android RTMP直播流媒体支持

软硬件环境

ubuntu 16.04
Android Studio 2.1.3
OTT BOx with android 5.1.1
nginx 1.11.3
nginx-rtmp-module
vitamio

前言

当下,直播已经成为网络热词,它不单单是指传统广播电视的实时播放,更是更为广泛的音视频实时分享的延伸。早先,直播数据源只可能来自于电视台及节目制作中心,但是放眼现在,基于计算机技术的高速发展,任何人都可以独自完成内容的制作,再利用身边的终端设备完成分享,你不仅仅是受众,同样可以成为主角。可以说,时下流行的”网红”文化,直播技术立下了汗马功劳。本文旨在搭建一个最简单的视频直播系统,包括服务器端及Android客户端,采用了nginx、nginx-rtmp、vitamio及ffmpeg。

rtmp协议

RTMP是Real Time Messaging Protocol的缩写,是被设计用来进行实时数据通信的网络协议。它是一个协议族,包括rtmpe、rtmpt、rtmps等,是直播技术中常用的协议

服务端配置

nginx添加rtmp支持

从http://nginx.org/en/download.html下载最新版1.11.3,然后从https://github.com/arut/nginx-rtmp-module下载nginx的rtmp补丁，下载的文件都放在目录/home/djstava(请根据实际情况自行修改)下，重新编译nginx

tar xvf nginx-1.11.3.tar.gz
cd nginx-1.11.3
mkdir build
./configure --prefix=/home/djstava/nginx-1.11.3/build --add-module=/home/djstava/nginx-rtmp-module
make
make install

修改配置文件nginx.conf

编辑/home/djstava/nginx-1.11.3/build/conf/nginx.conf,在文件末尾添加如下内容

rtmp {server {listen 1935;ping 30s;notify_method get;application myapp {live on;# sample play/publish handlers#on_play http://localhost:8080/on_play;#on_publish http://localhost:8080/on_publish;# sample recorder#recorder rec1 {#    record all;#    record_interval 30s;#    record_path /tmp;#    record_unique on;#}# sample HLS#hls on;#hls_path /tmp/hls;#hls_sync 100ms;}# Video on demand#application vod {#    play /var/Videos;#}# Video on demand over HTTP#application vod_http {#    play http://localhost:8080/vod/;#}}
}

启动nginx服务

/home/djstava/nginx-1.11.3/build/sbin/nginx

ffmpeg推送rtmp

找一个本地的视频文件进行推送，命令为

ffmpeg -re -i 大话西游之月光宝盒.BD1280超清国粤双语中英双字.mp4 -c copy -f flv rtmp://localhost/myapp/mystream

ffplay播放测试

如果没有安装ffplay的话，也可以用vlc

ffplay rtsmp://localhost/myapp/mystream

后记

前面的步骤都是在本机中进行的。可是在实际应用中，情况会复杂的多，nginx可能是一台服务器，ffmpeg推流的可能就是另一台服务器，这样的话，可将localhost换成对应的IP地址。如果数据源来自摄像头，同样可以通过ffmpeg进行推送，命令如下

ffmpeg -f dshow -i video="Integrated Camera" -vcodec libx264 -preset:v ultrafast -tune:v zerolatency -f flv rtmp://10.10.10.84/myapp/mystream1

Android客户端播放

之前已经写过一个基于vitamio的视频播放器，地址是https://github.com/djstava/DJMediaPlayer，我们就在它的基础上进行修改,找到MainActivity.Java

private String[] files = {"rtmp demo","apple demo"};

在listview的item被点击后发送包含播放地址的intent

Intent intent = new Intent(MainActivity.this, VitamioVideoViewActivity.class);
intent.putExtra("movieUrl", "rtmp://10.10.10.84/myapp/mystream");
startActivity(intent);

－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－

HLS

前言

之前的一篇博文http://www.xugaoxiang.com/2016/08/20/android-rtmp%E7%9B%B4%E6%92%AD/已经简单的介绍了如何利用nginx、nginx-rtmp-module和ffmpeg实现基于rtmp协议的直播.今天这篇继续直播这个话题,聊聊hls的应用.

