Skip to main content

Package media content for online streaming(DASH and HLS) using ffmpeg

Project description

📼 Python FFmpeg Video Streaming

Build Status Build status Scrutinizer Code Quality Downloads PyPI version Software License

Overview

This package uses the FFmpeg to package media content for online streaming such as DASH and HLS. You can also use DRM for HLS packaging. There are several options to open a file from clouds and save files to them as well.

Contents

Requirements

  1. This version of the package is only compatible with Python 3.6 or higher.

  2. To use this package, you need to install the FFmpeg. You will need both FFMpeg and FFProbe binaries to use it.

Installation

The latest version of ffmpeg-streaming can be acquired via pip:

pip install python-ffmpeg-video-streaming

Quickstart

Opening a Resource

There are two ways to open a file:

1. From a FFmpeg supported resource

You can pass a local path of video(or a supported resource) to the method(hls or dash):

video = '/var/www/media/videos/video.mp4'

For opening a file from a supported resource such as http, ftp, pipe, rtmp and etc. please see FFmpeg Protocols Documentation

For example:

video = 'https://www.aminyazdanpanah.com/PATH/TO/VIDEO.MP4'

2. From Clouds

You can open a file from a cloud by passing a tuple of cloud configuration to the method.

In this page, you will find some examples of opening a file from Amazon S3, Google Cloud Storage, Microsoft Azure Storage, and a custom cloud.

video = (google_cloud, download_options, None)

DASH

Dynamic Adaptive Streaming over HTTP (DASH), also known as MPEG-DASH, is an adaptive bitrate streaming technique that enables high quality streaming of media content over the Internet delivered from conventional HTTP web servers. Learn more

Create DASH files:

import ffmpeg_streaming

(
    ffmpeg_streaming
        .dash(video, adaption='"id=0,streams=v id=1,streams=a"')
        .format('libx265')
        .auto_rep()
        .package('/var/www/media/videos/dash/dash-stream.mpd')
)

Generate representations manually:

import ffmpeg_streaming
from ffmpeg_streaming import Representation

r_144p  = Representation(width=256, height=144, kilo_bitrate=95)
r_240p  = Representation(width=426, height=240, kilo_bitrate=150)
r_360p  = Representation(width=640, height=360, kilo_bitrate=276)
r_480p  = Representation(width=854, height=480, kilo_bitrate=750)
r_720p  = Representation(width=1280, height=720, kilo_bitrate=2048)
r_1080p = Representation(width=1920, height=1080, kilo_bitrate=4096)
r_2k    = Representation(width=2560, height=1440, kilo_bitrate=6144)
r_4k    = Representation(width=3840, height=2160, kilo_bitrate=17408)

(
    ffmpeg_streaming
        .dash(video, adaption='"id=0,streams=v id=1,streams=a"')
        .format('libx265')
        .add_rep(r_144p, r_240p, r_360p, r_480p, r_720p, r_1080p, r_2k, r_4k)
        .package('/var/www/media/videos/dash/dash-stream.mpd')
)

See DASH examples and DASH options for more information.

HLS

HTTP Live Streaming (also known as HLS) is an HTTP-based adaptive bitrate streaming communications protocol implemented by Apple Inc. as part of its QuickTime, Safari, OS X, and iOS software. Client implementations are also available in Microsoft Edge, Firefox and some versions of Google Chrome. Support is widespread in streaming media servers. Learn more

Create HLS files:

import ffmpeg_streaming

(
    ffmpeg_streaming
        .hls(video, hls_time=10, hls_allow_cache=1)
        .format('libx264')
        .auto_rep()
        .package('/var/www/media/videos/hls/hls-stream.m3u8')
)

Generate representations manually:

import ffmpeg_streaming
from ffmpeg_streaming import Representation

r_360p = Representation(width=640, height=360, kilo_bitrate=276)
r_480p = Representation(width=854, height=480, kilo_bitrate=750)
r_720p = Representation(width=1280, height=720, kilo_bitrate=2048)

(
    ffmpeg_streaming
        .hls(video, hls_time=10, hls_allow_cache=1)
        .format('libx264')
        .add_rep(r_360p, r_480p, r_720p)
        .package('/var/www/media/videos/hls/hls-stream.m3u8')
)

NOTE: You cannot use HEVC(libx265) and VP9 formats for HLS packaging.

