September 23, 2021

Conditional transcoding for ABR

Nimble Streamer Live Transcoder allows building transcoding pipelines of various complexity however it used to provide only straight-forward scenarios regardless of incoming stream parameters.

With conditional transcoding you can implement more complex scenarios based on the resolution of incoming stream. You may define the following conditions for decoder element:

  • width of the original stream;
  • height of the original stream;
  • comparison operators <, <=, =, >= or >;
If the defined condition is not met, the decoding will not start and its respective pipeline will not start.

The most popular use cases of such conditioning are:
  • situational up-scaling: if you have a scenario to transcode 4K input to 1080p and your input stream from your source happen to go down to 720p, this will prevent wasting your resources to create unnecessary high resolution.
  • set maximum output resolution, so whatever is published by your streamers, you produce a ladder limited from above.
All use cases fit perfectly with live ABR feature set of Nimble Streamer. It allows defining ABR streams with any number of streams and the output will provide any streams that are available at any given moment. So if some stream is not generated due to conditional transcoder, the output adaptive stream will still be available for playback with lower resolutions.

Here's how you can set up conditional transcoding.

We assume you're already familiar with Live Transcoder, you already have an active Transcoder license and you have it installed and registered

We'll create a scenario which allows down-scaling any incoming stream to 720p if the input video has larger frame size, then do the same for 480p resolution and create 360p regardless of the input. The audio will be passed through for all inputs.

Create a new scenario and drag and drop a new Video source element there.
Then Click on Filter settings to see fields with conditions:


As you can see from the screenshot, this decoder will be initiated if the height of original stream is greater. Then we put Scale filter and Video output element to downscale video to 720p into respective pipeline as shown below.



As a result, you'll get 720p only if the source stream is greater than 720p. If the original stream is 720p then you'll be able to use it in our ABR output as is, we'll show it below.

Now let's add 480p transcoding pipeline.



This Video source has condition of height>480. You'll get 480p only if the source stream is greater than 480p. This will cover extreme cases when the source drops the resolution for some reason.

The remaining 360 pipeline can be implemented without any conditions as we assume the resolution should not drop below 480p in vast majority of cases. Full transcoding scenario will look like this.




Now, having these conditional pipelines, you can proceed with creating output ABR stream.
  1. Go to Nimble Streamer top menu and click on Live Streams.
  2. On the next streams page you need to find your current transcoding server and click on the number in ABR column.
  3. You'll be brought to ABR streams page, there you need to click on Add ABR setting.
  4. You'll see a dialog which you need to fill as shown below.


Here you see an original stream as well as all 4 down-scaled sub-streams. If any of the sub-streams becomes un-available available, Nimble will still generate the correct playlist with active streams so your player will be able to choose between available resolutions.

You can find more details about setup and ABR streams' usage in this ABR setup article.


Overall, conditional transcoding can serve many purposes so feel free to experiment with it as much as you need. Let us know of your use cases which utilize this feature.

Related documentation


September 9, 2021

Nimble Streamer is Readers' Choice Awards 2021 finalist

Softvelum team is excited to be among the finalists of 2021 Streaming Media European Readers' Choice Awards in Hardware/software Server category with our Nimble Streamer media server.

We've been the finalist in 2016 Reader's Choice Awards as Best Innovation, and now we're here again as the best server, that's very exciting.

Huge thanks to all of our customers and partners who voted for us!






August 30, 2021

Generate NDI stream from local files via Server Playlist

We want to tell you about Server Playlist feature, which lets you create live streams from the video or audio files on your storage. This feature gives vigorous possibilities to your live streaming setup: from simply running a video in a loop as a backup stream to arranging a precise schedule for programs to run at a specific time. Multiplied by Nimble’s and Live Transcoder processing power, you can have any desired live stream egress ranging from re-publishing for any major streaming provider via RTMP or have a precise NDI stream to your equipment. Any supported output protocol can be used, including cutting edge SLDP, SRT and RIST.

As you finish reading this tutorial, you will be able to stream files from a hard disk drive as NDI streams and learn the basics of playlist editing for Nimble Streamer.

For our use case, we will just describe a couple of files in a playlist and loop them endlessly to a live stream, next we set up the Transcoder to make NDI output.


Step 1: Check Prerequisites

First, make sure you have the latest version of Nimble Streamer. You can install it using this instruction or upgrade it.

Second, make also sure that the Transcoder package is installed as described on Transcoder installation page.

Third, you need to have an active Live Transcoder license with respective subscription. Check the license at your account Settings menu under Transcoder licenses tab.

Also, we assume that all your video files are encoded with H.264 and have the same resolution and stereo AAC audio stream, so no additional encoding will be discussed.


Step 2: Make Nimble to know your playlists

Conception starts with a Playlist that uses a JSON syntax to describe files for making a live stream. JSON is powerful yet clear and various online tools can be used to validate it against errors.

To make Nimble know the playlist is engaged in work, specify a location of your playlist by adding the server_playlist_sync_url option to nimble.conf. Location can be either a local path on a hard disk drive with Nimble or HTTP(s) URL. 

Add the next string to /etc/nimble/nimble.conf file:

server_playlist_sync_url=/var/nimble/playlists/basic_playlist.json

and restart Nimble to apply changes. We will fill basic_playlist.json a moment later.

The playlist will be reloaded every 10 seconds, so you can make dynamic changes via some custom scripting and these changes will be accepted by Nimble. If you want any other update time interval, you may set it via the server_playlist_sync_interval option in nimble.conf or via the playlist itself.


Step 3: Write a playlist and get your live stream


The playlist’s grammar is based on key sections: Tasks, Blocks and Streams - which can have their own properties, e.g. Start time or Duration.

