A live roulette broadcast begins with a physical event: a dealer operates a wheel, and a ball settles into one of its pockets. The picture on a phone or computer carries that activity beyond the table. Between the camera and the display, equipment and software prepare, transmit, and reproduce the moving image.
That journey explains why “live” and “instantaneous” describe different things. A live stream follows an event in progress, with an interval between capture and playback. Understanding that interval also separates picture detail from timing. A sharp image tells us something about what is visible, without revealing exactly when the camera captured it or how long it took to reach the display.
The camera captures successive images
A camera supplies the view of the dealer and wheel for someone following live roulette online. The round takes place at the table, while its image appears within a game window. That window also provides the controls used to participate. The footage and controls share a display, although capturing the physical action and receiving a player’s input are different tasks.
The moving picture consists of successive images, called frames. Each frame contains a view captured at a particular moment in the round. Played in sequence, those images show the wheel’s movement and the dealer’s actions over time. A single frame supplies one view of that activity; the sequence carries its progression. What appears as continuous movement on screen therefore reaches the viewer through an ordered series of pictures. Two specifications describe that series. Resolution concerns the number of pixels making up each picture, while frame rate counts how many frames are presented each second. They describe image dimensions and the sequence used to show movement. Neither specification, by itself, states the elapsed time between an action at the table and its appearance on screen.
Encoding and playback carry the moving picture
Encoding prepares the captured video for transmission. Compression represents its picture information with less data than an uncompressed version requires. The authors of a 2025 study of live-streaming latency describe encoding methods that use information within individual pictures and relationships between pictures. Some frames provide reference information used to reconstruct others. The receiving software therefore works with a sequence, rather than treating every picture as an unrelated photograph.
The study examines how video is organized into groups of related pictures. A receiver joining that sequence needs the reference information required to decode later dependent pictures. Starting playback is therefore connected to the structure of the encoded video, as well as its delivery. The researchers investigated that relationship in particular streaming configurations; their results describe those systems, rather than identifying how any roulette provider organizes its broadcast.
The encoded information travels across a network to the receiving device. During a broadcast, capture, transmission, and playback overlap: the camera continues to capture later pictures while earlier pictures are being delivered and displayed. Each displayed picture has passed through the same stages, although the overall sequence is still growing. An ongoing event can therefore be transmitted without waiting for the whole event to finish before playback begins. The viewer receives the unfolding sequence as the system continues to supply it.
At the receiving end, buffering holds video information before playback. Decoding reconstructs the pictures from the encoded information, and rendering presents them on the display. Receiving the data and presenting its pictures are separate stages. The image on screen is the output of that process, showing the view captured earlier at the table.
The study describes the total capture-to-display interval as end-to-end latency, with encoding, transmission, buffering, and decoding contributing to it. Its researchers measured a difference between a source capture window and a playback window. That comparison addresses the age of the received picture relative to its source, rather than judging how smoothly the picture moves.
Picture detail and arrival time need separate measurements
In June 2026, What Hi-Fi?’s Tom Parsons tested ways of watching a World Cup match using identical televisions. He matched their software and picture settings, then compared reception through an aerial with streaming alternatives. Using the same event allowed him to compare when corresponding moments appeared on the different screens.
In his side-by-side broadcast test, BBC iPlayer arrived about 23 seconds after Freeview. He also found no meaningful timing difference between iPlayer’s HD and UHD streams. In that comparison, choosing the higher-resolution stream did not add an observable delay relative to the service’s lower-resolution version. These findings describe his football broadcast, without measuring a roulette feed’s timing.
Freeview was the reference in that comparison, so the 23 seconds describes the gap between two reception methods. It does not measure the whole journey from the match to the iPlayer screen. The aerial picture had already traveled through a broadcast route of its own. A comparison between two screens establishes their relative timing; a capture-to-display measurement starts further back, at the source. Both provide timing evidence, but they answer different questions about the route a moving picture takes before it becomes visible to someone watching remotely.
By the time the roulette picture appears, the receiving device has reconstructed an earlier view captured at the table. The wheel’s movement reaches the viewer through that sequence of pictures, with capture and delivery continuing as playback proceeds. The physical round supplies the action; the broadcast carries its changing appearance to another place, where the device turns the received information into the view on screen.