HLS

HLS(Http Live Streaming)是由Apple公司定义的用于实时流传输的协议，HLS基于HTTP协议实现，传输内容包括两部分，一是M3U8描述文件，二是TS媒体文件。

m3u8文件

#EXTM3U
#EXT-X-VERSION:3
#EXT-X-MEDIA-SEQUENCE:6119
#EXT-X-TARGETDURATION:14
#EXTINF:10.625,
6119.ts
#EXTINF:13.667,
6120.ts
#EXTINF:10.000,
6121.ts

如上,m3u8文件是一个描述文件,必须以#EXTM3U开头,之后是切片TS文件的序列.对于直播来讲,m3u8文件需要进行实时的更新,只保留若干个TS切片序列,防止本地存储撑爆硬盘.

多码率支持

针对应用网络多变及不稳定的情况,多数直播都会提供多码率支持,播放器会根据用户当前的网络状况,自动切换到对应的码率上,大大提升用户体验.在服务器端.为了提供多码率的支持,就需要多级m3u8文件.在主m3u8文件不再有TS序列,而是二级m3u8文件,如下所示

#EXTM3U
#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=1280000
low.m3u8
#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=2560000
mid.m3u8
#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=7680000
hi.m3u8

nginx-rtmp对HLS的支持

nginx-rtmp-module本身对rtmp和hls都有很好的支持,只需要在nginx.conf配置下就ok了


#user  nobody;
worker_processes  auto;rtmp_auto_push on;error_log  logs/error.log;
error_log  logs/error.log  notice;
error_log  logs/error.log  info;#pid        logs/nginx.pid;events {worker_connections  1024;
}rtmp {server {listen 1935;chunk_size 4000;# TV mode: one publisher, many subscribers#application mytv {# enable live streaming#live on;# record first 1K of stream#record all;#record_path /tmp/av;#record_max_size 1K;# append current timestamp to each flv#record_unique on;# publish only from localhost#allow publish 127.0.0.1;#deny publish all;#allow play all;#}# Transcoding (ffmpeg needed)#application big {#    live on;# On every pusblished stream run this command (ffmpeg)# with substitutions: $app/${app}, $name/${name} for application & stream name.## This ffmpeg call receives stream from this application &# reduces the resolution down to 32x32. The stream is the published to# 'small' application (see below) under the same name.## ffmpeg can do anything with the stream like video/audio# transcoding, resizing, altering container/codec params etc## Multiple exec lines can be specified.#    exec ffmpeg -re -i rtmp://localhost:1935/$app/$name -vcodec flv -acodec copy -s 32x32#-f flv rtmp://localhost:1935/small/${name};#}#application small {#    live on;#    # Video with reduced resolution comes here from ffmpeg#}#application webcam {#    live on;# Stream from local webcam#    exec_static ffmpeg -f video4linux2 -i /dev/video0 -c:v libx264 -an#-f flv rtmp://localhost:1935/webcam/mystream;#}#        application mypush {#            live on;# Every stream published here# is automatically pushed to# these two machines#push rtmp1.example.com;#push rtmp2.example.com:1934;
#        }#        application mypull {#            live on;# Pull all streams from remote machine# and play locally#pull rtmp://rtmp3.example.com pageUrl=www.example.com/index.html;
#        }#        application mystaticpull {#            live on;# Static pull is started at nginx start#pull rtmp://rtmp4.example.com pageUrl=www.example.com/index.html name=mystream static;
#        }# video on demand
#        application vod {#            play /opt/www/vod;
#        }#        application vod2 {#            play /var/mp4s;
#        }# Many publishers, many subscribers# no checks, no recording#application videochat {#   live on;# The following notifications receive all# the session variables as well as# particular call arguments in HTTP POST# request# Make HTTP request & use HTTP retcode# to decide whether to allow publishing# from this connection or not#   on_publish http://localhost:8080/publish;# Same with playing#   on_play http://localhost:8080/play;# Publish/play end (repeats on disconnect)#   on_done http://localhost:8080/done;# All above mentioned notifications receive# standard connect() arguments as well as# play/publish ones. If any arguments are sent# with GET-style syntax to play & publish# these are also included.# Example URL:#   rtmp://localhost/myapp/mystream?a=b&c=d# record 10 video keyframes (no audio) every 2 minutes#  record keyframes;#  record_path /tmp/vc;#  record_max_frames 10;#  record_interval 2m;# Async notify about an flv recorded#  on_record_done http://localhost:8080/record_done;#}# HLS# For HLS to work please create a directory in tmpfs (/tmp/hls here)# for the fragments. The directory contents is served via HTTP (see# http{} section in config)## Incoming stream must be in H264/AAC. For iPhones use baseline H264# profile (see ffmpeg example).# This example creates RTMP stream from movie ready for HLS:## ffmpeg -loglevel verbose -re -i movie.avi  -vcodec libx264#    -vprofile baseline -acodec libmp3lame -ar 44100 -ac 1#    -f flv rtmp://localhost:1935/hls/movie## If you need to transcode live stream use 'exec' feature.#application hls {live on;hls on;hls_path /opt/www/live;}# MPEG-DASH is similar to HLS#application dash {#    live on;#    dash on;#    dash_path /tmp/dash;#}}
}# HTTP can be used for accessing RTMP stats
http {server {listen      8081;location / {root /opt/www/;}# This URL provides RTMP statistics in XMLlocation /stat {rtmp_stat all;# Use this stylesheet to view XML as web page# in browserrtmp_stat_stylesheet stat.xsl;}location /stat.xsl {# XML stylesheet to view RTMP stats.# Copy stat.xsl wherever you want# and put the full directory path hereroot /home/djstava/Workshop/Web/nginx-rtmp-module/;}location /hls {# Serve HLS fragmentstypes {application/vnd.apple.mpegurl m3u8;video/mp2t ts;}root /opt/www/;add_header Cache-Control no-cache;}#location /dash {# Serve DASH fragments#    root /tmp;#    add_header Cache-Control no-cache;#}}
}