Now, copy this text to basic_playlist.json file using your preferred console editor:

{
  "Tasks": [
    {
      "Stream": "live/from_playlist",
      "Blocks": [
        {
          "Id":"1",
   "MaxIterations":0,
           "Streams" : [
            {"Type": "vod", "Source": "/var/nimble/mp4/bbb_1080p_h264_aac.mp4"},
            {"Type": "vod", "Source": "/var/nimble/mp4/TOS-1080p_aac.mp4"}
          ]
        }
      ]
    }
  ]
}

A few seconds later you will notice stream live/from_playlist appearing at Live Streams / Outgoing Live streams of your server.



This very basic example perfectly illustrates the mandatory fields and values of a playlist. It contains one Tasks, one Block, and two Stream sections.

Tasks object represents an array of alive streams, which must have a name set by a Stream field as application/name. Tasks also hold an array of Blocks that can be started at different times. It is achieved via Start property, but we are not discussing it for now in this tutorial. Each block must have a unique Id and a number of times to run. These repetitions are set by MaxIterations and each block runs only once by default. Setting MaxIterations to 0 means the infinite loop of a block.

Next comes the Streams array of objects which holds the full path and name of a file to play within the previously defined live stream. Source property is used for this. At the moment, Type property must have “vod” value, and this property is reserved for future improvements for mixing live streams with video files.

No more additional properties are needed for our simplified case but more details can be found at feature technical spec if needed.

Step 4: Turn Live Stream into NDI stream

Now, you see a live/from_playlist stream, so it can be processed via Transcoder to have NDI output.
Create a new Transcoder scenario and use live/from_playlist as Audio and Video Input.


As for Video and Audio output, click NDI radio button and name it from_playlist:




Click Save button and switch to a console to check available NDI streams with the nimble_ndi tool or any other NDI-supported software.





You did a great job by easily transforming your video files to an NDI stream using the Nimble Server Playlist feature and Live Transcoder.

We will continue to show more elegant scenarios and more complex playlists’ grammar in our later articles. Please feel free to share this article with anyone who may be interested in Nimble Streamer’s ability to playback files as a live stream.

Related documentation


August 8, 2021

Best practices for Larix Broadcaster production deployment

Larix Broadcaster is a universal mobile tool which is used in a variety of cases from a single event to complex multi-device production. Large-scale deployments show that you may successfully use modern mobile devices for live content remote contribution.

For example, as per Sports Video Group News, Larix was used in 2020 virtual drafts of NFL and MLB. They've released two articles regarding NFL Draft, with Larix Broadcaster mentioned as part of streaming setup: "NFL Draft 2020 ..." and "NFL Draft Reflections ...". The MLB Draft is covered in their respective MLB Draft 2020 article.
In this article we'll overview the best practices and recommendations for Larix Broadcaster deployments and mobile live streaming in general, based on our customers' experience.


Testing is crucial


First and most important advice we can give is: test as much as you can prior to your live event. This relates to hardware capabilities, software features, network and environment features.

This is especially important before making decisions about buying a significant amount of hardware, e.g. when you purchase a fleet of devices for your crew. In general it's good to have identical devices for all members of the team, so we highly recommend doing testing before such purchase.

So start with getting your hands on at least one device and try running your streaming scenarios on it. 

First, check the straight-forward scenario with some default settings (720p@30fps, H.264, default bitrate etc.) then move to those settings which you'd like to use in production if they differ. Short tests need to be combined with long runs.

Then try to push the limit. If your event is planned for 1 hour then run a 2-hours test, just to try the robustness of your device and your streaming setup. If your designated device has any energy-related or temperature problems, you'll see them at this point.

As part of our testing routine before major public releases, we run an hour-long test with periodic changing scenes, doing rotation, camera flip etc to imitate intensive work. Then we also do a work-day-long test with more static-scene streaming to make sure Larix won't cause troubles or overheat on the long run. We call all that the "endurance test".

Android vs iPhone


iPhone is the best option for live streaming so far. We run our tests on a range of devices from iPhone 6s to iPhone 11 Pro, and they all provide Larix with 1080p@60fps encoding. Starting from iPhone XR, latest models even provide 4K@60fps. Of course, you can surely use HEVC for all these modes. The image quality is also excellent, as you would expect.

What about Android? Most of the manufacturers provide only standard to minimum capabilities for third-party apps. We can confirm only one device which supports all the features of Android platform, that is Google Pixel 5. You can be sure Larix will pick up 1080p@60fps, concurrent cameras, 4K@30fps and all the latest platform features.

Android platform encourages the existence of what we call a "zoo of devices". This means a wide variance of shapes, sizes and features, however when it comes to encoding, a vendor will probably follow basic requirements like 1080p@30fps but most other image-related features won't be available for third-party apps. So you won't see 60fps, concurrent cameras or sometimes even 4K in Larix Broadcaster. Unlike iPhone, most Android devices do not provide exact frame rate of 25fps.

Read more about FPS support on Android and iOS in Q12 of our FAQ. It's a bit tricky on mobile devices in general.

If you prefer using Android and you don't need fancy image features, then most mid- to high-tier Android devices should work fine for you. But you need to run proper testing as described above. We use several devices by Samsung, Xiaomi and other brands and they work well in our tests. So far Pixel 5 is the best among them.

One more notice about Android: Larix Broadcaster can be used in background mode which will allow turning off the screen, thus saving energy and making the device less warm.

If you prefer an iPhone then basically any affordable model will do the job.

Audio, additional gear and tricks


Many of our customers use external audio equipment to obtain better sound for their live streams. We like good sound so besides "standard" features like sample rate or channel count, we allow selecting audio source as well as stream in "audio-only" mode when Larix streams only audio part of content. We also plan to extend this with gain control later on.

