使用 NAO 执行远程音频处理时出现问题

Posted 2023-02-25

【中文标题】使用 NAO 执行远程音频处理时出现问题

【英文标题】：Troubles performing remote audio processing with NAO

【发布时间】：2019-06-28 18:01:34

【问题描述】：

我正在尝试使用 Python naoqi SDK 的 2.1.4.13 版本通过 NAO v4 执行一些远程音频处理，但我在这样做时遇到了困难。我尝试使用 Alexandre Mazel 在NAO robot remote audio problems 提供的解决方案，但我仍然无法从缓冲区中检索音频数据。帖子中引用的代码在下面可用。我还按照https://www.generationrobots.com/media/NAO%20Next%20Gen/FeaturePaper(AudiosignalProcessing)%20(1).pdf中的远程处理音频数据的说明进行操作

我非常感谢任何帮助或解决方案，因为我已经被这个问题困扰了好几天了。

 # -*- coding: utf-8 -*-

###########################################################
# Retrieve robot audio buffer
# Syntaxe:
#    python scriptname --pip <ip> --pport <port>
# 
#    --pip <ip>: specify the ip of your robot (without specification it will use the NAO_IP defined some line below
#
# Author: Alexandre Mazel
###########################################################

NAO_IP = "10.0.252.126" # Romeo on table
#~ NAO_IP = "10.0.253.99" # Nao Alex Blue


from optparse import OptionParser
import naoqi
import numpy as np
import time
import sys


class SoundReceiverModule(naoqi.ALModule):
    """
    Use this object to get call back from the ALMemory of the naoqi world.
    Your callback needs to be a method with two parameter (variable name, value).
    """

    def __init__( self, strModuleName, strNaoIp ):
        try:
            naoqi.ALModule.__init__(self, strModuleName );
            self.BIND_PYTHON( self.getName(),"callback" );
            self.strNaoIp = strNaoIp;
            self.outfile = None;
            self.aOutfile = [None]*(4-1); # ASSUME max nbr channels = 4
        except BaseException, err:
            print( "ERR: abcdk.naoqitools.SoundReceiverModule: loading error: %s" % str(err) );

    # __init__ - end
    def __del__( self ):
        print( "INF: abcdk.SoundReceiverModule.__del__: cleaning everything" );
        self.stop();

    def start( self ):
        audio = naoqi.ALProxy( "ALAudioDevice", self.strNaoIp, 9559 );
        nNbrChannelFlag = 0; # ALL_Channels: 0,  AL::LEFTCHANNEL: 1, AL::RIGHTCHANNEL: 2; AL::FRONTCHANNEL: 3  or AL::REARCHANNEL: 4.
        nDeinterleave = 0;
        nSampleRate = 48000;
        audio.setClientPreferences( self.getName(),  nSampleRate, nNbrChannelFlag, nDeinterleave ); # setting same as default generate a bug !?!
        audio.subscribe( self.getName() );
        print( "INF: SoundReceiver: started!" );
        # self.processRemote( 4, 128, [18,0], "A"*128*4*2 ); # for local test

        # on romeo, here's the current order:
        # 0: right;  1: rear;   2: left;   3: front,  

    def stop( self ):
        print( "INF: SoundReceiver: stopping..." );
        audio = naoqi.ALProxy( "ALAudioDevice", self.strNaoIp, 9559 );
        audio.unsubscribe( self.getName() );        
        print( "INF: SoundReceiver: stopped!" );
        if( self.outfile != None ):
            self.outfile.close();


    def processRemote( self, nbOfChannels, nbrOfSamplesByChannel, aTimeStamp, buffer ):
        """
        This is THE method that receives all the sound buffers from the "ALAudioDevice" module
        """
        #~ print( "process!" );
        #~ print( "processRemote: %s, %s, %s, lendata: %s, data0: %s (0x%x), data1: %s (0x%x)" % (nbOfChannels, nbrOfSamplesByChannel, aTimeStamp, len(buffer), buffer[0],ord(buffer[0]),buffer[1],ord(buffer[1])) );
        #~ print( "raw data: " ),
        #~ for i in range( 8 ):
            #~ print( "%s (0x%x), " % (buffer[i],ord(buffer[i])) ),
        #~ print( "" );

