Recherche avancée

Sur d’autres sites (10378)

• There is no data in the inbound-rtp section of WebRTC. I don't know why

  13 juin 2024, par qyt

  I am a streaming media server, and I need to stream video to WebRTC in H.264 format. The SDP exchange has no errors, and Edge passes normally.

  


  These are the log debugging details from edge://webrtc-internals/. Both DTLS and STUN show normal status, and SDP exchange is also normal. I used Wireshark to capture packets and saw that data streaming has already started. The transport section (iceState=connected, dtlsState=connected, id=T01) also shows that data has been received, but there is no display of RTP video data at all.

  


  timestamp   2024/6/13 16:34:01
bytesSent   5592
[bytesSent_in_bits/s]   176.2108579387652
packetsSent 243
[packetsSent/s] 1.001198056470257
bytesReceived   69890594
[bytesReceived_in_bits/s]   0
packetsReceived 49678
[packetsReceived/s] 0
dtlsState   connected
selectedCandidatePairId CPeVYPKUmD_FoU/ff10
localCertificateId  CFE9:17:14:B4:62:C3:4C:FF:90:C0:57:50:ED:30:D3:92:BC:BB:7C:13:11:AB:07:E8:28:3B:F6:A5:C7:66:50:77
remoteCertificateId CF09:0C:ED:3E:B3:AC:33:87:2F:7E:B0:BD:76:EB:B5:66:B0:D8:60:F7:95:99:52:B5:53:DA:AC:E7:75:00:09:07
tlsVersion  FEFD
dtlsCipher  TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256
dtlsRole    client
srtpCipher  AES_CM_128_HMAC_SHA1_80
selectedCandidatePairChanges    1
iceRole controlling
iceLocalUsernameFragment    R5DR
iceState    connected


  


  video recv info

  


  inbound-rtp (kind=video, mid=1, ssrc=2124085007, id=IT01V2124085007)
Statistics IT01V2124085007
timestamp   2024/6/13 16:34:49
ssrc    2124085007
kind    video
transportId T01
jitter  0
packetsLost 0
trackIdentifier 1395f18c-6ab9-4dbc-9149-edb59a81044d
mid 1
packetsReceived 0
[packetsReceived/s] 0
bytesReceived   0
[bytesReceived_in_bits/s]   0
headerBytesReceived 0
[headerBytesReceived_in_bits/s] 0
jitterBufferDelay   0
[jitterBufferDelay/jitterBufferEmittedCount_in_ms]  0
jitterBufferTargetDelay 0
[jitterBufferTargetDelay/jitterBufferEmittedCount_in_ms]    0
jitterBufferMinimumDelay    0
[jitterBufferMinimumDelay/jitterBufferEmittedCount_in_ms]   0
jitterBufferEmittedCount    0
framesReceived  0
[framesReceived/s]  0
[framesReceived-framesDecoded-framesDropped]    0
framesDecoded   0
[framesDecoded/s]   0
keyFramesDecoded    0
[keyFramesDecoded/s]    0
framesDropped   0
totalDecodeTime 0
[totalDecodeTime/framesDecoded_in_ms]   0
totalProcessingDelay    0
[totalProcessingDelay/framesDecoded_in_ms]  0
totalAssemblyTime   0
[totalAssemblyTime/framesAssembledFromMultiplePackets_in_ms]    0
framesAssembledFromMultiplePackets  0
totalInterFrameDelay    0
[totalInterFrameDelay/framesDecoded_in_ms]  0
totalSquaredInterFrameDelay 0
[interFrameDelayStDev_in_ms]    0
pauseCount  0
totalPausesDuration 0
freezeCount 0
totalFreezesDuration    0
firCount    0
pliCount    0
nackCount   0
minPlayoutDelay 0


  


  wireshark,I have verified that the SSRC in the SRTP is correct.

  


  enter image description here

  


  This player works normally when tested with other streaming servers. I don't know what the problem is. Is there any way to find out why the web browser cannot play the WebRTC stream that I'm pushing ?