在rtmp标签下,指定hls application的根路径/opt/www/live,所有的TS切片文件都存放在这里

ffmpeg推流

推送本地文件

ffmpeg -re -i /opt/www/vod/dhxy1.mp4 -vcodec copy -acodec copy -f flv -y rtmp://192.168.1.88/hls/livestream1

推送成功后,你可以通过如下2个url播放对应的模拟实时流,请确保nginx服务已启动.

rtmp://192.168.1.88/hls/livestream1
http://192.168.1.88:8081/live/livestream1.m3u8

另外http://192.168.1.88:8081/stat页面可以显示当前服务的一些信息,如接入的客户端数量,音频视频的信息等等,见下图

推送UDP组播数据

ffmpeg -i udp://@224.0.0.2:9000 -vcodec libx264 -acodec aac -strict -2 -f flv -s 1280x720 -q 10 -ac 1 -ar 44100 rtmp://192.168.1.88/hls/livestream

在以UDP数据为输入源时,ffmpeg会报如下图中的错误信息

这时只需要重新修改下ffmpeg的推流命令就可以,如下

ffmpeg -i 'udp://@224.0.0.2:9000?fifo_size=2000000&overrun_nonfatal=1' -vcodec libx264 -acodec aac -strict -2 -f flv -s 1280x720 -q 10 -ac 1 -ar 44100 rtmp://192.168.1.88/hls/livestream

fifo_size的单位是字节,自己酌情增减.

－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－

多码率支持：

前言

Android RTMP直播(续)介绍了HLS协议相关的基础内容,本文将继续深入学习HLS的其它高级特性.