DRM (Encrypted HLS)

The encryption process requires some kind of secret (key) together with an encryption algorithm. HLS uses AES in cipher block chaining (CBC) mode. This means each block is encrypted using the ciphertext of the preceding block. Learn more

You must specify a path to save a random key to your local machine and also a URL(or a path) to access the key on your website(the key you will save must be accessible from your website). You must pass both these parameters to the encryption method:

Single Key

The following code generates a key for all TS files.

import ffmpeg_streaming

#A path you want to save a random key to your server
save_to = '/home/public_html/PATH_TO_KEY_DIRECTORY/random_key.key'

#A URL (or a path) to access the key on your website
url = 'https://www.aminyazdanpanah.com/PATH_TO_KEY_DIRECTORY/random_key.key'
# or url = '/PATH_TO_KEY_DIRECTORY/random_key.key'

(
    ffmpeg_streaming
        .hls(video, hls_time=10, hls_allow_cache=1)
        .encryption(url, save_to)
        .format('libx264')
        .auto_rep(heights=[480, 360, 240])
        .package('/var/www/media/videos/hls/hls-stream.m3u8')
)
Key Rotation

The code below, allows you to encrypt each TS file with a new encryption key. This can improve security and allows for more flexibility. You can also modify the code to use a different key for each set of segments(i.e. if 10 TS files has been generated then rotate the key) or you can generate a new encryption key at every periodic time(i.e. every 10 seconds).

import tempfile
from os.path import join
from random import randrange

import ffmpeg_streaming
from ffmpeg_streaming.key_info_file import generate_key_info_file

save_to = '/home/public_html/PATH_TO_KEY_DIRECTORY/key_rotation'
url = 'https://www.aminyazdanpanah.com/PATH_TO_KEY_DIRECTORY/key_rotation'
key_info_file_path = join(tempfile.gettempdir(), str(randrange(1000, 100000)) + '_py_ff_vi_st.tmp')
k_num = 1


def k_format(name, num):
    return str(name) + "_" + str(num)


def progress(per, ffmpeg):
    global k_num
    if ".ts' for writing" in ffmpeg:
        # A new TS file has been created!
        generate_key_info_file(k_format(url, k_num), k_format(save_to, k_num), key_info_path=key_info_file_path)
        k_num += 1

(
    ffmpeg_streaming
        .hls(video, hls_flags="periodic_rekey")
        .encryption(url, save_to, key_info_path=key_info_file_path)
        .format('libx264')
        .auto_rep()
        .package('/var/www/media/videos/hls/hls-stream.m3u8', progress=progress)
)

NOTE: It is very important to protect your key(s) on your website using a token or a session/cookie(It is highly recommended).

NOTE: However HLS supports AES encryption, that you can encrypt your streams, it is not a full DRM solution. If you want to use a full DRM solution, I recommend to try FairPlay Streaming solution which then securely exchange keys, and protect playback on devices.

See HLS examples and HLS options for more information.

Progress

You can get realtime information about transcoding by passing a callable method to the package method:

import sys
import ffmpeg_streaming


def progress(percentage, ffmpeg):
    # You can update a field in your database
    # You can also create a socket connection and show a progress bar to users
    sys.stdout.write("\rTranscoding...(%s%%)[%s%s]" % (percentage, '#' * percentage, '-' * (100 - percentage)))
    sys.stdout.flush()


(
    ffmpeg_streaming
        .hls(video)
        .format('libx264')
        .auto_rep()
        .package('/var/www/media/videos/hls/hls-stream.m3u8', progress=progress)
)
Output From a Terminal:

transcoding

Show a progress bar using tqdm

You can get realtime information about transcoding by passing a callable method to the package method:

import ffmpeg_streaming
from tqdm import tqdm

# initialize the tqdm object
bar = tqdm(total=100)
last_per = 0

def progress(percentage, ffmpeg):
    # update the progress bar
    global last_per
    if last_per != percentage:
        bar.update(percentage - last_per)
        last_per = percentage


(
    ffmpeg_streaming
        .hls(video)
        .format('libx264')
        .auto_rep()
        .package('/var/www/media/videos/hls/hls-stream.m3u8', progress=progress)
)

# close the progress bar
bar.close()

Saving Files

There are two ways to save your files.

1. To a Local Path

You can pass a local path to the package method. If there was no directory in the path, then the package auto makes the directory.

(
    ffmpeg_streaming
        .hls(video)
        .format('libx264')
        .auto_rep()
        .package('/var/www/media/videos/hls/hls-stream.m3u8', progress=progress)
)

It can also be null. The default path to save files is the input path.

(
    ffmpeg_streaming
        .hls(video)
        .format('libx264')
        .auto_rep()
        .package(progress=progress)
)

NOTE: If you open a file from a cloud and do not pass a path to save the file to your local machine, you will have to pass a local path to the package method.

2. To Clouds

You can save your files to a cloud by passing a tuple of cloud configuration to the package method.

In this page, you will find some examples of saving files to Amazon S3, Google Cloud Storage, Microsoft Azure Storage, and a custom cloud.

