第12篇:Linux Kazam屏幕录制+声音采集设置

Kazam+PulseAudio组合几乎成了所有Linux发布版的屏幕录制+声音采集的首选工具,但到了Ubuntu 20.04之后。 Kazam潜在一个Bug,那就是在“首选项”下,我看不到“扬声器”和“麦克风”。那么导致你在录制过程中无法同步采集声音。

如果你对Kazam的源代码了解的话,它底层还是基于Python写的,其音频模块位于如下路径的

/usr/lib/python3/dist-packages/kazam/pulseaudio/pulseaudio.py

造成该bug的原因是原pulseaudio.py模块中的相关代码仍然调用time.clock()函数。 但从Python3.3起该函数已弃用,并且Ubuntu20.04操作系统已经默认预装python3.7+,那么修复该Bug非常简单,只要模块中的所有time.clock()调用都必须替换为time.perf_counter()

下面是修改好的pulseaudio.py模块的完整代码,只要覆盖原有的模块即可。

# -*- coding: utf-8 -*-
#
#       pulseaudio.py
#
#       Copyright 2012 David Klasinc <bigwhale@lubica.net>
#
#       This program is free software; you can redistribute it and/or modify
#       it under the terms of the GNU General Public License as published by
#       the Free Software Foundation; either version 3 of the License, or
#       (at your option) any later version.
#
#       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 General Public License for more details.
#
#       You should have received a copy of the GNU General Public License
#       along with this program; if not, write to the Free Software
#       Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
#       MA 02110-1301, USA.

import time
import logging
logger = logging.getLogger("PulseAudio")

from kazam.pulseaudio.error_handling import *
from kazam.backend.prefs import *

try:
    from kazam.pulseaudio.ctypes_pulseaudio import *
except:
    raise PAError(PA_LOAD_ERROR, "Unable to load pulseaudio wrapper lib. Is PulseAudio installed?")

class pulseaudio_q:
    def __init__(self):
        """pulseaudio_q constructor.

        Initializes and sets all the necessary startup variables.

        Args:
            None
        Returns:
            None
        Raises:
            None
        """

        self.pa_state = -1
        self.sources = []
        self._sources = []
        self._return_result = []
        self.pa_status = PA_STOPPED

        #
        # Making sure that we don't lose references to callback functions
        #
        self._pa_state_cb = pa_context_notify_cb_t(self.pa_state_cb)
        self._pa_sourcelist_cb = pa_source_info_cb_t(self.pa_sourcelist_cb)
        self._pa_sourceinfo_cb = pa_source_info_cb_t(self.pa_sourceinfo_cb)
        self._pa_context_success_cb = pa_context_success_cb_t(self.pa_context_success_cb)

    def pa_context_success_cb(self, context, c_int, user_data):
        self._pa_ctx_success = c_int
        return

    def pa_state_cb(self, context, userdata):
        """Reads PulseAudio context state.

        Sets self.pa_state depending on the pa_context_state and
        raises an error if unable to get the state from PulseAudio.

        Args:
            context: PulseAudio context.
            userdata: n/a.

        Returns:
            Zero on success or raises an exception.

        Raises:
            PAError, PA_GET_STATE_ERROR if pa_context_get_state() failed.
        """
        try:
            state = pa_context_get_state(context)

            if state in [PA_CONTEXT_UNCONNECTED, PA_CONTEXT_CONNECTING, PA_CONTEXT_AUTHORIZING,
                         PA_CONTEXT_SETTING_NAME]:
                self.pa_state = PA_STATE_WORKING
            elif state == PA_CONTEXT_FAILED:
                self.pa_state = PA_STATE_FAILED
            elif state == PA_CONTEXT_READY:
                self.pa_state = PA_STATE_READY
                logger.debug("State connected.")
        except:
            raise PAError(PA_GET_STATE_ERROR, "Unable to read context state.")

        return  0

    def pa_sourcelist_cb(self, context, source_info, eol, userdata):
        """Source list callback function

        Called by mainloop thread each time list of audio sources is requested.
        All the parameters to this functions are passed to it automatically by
        the caller.

        Args:
            context: PulseAudio context.
            source_info: data returned from mainloop.
            eol: End Of List marker if set to non-zero there is no more date
            to read and we should bail out.
            userdata: n/a.

        Returns:
            self.source_list: Contains list of all Pulse Audio sources.
            self.pa_status: PA_WORKING or PA_FINISHED

        Raises:
            None
        """
        if eol == 0:
            logger.debug("pa_sourcelist_cb()")
            logger.debug("  IDX: {0}".format(source_info.contents.index))
            logger.debug("  Name: {0}".format(source_info.contents.name))
            logger.debug("  Desc: {0}".format(source_info.contents.description))
            self.pa_status = PA_WORKING
            self._sources.append([source_info.contents.index,
                                 source_info.contents.name.decode('utf-8'),
                                 " ".join(source_info.contents.description.decode('utf-8').split())])
        else:
            logger.debug("pa_sourcelist_cb() -- finished")
            self.pa_status = PA_FINISHED

        return 0

    def pa_sourceinfo_cb(self, context, source_info, eol, userdata):
        """Source list callback function

        Called by mainloop thread each time info for a single audio source is requestd.
        All the parameters to this functions are passed to it automatically by
        the caller. This is here for convenience.