We also see our customers using external microphones and external sound cards successfully, for example we test Larix with iRig HD2 and it works great. Some sound cards require different audio sample rates, i.e. strictly 48000Hz or 44100Hz. So if you experience issues please check that setting, and also switch audio sources as some devices are capricious about that.

Besides audio, additional streaming gear may include lightning kits and tripods, these are things that you need to pick up according to your case.

Take a look at The Ultimate Guide To DJ Livestreaming by Phil Morse who gives a good overview of audio hardware, other gear and best practices.

Notice that Larix Broadcaster provides image framing features like rule-of-thirds and safe margins. We recommend you to watch Video Camera and Webcam Framing for Live and VOD tutorial by Jan Ozer to get some creative ideas.

And of course don't forget about the power source. Your phone battery may be excellent but high-quality streaming will eat it very quickly. So either get some good power bank or just plug your phone into a charger. Notice that you'd like to reduce power consumption and have less heating. you can disable live rotation and image overlays, this simplifies the post-processing.

Streaming protocols


Today we have two general options when choosing a live streaming protocol: SRT and RTMP.

If you use popular public services like YouTube Live, Facebook Live or Twitch, you have no choice but to use RTMP. It's an old and well-implemented technology available in any service and live decoder. It does the job properly for a large number of use cases, especially when you stream in a controlled reliable network environment. Larix Broadcaster has solid RTMP support and it's tested with all major RTMP-powered solutions.

If you build your own streaming network or if you can choose a provider with SRT input, you should definitely consider SRT. This is especially important if your crew members stream in mobile networks. SRT was designed for operating in unreliable networks so even if your network has significant packet loss, this will be properly handled by re-sending the lost packets. This advantage comes with a downside: you need to set a latency parameter which defines a time buffer used for compensating the glitches, so keep in mind that you'll have that delay in your production input. We recommend using latency parameters as 4*RTT with your streaming destination, please read this article for more details.

Larix Broadcaster has full support for SRT publishing in Caller, Pull and Rendezvous modes and SRT is being properly tested with all major SRT solutions and tools. Most modern hardware decoders can process SRT (e.g. Haivision is the inventor of SRT and their devices have full support). There are also a lot of software solutions like our Nimble Streamer or Wowza Streaming Engine capable of SRT streaming, so if you decide to try something different from default live streaming services, you'll have plenty of options.

Larix also supports RTSP and RIST output even though those protocols are not used as widely as RTMP or SRT.

So work with your server-side team or your streaming service provider to see which option you may use.

Talkback


Remote contribution often means that a talent needs to be in contact with the studio. Larix Broadcaster supports Talkback, that is IFB audio return feed which you may get on your mobile device using several protocols. You can use any software or hardware solution to generate a live stream with voice instructions from the studio and use earpiece on a mobile device to hear it.

Take a look at our video setup instruction and watch mobile reporter use case demonstration with SRTMiniServer which supports talkback generation.

Larix Grove


If you operate a fleet of devices for live content contribution you need a way to provide your crew with connections' details. Every contributor will have separate connection URL and you'll need to explain each of them how to make the setup.

You may use Larix Grove format to create specially formed links and their QR codes to propagate your settings. On iPhone the QR code can be captured by the system camera app. On Android you may use any QR code reader for that. Once the code is recognized, the settings are imported into Larix Broadcaster and the stream can be started right after that.

The format is simple so you can either use our online wizard or make your own URL generator for that.

Test any further changes


So you've run your tests prior to your live event and Larix Broadcaster works perfectly, like you would expect. 

Now, freeze the settings and don't change anything until the event is over. No software updates, no settings change - just run the event with previously tested devices.

If you really need to make changes to either your mobile app or to server-side part between further events, then start your testing from the beginning to make sure that nothing is broken.

You can never over-test, the more you test the better.



We hope these recommendations are useful for you. If you have your own experience which might be useful for other streamers out there, please don't hesitate to send us a message about it. We appreciate all feedback from our customers.

P.S. If you speak Russian, check this article's translation here.

July 28, 2021

Streaming Media European Readers' Choice Awards 2021

The Streaming Media European Readers' Choice Awards of 2021 has finished voting, thank you for your participation.

Our company was represented in five categories with various server and mobile products:

Softvelum team greatly appreciates all the votes that we got from our customers and partners, your support helps us promote our products and move forward our innovations!

July 15, 2021

Server playlists in Nimble Streamer

Server playlists feature set of Nimble Streamer allows creating live streams by composing VOD files into playlists.

Basic workflow looks like this:

  1. Prepare files for further live streaming.
  2. Create JSON playlist which contains description of generated output live streams and their respective files' playback scenarios.
  3. Create Live Transcoder scenarios to re-align output if needed.
  4. Configure Nimble Streamer to use the designated JSON playlist.
  5. Nimble instance generates output live steams according to playlist.
  6. You can update the playlist any moment, it will be picked up by Nimble within playlist sync interval.

The playlist is a JSON file with a simple grammar which allows setting streaming scenarios of various complexity.

Notice that server playlist feature set requires Live Transcoder to operate so you'll need to subscribe for a license and install/register Transcoder on your server.

For more details, read Server playlists page on Nimble Streamer website.

June 30, 2021

Q2 2021 news: Advertizer, KLV, Zabbix, Larix overlays

Let's see what Softvelum team has to show from Q2 of 2021.

Before moving to our features and updates, check Video Delivery Landscape of 2021 created and maintained by Divitel, we are excited to see Softvelum there.

Also take a look at our latest Secure your account in 3 easy steps article showing how you can use users management, two-factor authentication and cloud server backup to make sure WMSPanel is secured from multiple directions.


