IBC '19: CMAF Ready for Wider Stage

Article Featured Image

The concept of a single set of files deliverable to all relevant end points has been the holy grail since the dawn of adaptive bitrate (ABR) streaming. The goal of the Common Media Application Format (CMAF) is to enable such a solution and slash encoding, storage, and bandwidth costs for companies who deploy it.

The good news is that the MPEG standard now carries momentum and the solid backing of Microsoft and Apple, both of whom were represented at an industry developers' forum in Amsterdam adjacent to the IBC Show. 

The forum was convened by the Web Application Video Ecosystem (WAVE), a Consumer Electronics Association initiative begun about the same time as CMAF to develop a global internet video ecosystem.

"CMAF can serve as the basis for DASH/HLS interoperability," said Krasimir Kolarov, acting chair of WAVE CMAF-IF who is also director, embedded media at Apple. "There is convergence."

The goal is to create a single format (fmp4 based) for multimedia content that can be delivered by various manifests (HLS or DASH) to a variety of client devices.

The potential for a lack of interop between DASH and HLS has concerned the industry for some time, but common sense has prevailed, with the forum working to ensure that the two protocols don't diverge.

"This codification and standardization can significantly improve cross-platform interoperability, reducing encoding, storage, distribution and continuing engineering costs – accelerating global web media growth," said Will Law, Chairman of WAVE Technical Working Group and chief architect, media cloud engineering at Akamai. "We're trying to get away from the competition like DVB versus ATSC. We want one OTT solution that works around the world."

The WAVE CMAF-Industry Forum outlined next steps for the fledging format, one of which was wider industry education about what is actually intended to achieve.

"There is market confusion that CMAF is a third delivery format, which it definitely is not," said Kolarov. "CMAF defines the standardized transport container for streaming VOD and linear media using DASH or HLS protocols. It is much more relevant to packaging and encoding. It is not another presentation format."

Another misconception is that CMAF solely targets live low latency.  A low latency mode within CMAF, which splits each segment into smaller units, or "chunks" of 500 milliseconds or lower, has been seized on by an industry desperate to for OTT delivery to match broadcast level delays. 

CMAF can certainly do this, but it is not the IF's main focus.

"When people hear CMAF they just think low latency," noted Thomas Stockhammer, director technical standards at Qualcomm and chair of CTA WAVE Device Playback Task Force (DPCTF). "We need to make clear that this is one application of CMAF but not its main feature."

Another key aim is tofacilitate wide interoperability in the growing market of CMAF-based solutions by using a common set of content and device playback specifications, as well as test conditions and material. 

"From an encoding point of view, regardless of whether you have HLS low latency low or the DASH version of ULL [ultra-low latency], the encode is still the same," Stockhammer said. "The same on playback … whether your content is DASH or HLS, both should work if you follow the CMAF implementation and device specs."

At its Worldwide Developer Conference earlier this year, Apple had put a potential fly in the ointment by announcing specs for a Low-Latency HLS. While reducing latency for live streaming is a common goal, this news interrupted the industry-wide effort to do so via the chunked transfer encoding of CMAF.

"Apple's low latency spec is still in draft stage," said Kolarov. "We're collecting feedback and will accommodate this feedback into the protocol. This is just how HLS was introduced in the beginning."

He confirmed, "Low latency CMAF support is on our roadmap. Since we [Kolarov's team] are dedicated to CMAF, we've been encouraging others at Apple to participate."

New technologies for possible inclusion in CMAF include HDR and audio profiles; sequencing and splicing of presentations; supplemental data brands; timed metadata and random access to tracks are also under consideration.

WAVE CMAF-IF has also agreed in principle with the 3GPP for the use of CMAF as a media streaming format over 5G.

Also under consideration is creation of a CMAF IF web site with free access to CTA WAVE content and device playback specs and testing.

"It can be the one place companies can go and have pointers for all the different specs being developed," Law said.

More meetings like the one in Amsterdam will promote CMAF activity at CES, NAB, and at events in Asia.

Other points on the agenda include efforts to develop a "heat map" of relevant standard bodies, industry for a, and target members, to encourage existing CTA WAVE members to promote CTA WAVE within their own companies and outreach to CTA WAVE membership for participation in outward-focused messaging for WAVE and CMAF-IF.

There is also potential for rebranding the WAVE specs to "CMAF-IF Content Spec" and "CMAF-IF Device Playback Spec."

To reiterate, the main goals of CMAF are: to reduce overhead and delivery costs through standardised encryption methods; to simplify complexities associated with streaming workflows and integrations (such as DRM, closed captioning, caching); and to support a single format that can be used to stream across any online streaming device.


Streaming Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues
Related Articles

IBC: We've Been Apart for Too Long

The international focus of events like IBC generate a sense of excitement and a critical mass of activity that drives the whole industry forward. That momentum has been noticeably lacking in 2021, and it's time to get the ball rolling again.