        Args:
            context: PulseAudio context.
            index: Source index
            source_info: data returned from mainloop.
            eol: End Of List marker if set to non-zero there is no more date
            to read and we should bail out.
            userdata: n/a.

        Returns:
            self.source_list: Contains list of all Pulse Audio sources.
            self.pa_status: PA_WORKING or PA_FINISHED

        Raises:
            None
        """
        if eol == 0:
            logger.debug("pa_sourceinfo_cb()")
            logger.debug("  IDX: {0}".format(source_info.contents.index))
            logger.debug("  Name: {0}".format(source_info.contents.name))
            logger.debug("  Desc: {0}".format(source_info.contents.description))
            self.pa_status = PA_WORKING
            cvolume = pa_cvolume()
            v = pa_volume_t * 32
            cvolume.channels = source_info.contents.volume.channels
            cvolume.values = v()
            for i in range(0, source_info.contents.volume.channels):
                cvolume.values[i] = source_info.contents.volume.values[i]


            self._return_result  = [source_info.contents.index,
                                    source_info.contents.name.decode('utf-8'),
                                    cvolume,
                                    " ".join(source_info.contents.description.decode('utf-8').split())]
        else:
            try:
                logger.debug("pa_sourceinfo_cb() -- Hit EOL")
                logger.debug("  EOL IDX: {0}".format(source_info.contents.index))
                logger.debug("  EOL Name: {0}".format(source_info.contents.name))
                logger.debug("  EOL Desc: {0}".format(source_info.contents.description))
            except:
                logger.debug("pa_sourceinfo_cb() -- EOL no data!")
            self.pa_status = PA_FINISHED
        logger.debug("pa_sourceinfo_cb() -- finished")
        return 0

    def start(self):
        """Starts PulseAudio threaded mainloop.

        Creates mainloop, mainloop API and context objects and connects
        to the PulseAudio server.

        Args:
            None

        Returns:
            None

        Raises:
            PAError, PA_STARTUP_ERROR - if unable to create PA objects.
            PAError, PA_UNABLE_TO_CONNECT - if connection to PA fails.
            PAError, PA_UNABLE_TO_CONNECT2 - if call to connect() fails.
            PAError, PA_MAINLOOP_START_ERROR - if not able to start mainloop.
        """
        try:
            logger.debug("Starting mainloop.")
            self.pa_ml = pa_threaded_mainloop_new()
            logger.debug("Getting API.")
            self.pa_mlapi = pa_threaded_mainloop_get_api(self.pa_ml)
            logger.debug("Setting context.")
            self.pa_ctx = pa_context_new(self.pa_mlapi, None)
            logger.debug("Set state callback.")
            pa_context_set_state_callback(self.pa_ctx, self._pa_state_cb, None)
        except:
            raise PAError(PA_STARTUP_ERROR, "Unable to access PulseAudio API.")

        try:
            logger.debug("Connecting to server.")
            if pa_context_connect(self.pa_ctx, None, 0, None):
                raise PAError(PA_UNABLE_TO_CONNECT, "Unable to connect to PulseAudio server.")
        except:
            raise PAError(PA_UNABLE_TO_CONNECT2, "Unable to initiate connection to PulseAudio server.")
        try:
            logger.debug("Start mainloop.")
            pa_threaded_mainloop_start(self.pa_ml)
            time.sleep(0.1)  # Mainloop needs some time to start ...
            pa_context_get_state(self.pa_ctx)
        except:
            raise PAError(PA_MAINLOOP_START_ERROR, "Unable to start mainloop.")

    def end(self):
        """Disconnects from PulseAudio server.

        Disconnects from PulseAudio server, it should be called after all the
        operations are finished.

        Args:
            None

        Returns:
            None

        Raises:
            PAError, PA_MAINLOOP_END_ERROR - if not able to disconnect.
        """
        try:
            logger.debug("Disconnecting from server.")
            pa_context_disconnect(self.pa_ctx)
            self.pa_ml = None
            self.pa_mlapi = None
            self.pa_ctx = None
        except:
            raise PAError(PA_MAINLOOP_END_ERROR, "Unable to end mainloop.")

    def get_audio_sources(self):
        try:
            logger.debug("get_audio_sources() called.")
            pa_context_get_source_info_list(self.pa_ctx, self._pa_sourcelist_cb, None)
            t = time.perf_counter()
            while time.perf_counter() - t < 5:
                if self.pa_status == PA_FINISHED:
                    self.sources = self._sources
                    self._sources = []
                    return self.sources
            raise PAError(PA_GET_SOURCES_TIMEOUT, "Unable to get sources, operation timed out.")
        except:
            logger.debug("Unable to get audio sources.")
            raise PAError(PA_GET_SOURCES_ERROR, "Unable to get sources.")

