Low latency

Video: Integrating CMAF Into A VOD Workflow

Posted on by Russell Trafford-Jones

CMAF is often seen as the best hope for streaming to match the latency of broadcast. Fully standards based, many see this as the best route over Apple’s LL-HLS. Another benefit of it over LL-HLS is that it’s already a completed standard with growing support.

This talk from Tomas Bacik starts by explaining CMAF to us. Standing for the Common Media Application Format, it’s based on the standardised ISOBMFF container format and whilst CMAF isn’t by default low-latency, it is flexible enough to deliver just that. However, as Tomas from CDN77 points out, there are other major benefits in terms of its use of the Common Encryption format, reduces storage fees and more.

MPEG DASH is a commonly found streaming format based on ISO BMFF. It has always had the benefit of supporting other codecs such as HEVC and AV1 over HLS which is an AVC-only specification. CMAF is an extension of MPEG DASH which goes one step further in that it can deal with both HLS-style manifest files (.hls) as well as MPEG DASH format (.mpd) inheriting, of course, the multi-codec ability of DASH itself.

Next is central theme of the talk, looking at VoD workflows showing how CMAF fits in and, indeed, changes workflows for the better. CMAF directly impacts packaging, storage and CDN which is where we focus now. Given that some devices can play HLS and some can play DASH, if you try to serve both, you will double your requirements of packaging, storage etc. Dynamic packaging allows for immediately repackaging your chunks into either HLS or DASH as needed. Whilst this reduces the storage requirements, it increases processing and also increases the time to first byte. As you might expect, using CMAF throughout, Tomas explains in this talk, allows you to package once and store once which solves these problems.

Tomas continues by explaining the DRM abilities of CMAF including AES-CBC and finishes by taking questions from the audience.

Watch now!
See Streamflow’s blog post supporting the talk
Speakers

Tomas Bacik Tomas Bacik
VP of Product Development, Streamflow by CDN77
CDN77

Video: Three Roads to Jerusalem

Posted on by Russell Trafford-Jones

With his usual entertaining vigour, Will Law explains the differences to the three approaches to low-latency streaming: DASH, LHLS and LL-HLS from Apple. Likening them partly to religions that all get you to the same end, we see how they differ and some of the reasons for that.

Please note: Since this video was recorded, Apple has released a new draft of LL-HLS. As described in this great article from Mux, the update’s changes are

  • “Delivering shorter sub-segments of the video stream (Apple call these parts) more frequently (every 0.3 – 0.5s)
  • Using HTTP/2 PUSH to deliver these smaller parts, pushed in response to a blocking playlist request
  • Blocking playlist requests, eliminating the current speculative manifest request polling behaviour in HLS
  • Smaller, delta rendition playlists, which reduces playlist size, which is important since playlists are requested more frequently
  • Faster rendition switching, enabled by rendition reports, which allows clients to see what is happening in another playlist without requesting it in its entirety”[0]

Read the full article for the details and implications, some of which address some points made in the talk.

Furthermore, THEOplayer have released this talk explaining the changes and discussing implementation.

Anyone who saw last year’s Chunky Monkey video, will recognise Will’s near-Oscar-winning animation style as he sets the scene explaining the contenders to the low-latency streaming crown.

We then look at a bullet list of features across each of the three low latency technologies (note Apple’s recent update) which leads on to a discussion on chunked transfer delivery and the challenges of line-rate delivery. A simple view of the universe would say that the ideal way to have a live stream, encoded at a constant bitrate, would be to stream it constantly at that bitrate to the receiver. Whilst this is, indeed, the best way to go, when we stream we’re also keeping one eye on whether we need to change the bitrate. If we get more bandwidth available it might be best to upgrade to a better quality and if we suddenly have contested, slow wifi, it might be time for an emergency drop down to the lowest bitrate stream.

When you are delivered a stream as individual files, you can measure how long they take to download to estimate your available bandwidth. If a file can be downloaded at 1Gbps, then it should always arrive at 1Gbps. Therefore if it arrives at less than 1Gbps we know that there is a bandwidth restriction and can make adjustments. Will explains that for streams delivered with chunked transfer or in real time such as in LL-HLS, this estimation no longer works as the files simply are never available at 1Gbps. He then explains some of the work that has been undertaken to develop more nuanced ways of estimating available bandwidth. It’s well worth noting that the smaller the files you transfer, the less accurate the bandwidth estimation as TCP takes time to speed up to line rate so small 320ms-length video segments are not ideal for maximising throughput.

Continuing to look at the differences, we next look at request rates with DASH at 20 requests per second compared to LL-HLS at 720. This leads naturally to an analysis of the benefits of HTTP/2 PUSH technology used in LL-HLS and the savings that can offer. Will explores the implications, and some of the problems, with last year’s version of the LL-HLS spec, some of which have been mitigated since.

The talk concludes with some work Akamai has done to try and establish a single, common workflow with examples and a GitHub repository. Will shows how this works and the limitations of the approach and finishes with a look at the commonalities in approaches.

