for the estimator. Parameters ---------- See StackedNoteTranscriber class...

Initialize parameters for the estimator. Parameters ---------- See Estimator class... notes_key : string or

Initialize parameters for the estimator. Parameters ---------- See Estimator class... multi_pitch_key : str

Initialize parameters for the estimator. Parameters ---------- See Estimator class... tablature_key : strin

Initialize parameters for the estimator. Parameters ---------- See Estimator class... stacked_key : string

Initialize parameters for the estimator. Parameters ---------- See MultiPitchWrapper class... multi_pitch_k

Initialize parameters for the evaluator. Parameters ---------- See Evaluator class... evaluators : list of

Initialize parameters for the evaluator. Parameters ---------- See Evaluator class for others... average_sl

Initialize parameters for the evaluator. Parameters ---------- See StackedEvaluator class...

Initialize parameters for the evaluator. Parameters ---------- See StackedEvaluator class for others... off