    def get_source_info_by_index(self, index):
        try:
            logger.debug("get_source_info_by_index() called. IDX: {0}".format(index))
            pa_context_get_source_info_by_index(self.pa_ctx, index, self._pa_sourceinfo_cb, None)
            t = time.perf_counter()
            while time.perf_counter() - t < 5:
                if self.pa_status == PA_FINISHED:
                    time.sleep(0.1)
                    ret = self._return_result
                    self._return_result = []
                    return ret
            raise PAError(PA_GET_SOURCE_TIMEOUT, "Unable to get source, operation timed out.")
        except:
            raise PAError(PA_GET_SOURCE_ERROR, "Unable to get source.")

    def set_source_volume_by_index(self, index, cvolume):
        try:
            pa_context_set_source_volume_by_index(self.pa_ctx, index, cvolume,
                                                  self._pa_context_success_cb, None)
            t = time.perf_counter()
            while time.perf_counter() - t < 5:
                if self.pa_status == PA_FINISHED:
                    return 1
            raise PAError(PA_GET_SOURCES_TIMEOUT, "Unable to get sources, operation timed out.")
        except:
            raise PAError(PA_GET_SOURCES_ERROR, "Unable to get sources.")

    def set_source_mute_by_index(self, index, mute):
        try:
            pa_context_set_source_mute_by_index(self.pa_ctx, index, mute,
                                                  self._pa_context_success_cb, None)
            t = time.perf_counter()
            while time.perf_counter() - t < 5:
                if self.pa_status == PA_FINISHED:
                    return 1
            raise PAError(PA_GET_SOURCES_TIMEOUT, "Unable to get sources, operation timed out.")
        except:
            raise PAError(PA_GET_SOURCES_ERROR, "Unable to get sources.")

    def cvolume_to_linear(self, cvolume):
        avg = 0
        for chn in range(cvolume.channels):
            avg = avg + cvolume.values[chn]
        avg = avg / cvolume.channels
        volume = pa_sw_volume_to_linear(uint32_t(int(avg)))
        return volume

    def cvolume_to_dB(self, cvolume):
        avg = 0
        for chn in range(cvolume.channels):
            avg = avg + cvolume.values[chn]
        avg = avg / cvolume.channels
        volume = pa_sw_volume_to_dB(uint32_t(int(avg)))
        return volume

    def linear_to_cvolume(self, index, volume):
        info = self.get_source_info_by_index(index)
        cvolume = pa_cvolume()
        v = pa_volume_t * 32
        cvolume.channels = info[2].channels
        cvolume.values = v()
        for i in range(0, info[2].channels):
            cvolume.values[i] = pa_sw_volume_from_linear(volume)
        return cvolume

    def dB_to_cvolume(self, channels, volume):
        cvolume = pa_cvolume()
        v = pa_volume_t * 32
        cvolume.channels = channels
        cvolume.values = v()
        value = pa_sw_volume_from_dB(volume)
        for i in range(0, channels):
            cvolume.values[i] = value
        return cvolume

将Kazam重启一次,我们看看Kazam的首选项


我们打开kazam,勾选[来自扬声器的声音],这里暂时不要点[捕获]按钮

接下来,我们先确定要录制的屏幕区域,例如,我经常选择区域录制,我们选择[区域]选项后,kazam会提醒我们需要拖动录制的区域大小,如下图。

声音的设定部分,PulseAudio安装之后,我们在任务栏的音量控制图标那里的音频列表

  • Speakers(Built-in Audio Analog Stereo),这是硬件层的物理扬声器,也就是你能听到的
  • Simultaneous output to Built-in Audio Analog Stereo,这是一个逻辑上的音频输出设备,这是PulseAudio驱动的虚拟扬声器

这样做的目地是让用户有多一个选择,我们在执行声音采集过程中,希望做到采集过程中做到安静采集,不想打扰身边其他人,此时我们只需要将虚拟扬声器设定为默认设备,并选择[通过次设备播放所有音频] 将操作系统的所有音频播放流重定向到该虚拟扬声器,此时我们将物理扬声器的音量调成静音即可。


不过过做上面的设定还是不够的的,我们打开音频控制的系统设置,将音频输出配置的如下选项都勾选上

  • 一个可以同步输出到所有本地声卡的虚拟输出设备
  • 当输出可用时,自动切换所有流

一切设定好后,点击[捕获]按钮,任务栏的录像按钮表示Kazam已经在后台执行屏幕已经在后台执行屏幕录像和声音采集了。

如果你发现你采取的视频没有声音或第一次使用kazam的话,请打开PulseAudio 音量控制面板,这个工具是安装PulseAudio时已经附带的,此时检测一下输入设备一栏,如果发现当前虚拟声卡输出的音量是跳动,并且录音一栏也有一个Python图标kazam的音量标也是跳动的。代码音声采集一切都是正常工作的。

我们看看取样的结果

  • 视频压制格式h.264,
  • 帧刷新率:50,对于视频录制格式建议不要低于50,因为会造成图像刷新时出现水平抖动的状况
  • 分辨率:跟你视频采集方式有关
  • 音频采样率:44100Hz,这个跟Kazam毫无关系,和操作系统的音频输出设置和PulseAudio的虚拟声卡设置有关。