Nimble Streamer

A number of updates about our software media server were released.

  • Nimble Advertizer is now able to serve per-session ads using your own session handler. It allows making per-session ads insertion for each individual user by using customer-side session handler. You can also now get per-session statistics to accumulate ads insertion metrics for advertisers' confidence. Read this article to learn more.
  • Nimble DVR now has SCTE-35 support. Just enable "Keep SCTE-35 markers" option to allow recording SCTE-35 markers from incoming stream into DVR. Once HLS DVR is played, those markers are streamed along with the content.
  • Speaking of DVR, check our new DVR setup tutorial video on Softvelum YouTube channel.
  • KLV metadata is now supported in Nimble Streamer, read this article for more details.
  • SRT: we've updated libsrt to the latest 1.4.3 version for Nimble Streamer SRT package. You can now use custom parameters local_ip  and local_port in Push mode. Read our main SRT setup article to see other details.
  • Also check ATEM Mini streams RTMP to Raspberry Pi with Nimble Streamer to make SRT output - an interesting video from our user shows Nimble in action for SRT streaming.
  • Larix Grove URL and QR code can be generated in Nimble Streamer / Live streams settings under Global, Applications and Interfaces tabs to make RTMP and RTSP ingest, and also under and MPEGTS In tab for SRT and RIST. This way you can easily test your Nimble Streamer publishing settings with Larix Broadcaster even easier.
  • Zabbix: last but not least, we've released templates and configs for Zabbix monitoring of Nimble Streamer and SRT streams. System administrators can now add their media server to their Zabbix to keep tracking live streaming parameters of Nimble instances as well as SRT-specific stats.
  • We've also added peak daily connections as part of daily stats in WMSPanel.
  • WMSPanel now shows NVidia GPU stats and average RAM cache in each server details.


Mobile products

This week we'll release SDK updates for our mobile products - Larix Broadcaster and Larix Player.

We'll release related mobile SDKs' updates within a few days.


That's all for now. Follow us via social networks to get our updates as they appear: Twitter, Facebook, Telegram, LinkedIn, YouTube and Reddit.

June 7, 2021

Zabbix monitoring of Nimble Streamer, SRT streams and NVidia GPU

A continuous monitoring is needed if you want to ensure the correct functioning of any server. Getting basic metrics such as RAM, CPU or bandwidth usage helps rapidly responding to performance issues.

We’d like to show a way for our customers to add Nimble Streamer server-specific metrics as well as SRT streams statistics to an open-source Zabbix monitoring system. This will allow you to keep extended metrics in your monitoring system, which sometimes are not even available in WMSPanel, like RAM cache status for certain periods of time.

SRT metrics will also be helpful not only for stream health checks but also for analysis of how this new protocol works for your environment.

TLDR: You can find Zabbix templates and agent configs for Nimble Streamer general and SRT monitoring in this GitHub repo.


1. Monitoring features and metrics


We provide two Zabbix templates:

  1. General server metrics
  2. SRT metrics

Each template is based on certain calls of Nimble Streamer native API which we'll describe in later sections.

Take a look at this quick demo video showing all steps and brief capabilities overview



1.1 General metrics

The first TemplateNimbleServer.yaml in our github collects Nimble Streamer server statistics, which are obtained via /manage/server_status method of Nimble Streamer native API. This template includes some graphs for the RAMCache and FileCache monitoring.



You can find a detailed description of each parameter on server_status API method page.


1.2. SRT metrics

The second TemplateNimbleSRT.yaml on our github is made for monitoring the SRT streams statistics using /manage/srt_sender_stats and /manage/srt_receiver_stats API methods respectively. It uses Low Level Zabbix’s Discovery (LLD) to dynamically discover incoming and outgoing SRT streams metrics. For each connection (either sender or receiver) an Application with detailed statistics is created. Brief description is available if you move a mouse over [?] icon.

Please note, that the Nimble native API doesn’t send full scope of session metrics from SRT lib, but you can find full list at SRT library docs page as a reference.

Nimble Streamer returns the following list of values by API call that are also used in the Nimble SRT template:

  • SRT Receiver/Sender Window Flow
  • SRT Receiver/Sender Window Flight
  • SRT Receiver/Sender Window Congestion
  • SRT Receiver/Sender Total Packets Received 
  • SRT Receiver/Sender Total Packets Lost
  • SRT Receiver/Sender Total Packets Dropped 
  • SRT Receiver/Sender Total Bytes Received
  • SRT Receiver/Sender Status 
  • SRT Receiver/Sender Session Time
  • SRT Receiver/Sender RTT
  • SRT Receiver/Sender Retry Count
  • SRT Receiver/Sender Rate
  • SRT Receiver/Sender Packets Belated
  • SRT Receiver/Sender NAKs Sent 
  • SRT Receiver/Sender Estimated mbpsMaxBandwidth
  • SRT Receiver/Sender Estimated mbpsBandwidth
  • SRT Receiver/Sender bytesLost
  • SRT Receiver/Sender bytesDropped 

At the moment we don't have any SRT-specific Graphs or Triggers or Screens with the data we gather using these templates, but we'll probably add some improvements in future versions. We also have some plans on extending Nimble API capabilities to show more metrics and protocols, let us know if you're interested.


1.3 NVidia GPU monitoring

If you have NVidia GPU you can gather its metrics in Zabbix via Nimble’s Native API /manage/server_status method. We provide a Zabbix NVidia template and related config that relies on it to gather statistics. For temperature monitoring, the nvidia-smi utility is used, so make sure you can run it successfully from the console. Nimble Server NVidia template is actually the same, but with the addition of NVidia GPU discovery rules. To install it, follow the same steps and prerequisites as for Nimble Server Template.

