Latency / Glass-to-Glass Delay

What is it?

Latency is the total delay between something happening in front of the camera and the viewer seeing it on their screen. "Glass-to-glass" makes the definition concrete: from the glass of the camera lens to the glass of the viewer's display. Every step of the chain — sending, processing, packaging, delivery, playback buffering — adds a little to it.

Practical example

A host says "Hello everyone, wave in the chat!" On a classic setup, viewers hear it 25 seconds later, type their waves, and the host sees responses about half a minute after asking — conversation feels broken. On a low-latency setup (3–5 seconds), the chat lights up almost immediately and the show feels like a real shared moment. Same host, same audience; the only difference is latency.

Key things to know (non-technical)

• Rough mental map: WebRTC ≈ under 1s · LL-HLS ≈ 2–5s · classic HLS ≈ 15–45s.
• Lower latency = more interactive, but usually more expensive and less forgiving of bad networks. There's always a trade-off triangle: latency vs. cost vs. stability.
• Different shows need different latency: an auction or quiz needs seconds; a concert replay-style broadcast can tolerate half a minute.
• Latency is also why "live" reactions on different platforms arrive at different times during simulcast.

In Tupic Live

Tupic Live can treat latency as a product setting, not a fixed fact: "Interactive mode" (low latency, for Q&A/shopping/gifting) vs. "Quality mode" (higher latency, maximum smoothness for performances) — letting creators pick what their show needs.