  


• PyQt6 6.7.0 - How to fix error : No QtMultimedia backends found

  4 février, par Belleroph0N

  Problem on Windows 10 and Windows 11 using Anaconda.

  


  Here is the full error message for PyQt6=6.7.0 :

  


  No QtMultimedia backends found. Only QMediaDevices, QAudioDevice, QSoundEffect, QAudioSink, and QAudioSource are available.
Failed to initialize QMediaPlayer "Not available"
Failed to create QVideoSink "Not available"


  


  Installed PyQt6 using a requirements file :

  


  PyQt6
PyQt6-WebEngine
requests
pyserial
pynput


  


  Here are a couple things I tried :

  


  

  1. Reroll version back to PyQt6=6.6.1. This results in an error as well : ImportError : DLL load failed while importing QtGui : The specified procedure could not be found.

  


  2. I thought that missing ffmpeg might be the issue so I installed it, but the issue persists.

  


  3. Tried the setup on Ubuntu (WSL2) and the issue disappears, but there is just a black screen and nothing gets displayed in the widget. (EDIT : Got this up and running, the problem was with differences in file paths in linux vs windows.)

  


  


  I am new to PyQt so any pointers will be helpful !

  


  Edit : Here is generic code (taken from here) that gives the same error :

  


  from PyQt6.QtGui import QIcon, QFont
from PyQt6.QtCore import QDir, Qt, QUrl, QSize
from PyQt6.QtMultimedia import QMediaPlayer
from PyQt6.QtMultimediaWidgets import QVideoWidget
from PyQt6.QtWidgets import (QApplication, QFileDialog, QHBoxLayout, QLabel, QStyleFactory,
        QPushButton, QSizePolicy, QSlider, QStyle, QVBoxLayout, QWidget, QStatusBar)


class VideoPlayer(QWidget):

    def __init__(self, parent=None):
        super(VideoPlayer, self).__init__(parent)

        self.mediaPlayer = QMediaPlayer()

        btnSize = QSize(16, 16)
        videoWidget = QVideoWidget()

        openButton = QPushButton("Open Video")   
        openButton.setToolTip("Open Video File")
        openButton.setStatusTip("Open Video File")
        openButton.setFixedHeight(24)
        openButton.setIconSize(btnSize)
        openButton.setFont(QFont("Noto Sans", 8))
        openButton.setIcon(QIcon.fromTheme("document-open", QIcon("D:/_Qt/img/open.png")))
        openButton.clicked.connect(self.abrir)

        self.playButton = QPushButton()
        self.playButton.setEnabled(False)
        self.playButton.setFixedHeight(24)
        self.playButton.setIconSize(btnSize)
        self.playButton.setIcon(self.style().standardIcon(QStyle.StandardPixmap.SP_MediaPlay))
        self.playButton.clicked.connect(self.play)

        self.positionSlider = QSlider(Qt.Orientation.Horizontal)
        self.positionSlider.setRange(0, 0)
        self.positionSlider.sliderMoved.connect(self.setPosition)

        self.statusBar = QStatusBar()
        self.statusBar.setFont(QFont("Noto Sans", 7))
        self.statusBar.setFixedHeight(14)

        controlLayout = QHBoxLayout()
        controlLayout.setContentsMargins(0, 0, 0, 0)
        controlLayout.addWidget(openButton)
        controlLayout.addWidget(self.playButton)
        controlLayout.addWidget(self.positionSlider)

        layout = QVBoxLayout()
        layout.addWidget(videoWidget)
        layout.addLayout(controlLayout)
        layout.addWidget(self.statusBar)

        self.setLayout(layout)