Here are examples of stats.




1.4 Pre-requisites

The templates that we provide are made using Zabbix version 5.2.6, but the latest version of Zabbix should work fine too.

Notice that we used Ubuntu Linux 20.04 for Nimble Streamer server instance and setup details are provided for that OS. There should not be any issues with any Linux distro if you change package manager commands and repo names.

We didn’t check the provided templates with Nimble Streamer and Zabbix Agent running on Windows but it should also work as soon as you're able to install and run some releavant version of jq and cURL on it. Contact us if you succeed using this monitoring method for Windows version of Nimble.

We also assume that you already have the Zabbix server installed. If not, please find setup details on Zabbix's comprehensive documentation web pages. Zabbix provides many installation choices. The virtual appliance works best for us but you may choose the one that best suits your environment. If you have specific issues with it, ask Zabbix support or the Zabbix community forum.


2. Link Templates and Gather Data


Here are the steps you need to complete in order to add Nimble-specific metrics to your Zabbix server:

  1. Enable Nimble Streamer HTTP API for the server instance which you want to monitor
  2. Install cURL if it is not yet installed
  3. Import provided Templates to your Zabbix Server
  4. Install Zabbix Agent to your Nimble server
  5. Update Zabbix Agent configs to make templates work
  6. Add Nimble Streamer host to Zabbix server
  7. Apply Nimble Streamer Templates to a specific host

Here are the details of each step.


2.1 Enable Nimble Streamer HTTP API

Add ‘management_listen_interfaces = *’ to /etc/nimble/nimble.conf file. We have used the default port 8082 to make Zabbix Agent receive metrics from Nimble. Please follow this documentation page. to find out more about enabling Nimble API.


2.2. Install cURL

Make sure that cURL is installed on Nimble server. It can be used to make sure API is set up fine and it will be used further.


2.3 Import templates

Next, import Nimble Templates into Zabbix.

Download template files from respective GitHub repo to some location. From the Zabbix’s pop-up menu, go to Configuration -> Templates and click the Import button at the right top of the Templates page.

At the Import page, make sure that the checkboxes look like on the following screenshot and choose the template file to import.


Complete importing process by clicking the Import button. "Imported successfully" message should appear at the top of the page.

Navigate to the Templates page again, type "Nimble" at Name field and click Apply to set filter for Nimble templates only. You will notice two Nimble templates available.


2.4 Install Zabbix agent

You need to install Zabbix Agent on your Nimble Streamer server to send your server’s metrics to Zabbix Server. The easiest method is to install the agent using official Zabbix repositories by executing the following commands:

wget https://repo.zabbix.com/zabbix/5.2/ubuntu/pool/main/z/zabbix-release/zabbix-release_5.2-1+ubuntu20.04_all.deb

sudo dpkg -i zabbix-release_5.2-1+ubuntu20.04_all.deb

sudo apt update 

sudo apt install zabbix-agent jq

sudo systemctl enable zabbix-agent

The "jq" JSON parser is needed to process data from Nimble API responses.


2.5 Update Zabbix agent configs

Now complete the setup of Zabbix Agent by editing the /etc/zabbix/zabbix.conf file with your favorite editor to specify a Zabbix Server IP or Domain name via"Server=" option.

Please also add hostname of the server at "Hostname=" option and use the same name as Host name field when adding a host to Zabbix.

Download Zabbix Agent configs for Nimble from GitHub repo and copy them to /etc/zabbix/zabbix_agentd.d/.

wget https://raw.githubusercontent.com/WMSPanel/zabbix/main/zbx_nimble.conf zbx_nimble.conf

wget https://raw.githubusercontent.com/WMSPanel/zabbix/main/zbx_nimble.conf zbx_nimble_srt.conf

sudo cp *.conf /etc/zabbix/zabbix_agentd.d/


Check if your firewall allows incoming connection to port 10050 and start Zabbix Agent:

sudo systemctl start zabbix-agent


2.6 Add Nimble Streamer host to Zabbix

Switch to Zabbix web interface and add your Nimble Streamer server by clicking Create Host at Configuration -> Hosts page as usual. 

Select Interface type as Agent, and specify your Nimble server’s IP or domain name to connect to. Use Nimble group name and the same Host name as you defined in zabbix.conf file.


2.7 Apply Nimble Streamer Templates to a host

At the Templates tab select Host group Nimble and select Templates.


Click Add button after selecting Templates to add host to Zabbix. The added host will appear at the top of the page.

If you did everything correctly, you will notice availability status (ZBX) will become green after a few moments:


You may get the following error if the hostname in zabbix_agent.conf file is not matching the hostname you’ve entered when adding Hosts:



The setup is finished. Check if the metrics are being received by Zabbix at Monitoring -> Latest data menu. 

Now you can use this data to compare with data received by other templates or methods, create unique Graphs to add to the Dashboard or custom Screens, and add some Triggers to track if your server’s health goes down.


That's it. We'll continue improving our Zabbix monitoring templates and configs, stay tuned for updates and contact us if you have any questions or suggestions.


Related documentation

Nimble Streamer native API, Nimble Streamer configuration methods, Softvelum github

May 17, 2021

Secure your account in 3 easy steps

WMSPanel cloud control panel provides extended control over your Nimble Streamer server instances. Softvelum customers utilize Nimble extensively to build their media delivery networks and streaming infrastructure, and they use WMSPanel to perform the setup easily via both web UI and API.

Web services have a lot of upside like convenience of operations. However there is a downside to it as well: if you compromise your account credentials, an abuser may take control over your assets and do significant damage.

