All Downloads are FREE. Search and download functionalities are using the official Maven repository.

marytts.tools.voiceimport.EndpointDetector Maven / Gradle / Ivy

The newest version!
/**
 * Copyright 2000-2009 DFKI GmbH.
 * All Rights Reserved.  Use is subject to license terms.
 *
 * This file is part of MARY TTS.
 *
 * MARY TTS is free software: you can redistribute it and/or modify
 * it under the terms of the GNU Lesser General Public License as published by
 * the Free Software Foundation, version 3 of the License.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public License
 * along with this program.  If not, see .
 *
 */
package marytts.tools.voiceimport;

import java.io.File;
import java.io.IOException;
import java.util.SortedMap;
import java.util.TreeMap;

import javax.sound.sampled.UnsupportedAudioFileException;

import marytts.util.data.audio.AudioConverterUtils;
import marytts.util.io.BasenameList;
import marytts.util.math.MathUtils;

/**
 * Identify and Remove End-ponints (intitial and final silences) from given set of wave files.
 * 
 * @author Sathish and Oytun
 * 
 */
public class EndpointDetector extends VoiceImportComponent {
	protected File textDir;
	protected File inputWavDir;
	protected File outputWavDir;
	protected String waveExt = ".wav";
	private BasenameList bnlist;

	protected DatabaseLayout db = null;
	protected int percent = 0;

	public String INPUTWAVDIR = "EndpointDetector.inputWaveDirectory";
	public String OUTPUTWAVDIR = "EndpointDetector.outputWaveDirectory";
	public String ENERGYBUFFERLENGTH = "EndpointDetector.energyBufferLength";
	public String SPEECHSTARTLIKELIHOOD = "EndpointDetector.speechStartLikelihood";
	public String SPEECHENDLIKELIHOOD = "EndpointDetector.speechEndLikelihood";
	public String SHIFTFROMMINIMUMENERGYCENTER = "EndpointDetector.shiftFromMinimumEnergyCenter";
	public String NUMENERGYCLUSTERS = "EndpointDetector.numEnergyClusters";
	public String MINIMUMSTARTSILENCEINSECONDS = "EndpointDetector.minimumStartSilenceInSeconds";
	public String MINIMUMENDSILENCEINSECONDS = "EndpointDetector.minimumEndSilenceInSeconds";

	public String getName() {
		return "EndpointDetector";
	}

	public SortedMap getDefaultProps(DatabaseLayout theDb) {
		this.db = theDb;
		if (props == null) {
			props = new TreeMap();
			props.put(INPUTWAVDIR, db.getProp(db.ROOTDIR) + "inputwav" + System.getProperty("file.separator"));

			props.put(OUTPUTWAVDIR, db.getProp(db.ROOTDIR) + "outputwav" + System.getProperty("file.separator"));

			props.put(ENERGYBUFFERLENGTH, "20");
			props.put(SPEECHSTARTLIKELIHOOD, "0.1");
			props.put(SPEECHENDLIKELIHOOD, "0.1");
			props.put(SHIFTFROMMINIMUMENERGYCENTER, "0.0");
			props.put(NUMENERGYCLUSTERS, "4");
			props.put(MINIMUMSTARTSILENCEINSECONDS, "1.0");
			props.put(MINIMUMENDSILENCEINSECONDS, "1.0");
		}
		return props;
	}

	protected void setupHelp() {
		props2Help = new TreeMap();

		props2Help.put(INPUTWAVDIR, "input wave files directory.");

		props2Help.put(OUTPUTWAVDIR, "output directory to store initial-end silences removed wave files."
				+ "Will be created if it does not exist");

		props2Help.put(ENERGYBUFFERLENGTH, "number of consecutive speech frames when searching for speech/silence start events"
				+ "Range [1, 1000], decrease to detect more events");

		props2Help.put(SPEECHSTARTLIKELIHOOD, "likelihood of speech starting event"
				+ "Range [0.0,1.0], decrease to get more silence before speech segments");

		props2Help.put(SPEECHENDLIKELIHOOD, "likelihood of speech ending event"
				+ "Range [0.0,1.0], decrease to get more silence after speech segments");

		props2Help.put(SHIFTFROMMINIMUMENERGYCENTER,
				"multiplied by lowest energy cluster mean to generate speech/silence energy threshold"
						+ "Range [0.0,5.0], decrease to get more silence in speech segments");