[0] From “Low Latency HLS 2: Judgment Day” https://mux.com/blog/low-latency-hls-part-2/

Watch now!
Speakers

Will Law Will Law
Chief Architect,
Akamai

Video: There and back again: reinventing UDP streaming with QUIC

Posted on by Russell Trafford-Jones

QUIC is an encrypted transport protocol for increased performance compared to HTTP but will this help video streaming platforms? Often conflated with HTTP/3, QUIC is a UDP-based way evolution of HTTP/2 which, in turn, was a shake-up of the standard HTTP/1.1 delivery method of websites. HTTP/3 uses the same well-known security handshake from TLS 1.3 that is well adopted now in websites around the world to provide encryption by default. Importantly, it creates a connection between the two endpoints into which data streams are multiplexed. This prevents the need to constantly be negotiating new connections as found in HTTP/1.x so helping with speed and efficiency. These are known as QUIC streams.

QUIC streams provide reliable delivery, explains Lucas Pardue from Cloudflare, meaning it will recover packets when they are lost. Moreover, says Lucas, this is done in an extensible way with the standard specifying a basic model, but this is extensible. Indeed, the benefit of basing this technology on UDP is that changes can be done, programmatically, in user-space in lieu of the kernel changes that are typically needed for improved TCP handling on which HTTP/1.1, for example, is based.

QUIC hailed from a project of the same name created by Google which has been taken in by the IETF and, in the open community, honed and rounded into the QUIC we are hearing about today which is notably different from the original but maintaining the improvements proved in the first release. HTTP/3 is the syntax which is a development on from HTTP/2 which uses the QUIC transport protocol underneath or as Lucas would say, “HTTP/3 is the HTTP application mapping to the QUIC transport layer.” Lucas is heavily involved within the IETF effort to standardise HTTP/3 and QUIC so he continues in this talk to explain how QUIC streams are managed, identified and used.

It’s clear that QUIC and HTTP/3 are being carefully created to be tools for future, unforeseen applications with clear knowledge that they have wide applicability. For that reason, we are already seeing projects to add datagrams and RTP into the mix, to add multiparty or multicast. In many ways mimicking what we already have in our local networks. Putting them on QUIC can enable them to work on the internet and open up new ways of delivering streamed video.

The talk finishes with a nod to the fact that SRT and RIST also deliver many of the things QUIC delivers and Lucas leaves open the question of which will prosper in which segments of the broadcast market.

The Broadcast Knowledge has well over 500 talks/videos on many topics so to delve further into anything discussed above, just type into the search bar on the right. Or, for those who like URLs, just add your search query to the end of this URL https://thebroadcastknowledge.com/tag/.

Lucas has already written in detail about his work and what HTTP 3 is on his Cloudflare blog post.

Watch now!
Speaker

Lucas Pardue Lucas Pardue
Senior Software Engineer,
Cloudflare

Video: Low Latency Streaming

Posted on by Russell Trafford-Jones

There are two phases to reducing streaming latency. One is to optimise the system you already have, the other is to move to a new protocol. This talk looks at both approaches achieving parity with traditional broadcast media through optimisation and ‘better than’ by using CMAF.

In this video from the Northern Waves 2019 conference, Koen van Benschop from Deutsche Telekom examines the large and low-cost latency savings you can achieve by optimising your current HLS delivery. With the original chunk sizes recommended by Apple being 10 seconds, there are still many services out there which are starting from a very high latency so there are savings to be had.

Koen explains how the total latency is made up by looking at the decode, encode, packaging and other latencies. We quickly see that the player buffer is one of the largest, the second being the encode latency. We explore the pros and cons of reducing these and see that the overall latency can fall to or even below traditional broadcast latency depending, of course, on which type (and which country’s) you are comparing it too.

While optimising HLS/DASH gets you down to a few seconds, there’s a strong desire for some services to beat that. Whilst the broadcasters themselves may be reticent to do this, not wanting to deliver online services quicker than their over-the-air offerings, online sports services such as DAZN can make latency a USP and deliver better value to fans. After all, DAZN and similar services benefit from low-second latency as it helps bring them in line with social media which can have very low latency when it comes to key events such as goals and points being scored in live matches.

Stefan Arbanowski from Fraunhofer leads us through CMAF covering what it is, the upcoming second edition and how it works. He covers its ability to use .m3u8 (from HLS) and .mpd (from DASH) playlist/manifest files and that it works both with fMP4 and ISO BMFF. One benefit from DASH is it’s Common Encryption standard. Using this it can work with PlayReady DRM, Fairplay and others.

Stefan then takes a moment to consider WebRTC. Given it proposes latency of less than one second, it can sound like a much better idea. Stefan outlines concerns he has about the ability to scale above 200,000 users. He then turns his attention back to CMAF and outlines how the stream is composed and how the player logic works in order to successfully play at low latency.

Watch now!
Speakers

Koen van Benschop Koen van Benschop
Senior Manager TV Headend and DRM,
Deutsche Telekom
Stefan Arbanowski Stefan Arbanowski
Director Future Applications and Media,
Fraunhofer FOKUS