Glass-to-glass latency
Glass-to-glass latency is the total delay between light hitting a camera lens and the matching image appearing on a viewer's screen. It adds up every step along the way: capture, encoding, upload, processing, CDN delivery, player buffering and decoding. Measuring it end to end shows the real delay a viewer experiences, not just one part of the chain.
Why it matters
It is the only latency number that matches what viewers feel. Teams that tune one step, such as the encoder, can still ship a slow stream because the player buffer or CDN adds the delay. Measuring the whole chain shows where seconds are lost before a big event.
How it works
- Point the camera at a running clock, such as a stopwatch or a timecode display.
- Stream that shot through your full setup: encoder, ingest, processing, CDN and player.
- Place the viewing screen next to the original clock and photograph both together.
- Subtract the time shown on screen from the real clock time to get the total delay.
- Repeat on different devices and networks, since player buffers and connections change the result.
How do you measure glass-to-glass latency?
You measure glass-to-glass latency by filming a running clock and comparing it with the same clock shown on the viewer's screen. Take a photo with both visible and subtract the two times. Burned-in timecode or a clock overlay added by the encoder makes the reading easier and more repeatable.
What is a good glass-to-glass latency for live streaming?
A good glass-to-glass latency depends on the use case. For example, about 2 to 5 seconds suits most sports and live shopping, 10 to 30 seconds is normal for standard HLS broadcasts, and under one second is needed for betting, auctions and two-way conversation. Pick the highest delay your product can tolerate, since lower targets cost more.
What adds the most delay to a live stream?
The player buffer and segment length usually add the most delay in a live stream. A player that waits for three 6-second segments starts about 18 seconds behind before any network time is counted. Encoder settings, transcoding and CDN delivery add smaller amounts. Shorter segments or partial segments cut the biggest share.
Where it's used
Engineers at live sports streamers and live event producers measure it before big broadcasts. Betting apps and game streaming platforms use it to confirm that viewers see action fast enough for chat and wagers.
By the numbers
- The HLS specification (RFC 8216) states that a typical target duration for a segment is 10 seconds, which is one reason standard HLS streams run well behind real time. (RFC 8216: HTTP Live Streaming (IETF))