(
    ffmpeg_streaming
        .dash('/var/www/media/video.mkv', adaption='"id=0,streams=v id=1,streams=a"')
        .format('libx265')
        .auto_rep()
        .package(clouds=[to_aws_cloud, to_azure_cloud, to_google_cloud],
                 progress=progress)
)

A path can also be passed to save a copy of files to your local machine.

(
    ffmpeg_streaming
        .dash('/var/www/media/video.mkv', adaption='"id=0,streams=v id=1,streams=a"')
        .format('libx265')
        .auto_rep()
        .package(output='/var/www/media/stream.mpd', clouds=[to_aws_cloud, to_google_cloud],
                 progress=progress)
)

NOTE: This option(Save To Clouds) is only valid for VOD (it does not support live streaming).

Schema: The relation is one-to-many.

Probe

You can extract the metadata of the video file using the following code:

from ffmpeg_streaming import FFProbe

ffprobe = FFProbe('/var/www/media/video.mp4')

NOTE: You can save these metadata to your database.

See the example for more information.

Live

You can pass a url(or a supported resource like ftp) to live method to upload all the segments files to the HTTP server(or other protocols) using the HTTP PUT method, and update the manifest files every refresh times.

If you want to save stream files to your local machine, please use the package method.

# DASH live
import ffmpeg_streaming

(
    ffmpeg_streaming
        .dash('/var/www/media/video.avi', adaption='"id=0,streams=v id=1,streams=a"')
        .format('libx265')
        .auto_rep()
        .live('http://YOUR-WEBSITE.COM/live-stream/out.mpd', progress=progress)
)

# HLS live
(
    ffmpeg_streaming
        .hls('/var/www/media/video.mp4', master_playlist_path='/var/www/stream/live-master-manifest.m3u8')
        .format('libx264')
        .auto_rep()
        .live('http://YOUR-WEBSITE.COM/live-stream/out.m3u8', progress=progress)
)

NOTE: In the HLS method, you must upload the master manifest to the server manually. (Upload the /var/www/stream/live-master-manifest.m3u8 file to the http://YOUR-WEBSITE.COM)

Please see FFmpeg Protocols Documentation for more information.

Conversion

You can convert your stream to a file or to another stream protocols. You should pass a manifest of a stream to the method:

1. HLS To DASH

import ffmpeg_streaming
from ffmpeg_streaming import Representation

r_360p = Representation(width=640, height=360, kilo_bitrate=276)

(
    ffmpeg_streaming
        .dash('https://www.aminyazdanpanah.com/PATH/TO/HLS-MANIFEST.M3U8')
        .format('libx265')
        .add_rep(r_360p)
        .package('/var/www/media/dash-stream.mpd', progress=progress)
)

2. DASH To HLS

import ffmpeg_streaming

(
    ffmpeg_streaming
        .hls('https://www.aminyazdanpanah.com/PATH/TO/DASH-MANIFEST.MPD')
        .format('libx264')
        .auto_rep(heights=[360, 240])
        .package('/var/www/media/hls-stream.m3u8', progress=progress)
)

3. Stream(DASH or HLS) To File

import ffmpeg_streaming

(
    ffmpeg_streaming
        .stream2file('https://www.aminyazdanpanah.com/PATH/TO/MANIFEST.MPD or M3U8')
        .format('libx264')
        .save('/var/www/media/new-video.mp4', progress=progress)
)

Several Open Source Players

You can use these libraries to play your streams.

As you may know, IOS does not have native support for DASH. Although there are some libraries such as Viblast and MPEGDASH-iOS-Player to support this technique, I have never tested them. So if you know any IOS player that supports DASH Stream and also works fine, please add it to the above list.

NOTE: You should pass a manifest of stream(e.g. https://www.aminyazdanpanah.com/PATH_TO_STREAM_DIRECTORY/dash-stream.mpd or /PATH_TO_STREAM_DIRECTORY/hls-stream.m3u8 ) to these players.

Contributing and Reporting Bugs

I'd love your help in improving, correcting, adding to the specification. Please file an issue or submit a pull request.

  • Please see Contributing File for more information.
  • If you have any questions or you want to report a bug, please just file an issue
  • If you discover a security vulnerability within this package, please see SECURITY File for more information.

NOTE: If you have any questions about this package or FFmpeg, please DO NOT send an email to me (or submit the contact form on my website). Emails regarding these issues will be ignored.

Credits

License

The MIT License (MIT). Please see License File for more information.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

python-ffmpeg-video-streaming-0.0.25.tar.gz (21.6 kB view hashes)

Uploaded Source

Built Distribution

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page