服务端多码率支持

nginx.conf


#user  nobody;
worker_processes  auto;rtmp_auto_push on;error_log  logs/error.log;
error_log  logs/error.log  notice;
error_log  logs/error.log  info;#pid        logs/nginx.pid;events {worker_connections  1024;
}rtmp {server {listen 1935;chunk_size 4000;# TV mode: one publisher, many subscribers#application mytv {# enable live streaming#live on;# record first 1K of stream#record all;#record_path /tmp/av;#record_max_size 1K;# append current timestamp to each flv#record_unique on;# publish only from localhost#allow publish 127.0.0.1;#deny publish all;#allow play all;#}# Transcoding (ffmpeg needed)#application big {#    live on;# On every pusblished stream run this command (ffmpeg)# with substitutions: $app/${app}, $name/${name} for application & stream name.## This ffmpeg call receives stream from this application &# reduces the resolution down to 32x32. The stream is the published to# 'small' application (see below) under the same name.## ffmpeg can do anything with the stream like video/audio# transcoding, resizing, altering container/codec params etc## Multiple exec lines can be specified.#    exec ffmpeg -re -i rtmp://localhost:1935/$app/$name -vcodec flv -acodec copy -s 32x32#-f flv rtmp://localhost:1935/small/${name};#}#application small {#    live on;#    # Video with reduced resolution comes here from ffmpeg#}#application webcam {#    live on;# Stream from local webcam#    exec_static ffmpeg -f video4linux2 -i /dev/video0 -c:v libx264 -an#-f flv rtmp://localhost:1935/webcam/mystream;#}#        application mypush {#            live on;# Every stream published here# is automatically pushed to# these two machines#push rtmp1.example.com;#push rtmp2.example.com:1934;
#        }#        application mypull {#            live on;# Pull all streams from remote machine# and play locally#pull rtmp://rtmp3.example.com pageUrl=www.example.com/index.html;
#        }#        application mystaticpull {#            live on;# Static pull is started at nginx start#pull rtmp://rtmp4.example.com pageUrl=www.example.com/index.html name=mystream static;
#        }# video on demand
#        application vod {#            play /opt/www/vod;
#        }#        application vod2 {#            play /var/mp4s;
#        }# Many publishers, many subscribers# no checks, no recording#application videochat {#   live on;# The following notifications receive all# the session variables as well as# particular call arguments in HTTP POST# request# Make HTTP request & use HTTP retcode# to decide whether to allow publishing# from this connection or not#   on_publish http://localhost:8080/publish;# Same with playing#   on_play http://localhost:8080/play;# Publish/play end (repeats on disconnect)#   on_done http://localhost:8080/done;# All above mentioned notifications receive# standard connect() arguments as well as# play/publish ones. If any arguments are sent# with GET-style syntax to play & publish# these are also included.# Example URL:#   rtmp://localhost/myapp/mystream?a=b&c=d# record 10 video keyframes (no audio) every 2 minutes#  record keyframes;#  record_path /tmp/vc;#  record_max_frames 10;#  record_interval 2m;# Async notify about an flv recorded#  on_record_done http://localhost:8080/record_done;#}# HLS# For HLS to work please create a directory in tmpfs (/tmp/hls here)# for the fragments. The directory contents is served via HTTP (see# http{} section in config)## Incoming stream must be in H264/AAC. For iPhones use baseline H264# profile (see ffmpeg example).# This example creates RTMP stream from movie ready for HLS:## ffmpeg -loglevel verbose -re -i movie.avi  -vcodec libx264#    -vprofile baseline -acodec libmp3lame -ar 44100 -ac 1#    -f flv rtmp://localhost:1935/hls/movie## If you need to transcode live stream use 'exec' feature.#application hls {live on;hls on;hls_path /opt/www/live;hls_nested on;hls_variant _low BANDWIDTH=800000;hls_variant _mid BANDWIDTH=1200000;hls_variant _hi  BANDWIDTH=2000000;}# MPEG-DASH is similar to HLS#application dash {#    live on;#    dash on;#    dash_path /tmp/dash;#}}
}# HTTP can be used for accessing RTMP stats
http {server {listen      8081;location / {root /opt/www/;}# This URL provides RTMP statistics in XMLlocation /stat {rtmp_stat all;# Use this stylesheet to view XML as web page# in browserrtmp_stat_stylesheet stat.xsl;}location /stat.xsl {# XML stylesheet to view RTMP stats.# Copy stat.xsl wherever you want# and put the full directory path hereroot /home/djstava/Workshop/Web/nginx-rtmp-module/;}location /hls {# Serve HLS fragmentstypes {application/vnd.apple.mpegurl m3u8;video/mp2t ts;}root /opt/www/;add_header Cache-Control no-cache;}#location /dash {# Serve DASH fragments#    root /tmp;#    add_header Cache-Control no-cache;#}}
}

主要看看application hls的内容

application hls {live on;hls on;hls_path /opt/www/live;hls_nested on;hls_variant _low BANDWIDTH=800000;hls_variant _mid BANDWIDTH=1200000;hls_variant _hi  BANDWIDTH=2000000;}

这里设定当带宽分别为800k 1200k 2000k的时候,终端都播放相对应的m3u8索引文件

ffmpeg推流

这里需要利用ffmpeg推送3路不同的流,对应上面提到的低中高

ffmpeg -re -i ~/Videos/xjcy.mp4 -vcodec copy -acodec copy -b:v 800k -b:a 32k -f flv rtmp://10.10.10.59/hls/livestream_low

ffmpeg -re -i ~/Videos/xjcy.mp4 -vcodec copy -acodec copy -b:v 1200k -b:a 64k -f flv rtmp://10.10.10.59/hls/livestream_mid

ffmpeg -re -i ~/Videos/xjcy.mp4 -vcodec copy -acodec copy -b:v 2000k -b:a 128k -f flv rtmp://10.10.10.59/hls/livestream_hi

