Low Latency, Real-Time Streaming, and the State of Innovation: Reflections from a Season of Streaming Technology Conferences
By Romain Bouqueau, Founder of Motion Spell
Last October, I had the chance to attend Demuxed 2024 in San Francisco. While it was great to reconnect with so many brilliant people in our industry, my primary goal in the U.S. was to visit customers — and to run two training sessions, including one on the future of low-latency streaming.
I’ve been a long-time advocate of innovation in this space. At Motion Spell, our work on GPAC has helped shape key building blocks for low latency and interactivity. GPAC now includes a powerful JavaScript engine, a flexible 2D/3D compositor, and advanced metadata processing tools — all of which are becoming increasingly relevant as the streaming landscape evolves.
But while technical capabilities are advancing, the path forward is anything but straightforward. Later in February, my visit to present two research papers at Mile High Video in Denver just made this feeling stronger.
WebRTC: A Case Study in Missed Opportunity
One of the topics that came up in both informal conversations and my training was WebRTC. I’ve long thought WebRTC was technically the right move. It brought peer-to-peer real-time communication to the browser, which was revolutionary.
But here’s the problem: Google essentially treated interoperability the same way Adobe did with RTMP — a proprietary black box approach that ultimately harmed the ecosystem. Outside of Google, the WebRTC community was vibrant and full of promise. But over the years, it was rocked by internal fragmentation and key events, like the untimely passing of Dr. Alex Gouaillard — one of the few truly independent voices pushing for open innovation.
There was hope that WebRTC could move beyond video conferencing to challenge traditional ABR-based delivery systems like DASH, HLS, or CMAF. This led to the creation of WHIP and WHEP — extensions that aim to bridge WebRTC into broader streaming workflows.
But as we’ve seen recently — including public disputes like this WHIP/WHEP fork debate — the situation is now divisive, politically fraught, and technically stalled. The IETF is often seen as the ultimate standardization body, but WebRTC is proof that technical standards don’t matter if the community and deployment strategies don’t align with real-world needs.
My personal take? WebRTC can’t be saved — not because it’s broken, but because its governance never focused on solving the right problems.
MoQ and the QUIC Future
Another hot topic at Demuxed was the rise of Media over QUIC (MoQ). In my training, I covered MoQ alongside HTTP/3, WebTransport, and the broader QUIC ecosystem.
On paper, MoQ offers a bold vision: ultra-low-latency, multicast-friendly, end-to-end media transport over modern internet infrastructure. In practice? It’s early.
I had a fascinating chat with Luke Curley, who recently forked MoQ into a new implementation. The move sparked questions about governance, fragmentation, and what kind of technical or business consensus MoQ will really need to succeed.
Then comes the elephant in the room: How much QUIC-delivered content is actually feasible today? Google might report 60% UDP penetration for their own services, but:
- What about corporate firewalls?
- What about the rest of the open internet?
- How many CDNs are QUIC-ready at scale?
The Packaging Layer Still Matters
With all the protocol-level experimentation, one thing remains consistent: you still need a stable, standards-based packaging layer.
This is why CMAF continues to matter — but with a caveat. At Motion Spell, we believe there’s a need for a simplified CMAF profile, tailored for real-time use cases:
- Stronger profile constraints leading to a single way to code any feature
- Stronger conformance including test vectors and a conformance checker
GPAC is already well-aligned with this vision. It’s built to ingest, process, and repackage live content in versatile ways — and can integrate with future transport layers like QUIC or MoQ when they’re ready.
Where Do We Go From Here?
If there’s one thing Demuxed 2024 made clear, it’s this: the low-latency streaming space is in flux. If there’s one thing Mile High Video 2025 made clear, it’s this: the streaming space moves where the money is, and the hot bet is on live sports, gaming, and online betting. The tools are getting more powerful. The protocols are evolving. But we still haven’t reached consensus on what “next-gen streaming” should actually look like.
In the meantime:
- WebRTC’s promise remains unfulfilled
- MoQ isn’t production-ready, and QUIC is still more theory than deployment
- And yet — innovation continues, just beneath the surface. Want to know who uses GPAC? Check out my conversation with Russell Trafford-Jones and Allan McLennan (here).
At Motion Spell, we’re staying pragmatic. We’ll support what’s real today, and we’ll help our clients prepare for what’s coming. Whether it’s GPAC’s packaging engine, our work on multicast-ABR, or our deep involvement in open standards — we’re committed to helping the ecosystem move forward.
Want to explore what low latency streaming could look like in your project?
Let’s talk. We’re building the future — one codec, one protocol, one filter at a time.
Contact us!