		props2Help.put(NUMENERGYCLUSTERS, "number of energy clusters"
				+ "Range [1,20], decrease to get more silence in speech segments");

		props2Help.put(MINIMUMSTARTSILENCEINSECONDS, "minimum silence in the beginning of the output files in seconds"
				+ "Range [0.0,30.0], increase to get more silence in the beginning");

		props2Help.put(MINIMUMENDSILENCEINSECONDS, "minimum silence at the end of the output files in seconds"
				+ "Range [0.0,30.0], increase to get more silence at the end");
	}

	public boolean compute() throws IOException, UnsupportedAudioFileException {
		// Check existance of input directory
		inputWavDir = new File(getProp(INPUTWAVDIR));
		if (!inputWavDir.exists()) {
			throw new Error("Could not find input Directory: " + getProp(INPUTWAVDIR));
		}

		// Check existance of output directory
		// if not exists, create a new directory
		outputWavDir = new File(getProp(OUTPUTWAVDIR));
		if (!outputWavDir.exists()) {
			System.out.print(OUTPUTWAVDIR + " " + getProp(OUTPUTWAVDIR) + " does not exist; ");
			if (!outputWavDir.mkdir()) {
				throw new Error("Could not create OUTPUTWAVDIR");
			}
			System.out.print("Created successfully.\n");
		}

		// Automatically collect all ".wav" files from given directory
		bnlist = new BasenameList(inputWavDir + File.separator, waveExt);

		int energyBufferLength = Integer.valueOf(getProp(ENERGYBUFFERLENGTH));
		energyBufferLength = MathUtils.CheckLimits(energyBufferLength, 1, 1000);

		double speechStartLikelihood = Double.valueOf(getProp(SPEECHSTARTLIKELIHOOD));
		speechStartLikelihood = MathUtils.CheckLimits(speechStartLikelihood, 0.0, 1.0);

		double speechEndLikelihood = Double.valueOf(getProp(SPEECHENDLIKELIHOOD));
		speechEndLikelihood = MathUtils.CheckLimits(speechEndLikelihood, 0.0, 1.0);

		double shiftFromMinimumEnergyCenter = Double.valueOf(getProp(SHIFTFROMMINIMUMENERGYCENTER));
		shiftFromMinimumEnergyCenter = MathUtils.CheckLimits(shiftFromMinimumEnergyCenter, 0.0, 5.0);

		int numClusters = Integer.valueOf(getProp(NUMENERGYCLUSTERS));
		numClusters = MathUtils.CheckLimits(numClusters, 1, 20);

		double minimumStartSilenceInSeconds = Double.valueOf(getProp(MINIMUMSTARTSILENCEINSECONDS));
		minimumStartSilenceInSeconds = MathUtils.CheckLimits(minimumStartSilenceInSeconds, 0.0, 30.0);

		double minimumEndSilenceInSeconds = Double.valueOf(getProp(MINIMUMENDSILENCEINSECONDS));
		minimumEndSilenceInSeconds = MathUtils.CheckLimits(minimumEndSilenceInSeconds, 0.0, 30.0);
		//

		System.out.println("Removing endpoints for " + bnlist.getLength() + " wave files");

		for (int i = 0; i < bnlist.getLength(); i++) {
			percent = 100 * i / bnlist.getLength();
			String inputFile = inputWavDir + File.separator + bnlist.getName(i) + waveExt;
			String outputFile = outputWavDir + File.separator + bnlist.getName(i) + waveExt;

			AudioConverterUtils.removeEndpoints(inputFile, outputFile, energyBufferLength, speechStartLikelihood,
					speechEndLikelihood, shiftFromMinimumEnergyCenter, numClusters, minimumStartSilenceInSeconds,
					minimumEndSilenceInSeconds);

			System.out.println("    " + bnlist.getName(i));
		}

		System.out.println("...Done.");

		return true;
	}

	/**
	 * Provide the progress of computation, in percent, or -1 if that feature is not implemented.
	 * 
	 * @return -1 if not implemented, or an integer between 0 and 100.
	 */
	public int getProgress() {
		return percent;
	}

}




© 2015 - 2025 Weber Informatics LLC | Privacy Policy