        aSoundDataInterlaced = np.fromstring( str(buffer), dtype=np.int16 );
        #~ print( "len data: %s " % len( aSoundDataInterlaced ) );
        #~ print( "data interlaced: " ),
        #~ for i in range( 8 ):
            #~ print( "%d, " % (aSoundDataInterlaced[i]) ),
        #~ print( "" );
        aSoundData = np.reshape( aSoundDataInterlaced, (nbOfChannels, nbrOfSamplesByChannel), 'F' );
        #~ print( "len data: %s " % len( aSoundData ) );
        #~ print( "len data 0: %s " % len( aSoundData[0] ) );
        if( False ):
            # compute average
            aAvgValue = np.mean( aSoundData, axis = 1 );
            print( "avg: %s" % aAvgValue );
        if( False ):
            # compute fft
            nBlockSize = nbrOfSamplesByChannel;
            signal = aSoundData[0] * np.hanning( nBlockSize );
            aFft = ( np.fft.rfft(signal) / nBlockSize );
            print aFft;
        if( False ):
            # compute peak
            aPeakValue = np.max( aSoundData );
            if( aPeakValue > 16000 ):
                print( "Peak: %s" % aPeakValue );
        if( True ):
            bSaveAll = True;
            # save to file
            if( self.outfile == None ):
                strFilenameOut = "/out.raw";
                print( "INF: Writing sound to '%s'" % strFilenameOut );
                self.outfile = open( strFilenameOut, "wb" );
                if( bSaveAll ):
                    for nNumChannel in range( 1, nbOfChannels ):
                        strFilenameOutChan = strFilenameOut.replace(".raw", "_%d.raw"%nNumChannel);
                        self.aOutfile[nNumChannel-1] = open( strFilenameOutChan, "wb" );
                        print( "INF: Writing other channel sound to '%s'" % strFilenameOutChan );

            #~ aSoundDataInterlaced.tofile( self.outfile ); # wrote the 4 channels
            aSoundData[0].tofile( self.outfile ); # wrote only one channel
            #~ print( "aTimeStamp: %s" % aTimeStamp );
            #~ print( "data wrotten: " ),
            #~ for i in range( 8 ):
                #~ print( "%d, " % (aSoundData[0][i]) ),
            #~ print( "" );            
            #~ self.stop(); # make naoqi crashes
            if( bSaveAll ):
                for nNumChannel in range( 1, nbOfChannels ):
                    aSoundData[nNumChannel].tofile( self.aOutfile[nNumChannel-1] ); 


    # processRemote - end


    def version( self ):
        return "0.6";

# SoundReceiver - end


def main():
    """ Main entry point

    """
    parser = OptionParser()
    parser.add_option("--pip",
        help="Parent broker port. The IP address or your robot",
        dest="pip")
    parser.add_option("--pport",
        help="Parent broker port. The port NAOqi is listening to",
        dest="pport",
        type="int")
    parser.set_defaults(
        pip=NAO_IP,
        pport=9559)

    (opts, args_) = parser.parse_args()
    pip   = opts.pip
    pport = opts.pport

    # We need this broker to be able to construct
    # NAOqi modules and subscribe to other modules
    # The broker must stay alive until the program exists
    myBroker = naoqi.ALBroker("myBroker",
       "0.0.0.0",   # listen to anyone
       0,           # find a free port and use it
       pip,         # parent broker IP
       pport)       # parent broker port


    # Warning: SoundReceiver must be a global variable
    # The name given to the constructor must be the name of the
    # variable
    global SoundReceiver
    SoundReceiver = SoundReceiverModule("SoundReceiver", pip)
    SoundReceiver.start()

    try:
        while True:
            time.sleep(1)
    except KeyboardInterrupt:
        print
        print "Interrupted by user, shutting down"
        myBroker.shutdown()
        sys.exit(0)



if __name__ == "__main__":
    main()

【问题讨论】：

“但我仍然无法从缓冲区中检索音频数据。”是否抛出错误？缓冲区是空的吗？你能访问缓冲区吗？你有什么具体问题？

嗨安德鲁，没有错误被抛出。根据 NAOqi 文档，模块 ALAudioDevice 应该通过 processRemote 函数将音频数据发送到缓冲区。但是，它似乎没有将这些数据传递给函数。我可以说，因为当我尝试将原始数据转换为 wav 文件时，生成的 wav 文件是空的。

【参考方案1】：

也许这个项目可以帮助你定位：

https://github.com/UNSWComputing/rUNSWift-2015-release/wiki/Whistle-Detection

它提供了一个whistle_detector.py Python 模块，

也可以在 2.1 工具链/Nao V4 下的 Nao 上运行。

这个项目也值得一看：

https://www.ibm.com/blogs/watson/2016/07/getting-robots-listen-using-watsons-speech-text-service/