        #help(self.mediaPlayer)
        self.mediaPlayer.setVideoOutput(videoWidget)
        self.mediaPlayer.playbackStateChanged.connect(self.mediaStateChanged)
        self.mediaPlayer.positionChanged.connect(self.positionChanged)
        self.mediaPlayer.durationChanged.connect(self.durationChanged)
        self.mediaPlayer.errorChanged.connect(self.handleError)
        self.statusBar.showMessage("Ready")

    def abrir(self):
        fileName, _ = QFileDialog.getOpenFileName(self, "Select Media",
                ".", "Video Files (*.mp4 *.flv *.ts *.mts *.avi)")

        if fileName != '':
            self.mediaPlayer.setSource(QUrl.fromLocalFile(fileName))
            self.playButton.setEnabled(True)
            self.statusBar.showMessage(fileName)
            self.play()

    def play(self):
        if self.mediaPlayer.playbackState() == QMediaPlayer.PlaybackState.PlayingState:
            self.mediaPlayer.pause()
        else:
            self.mediaPlayer.play()

    def mediaStateChanged(self, state):
        if self.mediaPlayer.playbackState() == QMediaPlayer.PlaybackState.PlayingState:
            self.playButton.setIcon(
                    self.style().standardIcon(QStyle.StandardPixmap.SP_MediaPause))
        else:
            self.playButton.setIcon(
                    self.style().standardIcon(QStyle.StandardPixmap.SP_MediaPlay))

    def positionChanged(self, position):
        self.positionSlider.setValue(position)

    def durationChanged(self, duration):
        self.positionSlider.setRange(0, duration)

    def setPosition(self, position):
        self.mediaPlayer.setPosition(position)

    def handleError(self):
        self.playButton.setEnabled(False)
        self.statusBar.showMessage("Error: " + self.mediaPlayer.errorString())

if __name__ == '__main__':
    import sys
    app = QApplication(sys.argv)
    player = VideoPlayer()
    player.setWindowTitle("Player")
    player.resize(900, 600)
    player.show()
    sys.exit(app.exec())


  


  The videos I want to play are in the same folder as this .py file.
The conda env (python 3.9.2) I am working on has the following packages :

  


  certifi                     2024.6.2
charset-normalizer          3.3.2
idna                        3.7
pip                         24.0
pynput                      1.7.6
PyQt6                       6.7.0
PyQt6-Qt6                   6.7.1
PyQt6-sip                   13.6.0
PyQt6-WebEngine             6.7.0
PyQt6-WebEngine-Qt6         6.7.1
PyQt6-WebEngineSubwheel-Qt6 6.7.1
pyserial                    3.5
requests                    2.31.0
setuptools                  69.5.1
six                         1.16.0
urllib3                     2.2.1
wheel                       0.43.0


  


  PS : MacOS seems to have the same issue.

  


• ffmpeg fast seek large MP4 over HTTP

  28 juillet 2024, par Gmanicus

  I'm attempting to download snapshots from a video provided by the U.S House of Representatives :

  


  https://houseliveprod-f9h4cpb9dyb8gegg.a01.azurefd.net/east/2024-04-11T08-55-12_Download/video_3000000_1.mp4


  


  I am using fluent-ffmpeg in Node to execute this command :

  


  ffmpeg('https://houseliveprod-f9h4cpb9dyb8gegg.a01.azurefd.net/east/2024-04-11T08-55-12_Download/video_3000000_1.mp4')
  .inputOption(`-ss 03:33:33`)
  .outputOptions([
     '-vframes 1'
  ])
  .output('test.png')

// Effectively:
// ffmpeg -ss 03:33:33 -i  -y -vframes 1 test.png


  


  My intention is to fast-seek to the desired timestamp and take a snapshot over HTTP. However, when doing so, the performance is not great. A snapshot takes about 10s per 3hrs of video and seems to increase fairly linearly at that rate.

  


  However, when using ffmpeg on the same video locally, it's super fast ! Sub-500ms regardless of the desired timestamp.

  


  Is there some magic that could be done via ffmpeg options or perhaps some sort of technique with manual requests to get a snapshot at the desired segment of video more efficiently ?