推送开始后,hls的root目录下就会生成相应的文件内容,如下图所示

此时livestream.m3u8文件内容为

#EXTM3U
#EXT-X-VERSION:3
#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=800000
livestream_low/index.m3u8
#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=1200000
livestream_mid/index.m3u8
#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=2000000
livestream_hi/index.m3u8

相应的,播放链接为 http://10.10.10.59:8081/live/livestream.m3u8,播放器需要做的就是根据自身的网络状况,切换到其它的索引文件.

直播节目的录制

直播进行的同时一般都会有本地录制的需求,nginx-rtmp-module提供了这个功能,接下来实践一下.还是看nginx.conf配置文件

application hls {live on;hls on;hls_path /opt/www/live;hls_nested on;hls_variant _low BANDWIDTH=800000;hls_variant _mid BANDWIDTH=1200000;hls_variant _hi  BANDWIDTH=2000000;recorder all {record all;record_suffix -%Y-%m-%d-%H_%M_%S.flv;record_max_size 200000K;record_path /opt/www/record;}}

record all录制所有内容,也可以只录音频或者视频.

推流后/opt/www/record路径下就会自动生成带对应时间戳的flv文件,用vlc测试播放OK.

时移电视

要想实现时移电视(这里指的是服务器端)的话,首先需要在服务器上保留足够的切片文件,比如说你提供1小时的时移,就意味着要有1小时的切片文件,而且索引文件中包含前1小时的切片序列.

application hls {live on;hls on;hls_path /opt/www/live;hls_continuous on;hls_sync 100ms;hls_nested on;hls_playlist_length 5m;hls_fragment 10s;hls_variant _low BANDWIDTH=800000;hls_variant _mid BANDWIDTH=1200000;hls_variant _hi  BANDWIDTH=2000000;#exec /home/djstava/Workshop/Web/nginx-1.11.3/build/test.sh;#exec_kill_signal term;#recorder all {#    record all;#    record_suffix -%Y-%m-%d-%H_%M_%S.flv;#    record_max_size 6200000K;#    record_path /opt/www/record;#}}

hls_fragment指的是切片文件的长度,这里是10秒,hls_playlist_length指的是索引文件的长度,我这里设的是5分钟.推流开始后,你到切片生成的目录,会发现*.m3u8文件包含了30个ts序列.所以,在上面这种情况下,就只能进行5分钟的时移,当播放进度到达当前直播点时则继续回到直播状态.

执行外部shell脚本

比如有个脚本test.sh,内容如下

#!/bin/bash

on_die ()
{# kill all childrenpkill -KILL -P $$
}trap 'on_die' TERM
ffmpeg -re -i /home/djstava/Videos/ygdx.mp4  -vcodec copy -acodec copy -f flv rtmp://10.10.10.48/hls/ygdx &
wait

我这里把它放在hls application中执行,则nginx.conf应如下

application hls {live on;hls on;hls_path /opt/www/live;hls_continuous on;hls_sync 100ms;hls_nested on;hls_playlist_length 5m;hls_fragment 10s;hls_variant _low BANDWIDTH=800000;hls_variant _mid BANDWIDTH=1200000;hls_variant _hi  BANDWIDTH=2000000;exec /home/djstava/Workshop/Web/nginx-1.11.3/build/test.sh;exec_kill_signal term;#recorder all {#    record all;#    record_suffix -%Y-%m-%d-%H_%M_%S.flv;#    record_max_size 6200000K;#    record_path /opt/www/record;#}}

当hls服务正常启动时(如上面写过的ffmpeg推流动作),外部脚本test.sh也被执行了.脚本中捕捉了退出的中断信号,也就说,如果ffmpeg推流动作中断了,那么test.sh脚本也就不再执行了.

制作RAMDISK

为了提高HLS的读写效率,可以把切片和索引文件操作放在内存中进行.

mount -t tmpfs -o size=512m tmpfs /opt/www/live

转自：http://blog.csdn.net/djstavav/article/details/53021361