它使用一种方法是调用linux命令arecord而不是使用ALAudioDevice

【讨论】：

非常感谢。这应该足以让我指出正确的方向。我想说明我使用的是 Windows 操作系统，那么有什么类似于 Windows 的 arecord 的吗？

欢迎您！第二个项目的脚本在运行 linux 的机器人上执行。因此，命令“arecord”也在机器人 linux 上执行。您的开发设备的操作系统无关紧要。

【参考方案2】：

您是否尝试过documentation 中的示例？

尝试在机器人上运行：

#! /usr/bin/env python
# -*- encoding: UTF-8 -*-

"""Example: Get Signal from Front Microphone & Calculate its rms Power"""


import qi
import argparse
import sys
import time
import numpy as np


class SoundProcessingModule(object):
    """
    A simple get signal from the front microphone of Nao & calculate its rms power.
    It requires numpy.
    """

    def __init__( self, app):
        """
        Initialise services and variables.
        """
        super(SoundProcessingModule, self).__init__()
        app.start()
        session = app.session

        # Get the service ALAudioDevice.
        self.audio_service = session.service("ALAudioDevice")
        self.isProcessingDone = False
        self.nbOfFramesToProcess = 20
        self.framesCount=0
        self.micFront = []
        self.module_name = "SoundProcessingModule"

    def startProcessing(self):
        """
        Start processing
        """
        # ask for the front microphone signal sampled at 16kHz
        # if you want the 4 channels call setClientPreferences(self.module_name, 48000, 0, 0)
        self.audio_service.setClientPreferences(self.module_name, 16000, 3, 0)
        self.audio_service.subscribe(self.module_name)

        while self.isProcessingDone == False:
            time.sleep(1)

        self.audio_service.unsubscribe(self.module_name)

    def processRemote(self, nbOfChannels, nbOfSamplesByChannel, timeStamp, inputBuffer):
        """
        Compute RMS from mic.
        """
        self.framesCount = self.framesCount + 1

        if (self.framesCount <= self.nbOfFramesToProcess):
            # convert inputBuffer to signed integer as it is interpreted as a string by python
            self.micFront=self.convertStr2SignedInt(inputBuffer)
            #compute the rms level on front mic
            rmsMicFront = self.calcRMSLevel(self.micFront)
            print "rms level mic front = " + str(rmsMicFront)
        else :
            self.isProcessingDone=True

    def calcRMSLevel(self,data) :
        """
        Calculate RMS level
        """
        rms = 20 * np.log10( np.sqrt( np.sum( np.power(data,2) / len(data)  )))
        return rms

    def convertStr2SignedInt(self, data) :
        """
        This function takes a string containing 16 bits little endian sound
        samples as input and returns a vector containing the 16 bits sound
        samples values converted between -1 and 1.
        """
        signedData=[]
        ind=0;
        for i in range (0,len(data)/2) :
            signedData.append(data[ind]+data[ind+1]*256)
            ind=ind+2

        for i in range (0,len(signedData)) :
            if signedData[i]>=32768 :
                signedData[i]=signedData[i]-65536

        for i in range (0,len(signedData)) :
            signedData[i]=signedData[i]/32768.0

        return signedData


if __name__ == "__main__":
    parser = argparse.ArgumentParser()
    parser.add_argument("--ip", type=str, default="127.0.0.1",
                        help="Robot IP address. On robot or Local Naoqi: use '127.0.0.1'.")
    parser.add_argument("--port", type=int, default=9559,
                        help="Naoqi port number")

    args = parser.parse_args()
    try:
        # Initialize qi framework.
        connection_url = "tcp://" + args.ip + ":" + str(args.port)
        app = qi.Application(["SoundProcessingModule", "--qi-url=" + connection_url])
    except RuntimeError:
        print ("Can't connect to Naoqi at ip \"" + args.ip + "\" on port " + str(args.port) +".\n"
               "Please check your script arguments. Run with -h option for help.")
        sys.exit(1)
    MySoundProcessingModule = SoundProcessingModule(app)
    app.session.registerService("SoundProcessingModule", MySoundProcessingModule)
    MySoundProcessingModule.startProcessing()

【讨论】：

是的，我试过使用这个例子。但是，它不起作用，因为我使用的是与我正在使用的 NAO 机器人兼容的旧版本的 python naoqi SDK