For many years before MPEG-DASH came on to the scene, there was a three-way tussle between Smooth Streaming, Apple HLS and Adobe HDS, all of which are very similar in technology, and all of which use http as the protocol. But despite the similarities, three standards meant that there had to be three different players.
But now two of these standards are getting long in the tooth, and while Silverlight and Flash are not quite defunct yet, they are no longer supported in Chrome and will be absent from most browsers within two years. So it has become necessary for service providers and content producers to look for new formats.
MPEG-DASH has been thrust into this vacuum, and is seen by some as the new standard for the future, replacing those that have fallen out of favour and been abandoned for security reasons, and providing a way to standardise and rationalise across the industry.
Accordingly, Vimond has added MPEG-DASH support to its product line, and this gives users the ability to continue to use any of existing formats while being ready to switch to MPEG-DASH as and when required.
Some fundamental differences between MPEG-DASH and the previous generation of players are apparent. Since security has been a major issue with Flash among others, it’s good to note that there are no specific security implications so far with MPEG-DASH. Each implementation is native within the browser or player; our support for the standard is based on the DASH industry forum’s reference player, and there is no need for browser plug-ins or other third-party modifications – the browser itself is enough. This alone makes things much simpler for both consumers and the service providers.
In production, the MPEG-DASH based workflow is very simple and streamlined. Ingest is via the unified streaming platform using mp4 files as input, and MPEG-DASH, HLS, and the other formats are created as output. MPEG-DASH files are typically smaller than HDS as output files – a useful advantage.
One of the major upsides with MPEG-DASH is that it can easily adapt to new and emerging video and audio codecs, such as VP10, in contrast to the previous generation of streaming standards, which were tied to specific video and audio codecs. A feature like this is a powerful incentive to developers to work with MPEG-DASH, and that is an important factor in creating enough momentum behind the standard.
But, set against these benefits, some aspects of MPEG-DASH are not completely ideal.
As is common when any new standard begins to emerge, not everything is totally defined and there is still room for manoeuvre as various interested parties in the industry lobby for the flavour of the standard that they prefer. As a result MPEG-DASH specifications are still very wide, and this can lead to compatibility problems with different players. With time these may disappear as the standard becomes tighter, but meanwhile manufacturers have to find a way of ensuring that customers can operate successfully right now. In practice the reality is that supporting MPEG-DASH means supporting a few different variations of it. To achieve this, we have to make some adjustments in the manifest files that we receive from the server, so that they will play correctly in all devices. Individual problems in each specific player have to be rooted out and fixed. Then, as new refinements of the standard are adopted, and as new versions of Unified Streaming are released, we have to check to maintain compatibility so that no new glitches are introduced by the changes.
Some other practical aspects of MPEG-DASH are not clear yet; recent unresolved questions over licensing issues are an example. These and a natural tendency to ‘see which way the wind is blowing’ means that there is still a way to go before general agreement is reached on MPEG-DASH as the first choice format.
From a manufacturer’s point of view, if MPEG-DASH were to replace HLS, Smooth Streaming and all the other formats, it would be a very welcome development. Unfortunately at this stage, it’s impossible to say if enough industry momentum will gather behind the new format to guarantee its progress towards a position as the dominant standard. So at the moment, MPEG-DASH is just another standard – not the replacement that so many people hoped it would be, and still hope it will be one day.
Taking into consideration the natural life-cycles of the older standards, it’s clear that Smooth Streaming and HDS are on the way out, so the most likely near-future is one in which we see HLS and MPEG-DASH running side-by-side for a while, and the industry continuing to live with multiple protocols for some time yet.
Anders Instefjord
Streaming Specialist Vimond Media Solutions