Besides evil intentions, people just make mistakes sometimes, so you need to improve your account security to avoid them.

Here are some general practices which we highly recommend for all of WMSPanel accounts and users to improve security and robustness of your account.


1. Users management: admins vs. non-admins


First, let's check what you can do on a company account and user level.

There are two types of users in WMSPanel: admins and non-admins.

Admins can do the following.

  • Install servers and register them in WMSPanel.
  • Add and change all Nimble Streamer settings on all servers.
  • View all stats for all servers.
  • Enable and disable statistical metrics.
  • Create and change subscriptions and view invoices.
  • Add, change and remove users.
  • Track users' activities log.
When you create an account in WMSPanel, your login becomes an admin user.

Non-admins cannot do much unless you allow them to:

  • They only view the stats in the data slices where they are assigned by admins.
  • Admins can grant non-admins specific permissions for specific servers, e.g. change live streaming settings only on a designated Nimble Streamer server instance. This article explains how it works.
  • Admins may also assign a group of non-admins to control a separate group of servers using data slices. This article describes the approach and its setup.

As you can see, there is no need to make some people the admin users while you can make them non-admins and grant only some limited permissions.

The rule of thumb is: don't grant too many permissions unless you really need to do that.


2. Two-factor authentication: a must-have


Being a user with WMSPanel login, you have to make sure your credentials are not exposed to anyone else. Doesn't matter if you are a full-scale admin or a non-admin who wants to view stats. However, if your credentials are obtained by malware or as a result of some sophisticated targeted attack, you need the second line of defense.

So you must enable two-factor authentication for your WMSPanel user. This is a modern de-facto standard for operations on the Internet so you must be familiar with it. So read this article to learn more about enabling 2FA.

Or just go to Settings menu, open Security tab and follow the instructions there.


3. Nimble config cloud backups: "undoing" mistakes


If you use a Nimble Streamer instance, one of your most valuable assets is its streaming settings. This is what you do as a streaming infrastructure architect and engineer: set up Nimble, test it with your source streams or files, launch it in products and make changes to those settings if necessary.

However, people make mistakes. Whatever you do to secure the users, those users can accidentally remove a server, erase some setting or make some experiment which would ruin the setup. You need to be able to overturn events like that.

We created Cloud backups of Nimble Streamer configuration to cover this use case. It allows making both manual and automated "snapshots" of server instance configuration which is then stored in WMSPanel cloud infrastructure. The key feature is that those backups cannot be erased or changed by any end user - admin or non-admin. When the backup is set and enabled for a specific server, those backups will allow restoring the state of settings of that server.

Cloud backups cost just 1 USD per month per backup.


These are simple rules to make your WMSPanel experience more secure and reliable.

Let us know if you have any questions.


May 13, 2021

KLV metadata in Nimble Streamer

KLV is a standard used for embedding meta information, usually into video feeds. Some of our customers use it for their use cases so they've been asking us for its support.

So now Nimble Streamer supports KLV metadata passthrough. It works for the following protocols:

  • MPEG-TS-based input: MPEG-TS over UDP and over HTTP, SRT and RIST.
  • Supported outputs are MPEG-TS over UDP, SRT and RIST.

So if your MPEGTS input has KLV streams in it, you'll be able to pass them through.

To enable this feature, you need to add max_forwarded_klv_streams parameter into Nimble config file to set the maximum number of expected streams.

max_forwarded_klv_streams=4

Don't forget to re-start Nimble instance to make it work.

You can use ffprobe to see how many KLV streams there are in your input MPEGTS streams.


Let us know if you have any questions and if you have any specific cases for KLV.


Related documentation

Nimble Streamer MPEGTS feature set

Advertizer per-session ads insertion and stats

Nimble Advertizer is a wide feature set for server-side ads insertion for Nimble Streamer software media server. It allows adding pre-roll and mid-roll ads into live and VOD streams according to customer business logic.

Dynamic ads can be inserted into HLS, RTMP, SLDP and Icecast outgoing live streams, along with HLS VOD output.

Previously the Advertizer allowed setting rules for ads insertion per application, per stream, as well as define per-user rules based on pay-per-view framework of Nimble Streamer.

With recent update it's now possible to apply Advertizer to more use cases:

  1. Make per-session ads insertion for each individual user by using customer-side session handler.
  2. Get per-session statistics to accumulate ads insertion metrics for advertisers' confidence.

In current article we'll focus on per-session mechanics without going into common details of Advertizer functionality like pre-requisites or general config grammar. You can find full set of details about Advertizer set setup and usage in Advertizer tech spec. You can also take a look Advertizer demo page showing the playback of all supported protocols with inserted ads.

Notice that per-session ads insertion is available for HLS, SLDP and Icecast outgoing live streams and for HLS VOD output. RTMP output is not supported for per-session scenarios.


1. Workflow


Let's see how Advertizer works with regular workflow and how it can be adjusted with per-session approach.

In a regular workflow the steps are straight-forward:

  1. Nimble Advertizer calls ads handler web application to get a business logic description.
  2. Advertising handler returns ads content description and applicability rules of what ads need to be inserted in which streams.
  3. Nimble Streamer gets the ads files and inserts their content them into output streams according to.

In per-session workflow, the second and third steps will have additional layer or activities.
  1. Your advertising handler (main handler) will now need to return the URL of per-session handler in addition to ads content description and applicability rules.
  2. Nimble Advertizer will call the per-session handler to send rules request and session info.
  3. Per-session handler gets the request and session info, and then makes the decision about who needs to watch which ads. It may also save the stats for further analysis.
  4. Per-session handler returns a set of rules specific to individual users.
  5. Nimble Advertizer serves ads according to new rules received by per-session handler, with session rules having priority over rules provided by the main handler.
Let's see how you can follow the described steps.

2. Enabling per-session handler


Per-session handler, just like main Advertizer handler, is a REST controller application which is called by Nimble Advertizer. You may use any language and framework to create it. Handler response must be a valid JSON text. Handler app must be available via HTTP/HTTPS protocol and accessible from Nimble instance.

Per-session handler URL needs to be returned in main handler's response in a new section called "session_handler" like shown below.

{

  "session_handler": {

    "apps": ["local", "remote", "live"],

    "url":"http://127.0.0.1:8085/session-handler"

  },

  "contents": [

    {"id":"1","uri":"http:\/\/127.0.0.1:8085\/ads\/1_180.mp4", "height":"180"},

    {"id":"2","uri":"http:\/\/127.0.0.1:8085\/ads\/1_240.mp4", "height":"240"},

    {"id":"3","uri":"http:\/\/127.0.0.1:8085\/ads\/1_360.mp4", "height":"360"},

    {"id":"4","uri":"http:\/\/127.0.0.1:8085\/ads\/1_360.mp4", "height":"480"},

    {"id":"5","uri":"http:\/\/127.0.0.1:8085\/ads\/pre-roll.aac.mp4"}

  ],

  "rules": [

  ]

}

This and other examples are available in Advertizer github repo, like this one.

The session_handler section may contain optional elements, here's full example:

"session_handler": {

  "url": "https://server/handler",

  "apps": ["app1", "app2"],

  "timeout": "1000",

  "onerror": "skip"

}

The elements mean the following:
  • url is the URL of per-session handler.
  • apps is a list of applications which are defined in Nimble Streamer, and which ads insertion will be applied to.
  • timeout is a period of time which Nimble can wait for per-session handler response. It's measured in milliseconds and it's 1000 by default.
  • onerror defines what action needs to be made is response is not received on time. It's "skip" by default which means that the playback will continue. The "stop" value means that the playback must be stopped and not processed since session handler has failed.

Let's see what happens when Nimble Streamer calls per-session handler.


3. Sending request to handler


First, open per-session-handler-request.json example from our github to illustrate the description below.

The request has two sections: session_info and rules_request. The handler is called in any of these events:

  • Each 30 seconds to send session_info part, you can use advertising_session_rules_request_interval to set it.
  • On each new connection, to send rules_request part. Also session_info can be sent with collected session information if available.
  • Each request_interval time slot, as defined in section 4.

Let's see what each section provides.

session_info is included in the request to show the statistics of ads viewership.

It has subsections for all sessions served by Advertizer since the last handler call, each session has its subsection. Here are the elements of each session:

  • session is a session ID.
  • app and stream show the stream which was served.
  • client_ip is the IP of a viewer.
  • user_agent is a data from viewer's User Agent header.
  • state is either active or inactive.
  • views contain data about ad viewership. It has 3 subelements: rule has ID of an ads rule applied, content has ID of ads content shown and uri is a URL of content applied.
  • userif you use pay-per-view framework, you will get user element which indicates viewer's user ID.

rules_request is included in requests to get response about what Nimble must do for sessions. Each session has individual section. It has the following elements.

  • session is s session ID.
  • protocol is a protocol type of a stream. This can be "hls", "sldp" or "icecast".
  • app and stream show the stream which is being served.
  • client_ip is the IP of a viewer.
  • user_agent is a data from viewer's User Agent header.
  • stream_time is a duration of a session prior to the moment when the request was sent

With these two sections, your per-session handler can make decisions about the ads content to be shown to your viewers.

Here's what per-session handler is expected to return to Nimble Advertizer.


4. Getting response from session handler


Take a look at response example below.

{
  "rules":
  [
    {
      "time_offset": 0,
      "time_sync": "stream",
      "type": "session",
      "id": 101,
      "contents": [{"id": ["1", "2", "3", "4"]}]
    }
  ],
  "rules_response":
  [
    {
      "rules": [101],
      "session": 5,
      "request_interval": 10
    }
  ]
}

You may also find it in Advertizer github.

Per-session handler response has two sections.

rules section defines rules for ads insertion which will be applied to sessions from rules_response section. The grammar for this section is the same as for main handler response describe in Advertizer spec. The only difference is that "type" element must always be "session".

These rules are also appended to the rules from the main handler response, so you may combine both per-session and default approaches to ads insertion. If there are per-session and main handler rules with the same rule ID, then session rule will be applied as it has priority.

rules_response section describes which ads insertion rules need to be applied to a particular session.

  • rules is the list of ads rules to be applied.
  • session is the ID of the session where rules are applied
  • request_interval defines how often after that Nimble Advertizer needs to request rules for this session.

The rules from response will be applied to current session once it's received by Advertizer. Some delay is possible in case of HLS just because of chunks download time and player reaction.



If you have any questions about Advertizer or per-session ads insertion, let us know.


Related documentation

Nimble AdvertizerNimble Advertizer spec, github repo for Advertizer,

March 30, 2021

Q1 2021 news: Cloud backups, Advertizer, Talkback, tutorials and more

Softvelum team kept working on products' improvements during Q1 of 2021 so let's see what we've got.


WMSPanel cloud service is the best way to control Nimble Streamer as it provides web UI for operating the vast majority of its features. Nimble settings are defined in the panel and synced up with Nimble instance.

Our customers always needed some way to back up their server settings, so in Q1 we introduced Nimble Cloud Backup. With cloud backups you can save streaming configs of your Nimble Streamer instances into distributed cloud database to improve redundancy. You can do it manually or automatically. If a customer needs to get back to some version of their cloud backup, he or she can restore it within WMSPanel account as a new server instance. After a backup is restored, a customer can initiate respective server instance to obtain these settings from WMSPanel, thus make new server to move into previous state. Learn more about Cloud Backups and try it in action.


Nimble Streamer Advertizer now has VOD server-side ads insertion. At the moment, HLS output with MPEG-TS and fMP4 containers is supported. Take a look at VOD SSAI overview article and also read tech spec page for full details.

We are working on other improvements for server-side ads insertion so stay tuned for updates.


SLDP protocol is actively used by our customers to provide low latency playback, especially after Flash has been officially abandoned and has put RTMP playback at denial. One of the use cases that our customers have discovered is browser-based videowall.

So we've made mosaic videowall demo page which allows playing 4 streams simultaneously. You can make your own mosaic using reference simplified version on our github. You may change it accordingly to make as many players as you need.


This quarter we also posted a number of video tutorials on our YouTube channel.

  • Converting NDI to Apple Low Latency HLS. There we demonstrate how you can use Nimble NDI feature set to get NDI input and transform it into Apple Low Latency HLS output. This can be a great case for customers who need to deliver live streams with low latency, check this brief description as well.
  • NDI to SRT to AWS Elemental MediaConnect via Nimble Streamer. Amazon recently introduced SRT support in their MediaConnect service. In our video we show how to take NDI input and use SRT feature set to stream live content out into AWS.
  • Widevine EZDRM setup in Nimble Streamer. Nimble DRM allows protecting live, DVR and VOD content using Widevine, Playready and FairPlay using various key management providers. This video uses EZDRM to show how easily you may set up Nimble DRM.
  • RTMP setup in Nimble Streamer. RTMP feature set is one of the most popular in Nimble Streamer so there we show all major scenarios related to the protocol.

Subscribe to our YouTube channel to get other upcoming updates.


Larix Broadcaster has got a number of significant updates this quarter.

Larix Broadcaster now supports Talkback, which is the ability to get an audio return feed. So while you generate a stream, you may get audio stream via SRT, RTMP, SLDP or even Icecast. This is a great feature for those creators who need to get a word from their studios while streaming live. Other updates of Larix include SRT Listen and Rendezvous support and some improvements for RIST protocol.

The talkback feature already got some attention. Take a look at How to: mobile reporter video demonstrating SRTMiniServer working with Larix Broadcaster by getting its live stream and returning talk back feed.

Larix Broadcaster SDK for Android and SDK for iOS pages now have architecture overview of Larix Broadcaster to make it easier to understand for beginners. They also refer to apps' sample code available on github.

Larix Player for Android and iOS was also updated with fresh libsrt and got SRT playback in Listen and Rendezvous modes.

You can learn more about Talkback, Larix SDKs updates and other features by reading this blog post.

Also, since we've mentioned MediaConnect, we've made Stream to AWS MediaConnect via SRT from Larix Broadcaster article describing the process in case you need to stream there from mobile.


That's all for now. Follow us via social networks to get our updates as they appear: Twitter, Facebook, Telegram, LinkedIn, YouTube and Reddit.

March 23, 2021

NDI to SRT to AWS Elemental MediaConnect via Nimble Streamer

AWS Elemental MediaConnect can now receive SRT streams which means a lot of streaming software is able to deliver content there.

Being an active SRT Alliance member, Softvelum is excited to see this shift. So we decided to make a simple video where we show how Nimble Streamer can get NDI stream, transform it into SRT and publish to MediaConnect.



The described use case gives flexibility for various live production teams which use NDI as their primary format and who want to deliver their content over Amazon network.

Feel free to try this approach in action.

Take a look at related materials as well:


March 22, 2021

Stream to AWS MediaConnect via SRT from Larix Broadcaster

AWS Elemental MediaConnect recently announced their support for SRT ingest. This makes it capable of receiving streams from all modern encoders, media servers and other SRT-powered software.

Larix Broadcaster has full support of SRT besides RTMP. RTSP and RIST, it also allows streaming simultaneously to multiple destinations.

Here's a brief instruction for how you can stream from Larix to MediaConnect using SRT.


Set up MediaConnect


First, log into your AWS account, go to MediaConnect control page and open Flows section.

Then click on Create flow to enter flow creation dialog.


First you enter a name for your flow and select availability zone if you need to. The Protocol name needs to be SRT listener. Another important field to specify is Inbound port which will be used for further connection setup. Also, White list CIDR block must be specified to allow your devices streaming into Media Connect.

Once you save the flow, you'll see the Inbound IP address, you'll need it for further connection setup.


Click on Start for your flow to get it ready to receive the content.

Now let's proceed with Larix Broadcaster setup.


Set up Larix Broadcaster via Larix Grove


If you haven't installed Larix Broadcaster yet, you can do it via Google Play and AppStore.

Now, in order to set up Larix easily, we'll use Larix Grove. Larix Grove format allows propagating streaming setting across devices using special URL. It can be distributed via QR code for easier import in your device. So we'll create QR code with our SRT ingest point.

Go to Larix Grove wizard, enter srt://ipaddress:port in URL field and select Caller mode for this connection. Click on QR code button to get the image.


Having the QR code, scan it it on our device using any app capable of that. It will open Larix Broadcaster and will import the connection settings automatically. All you'll need to do  after that it to tap on big red start button to initiate the stream.


Check MediaConnect input


Now go to MediaConnect, you'll see when your input becomes active.


You will see charts which means you have the stream up and running. You may also change flow setup if you need to make changes specific to SRT protocol, like set up latency, maxbw or other parameters.



You can use AWS Elemental MediaConnect documentation to see what you can do next with your stream.


Related documentation

SRT capabilities of Softvelum productsLarix Broadcaster, AWS Elemental MediaConnect docs