How Latency Affects Dialogue and Timing in Online Performance

What latency does to dialogue and timing

Latency is the delay between an action and its reception over a network. In online performance it is the gap between a gesture, a line or a musical note and the moment another participant sees or hears it. Jitter is the variation in that delay; lag is a looser everyday term. A production may have low average latency and still be difficult to play if the delay keeps changing. Latency matters to theatre because timing is not decoration: pauses, overlaps, cue pickups and silences carry dramatic meaning and comic rhythm. When the feedback loop between performers is delayed, those timings become approximate and must be re-authored for the network.

The most immediate effect is on turn-taking. A 2020 study of 25 video consultations in the Journal of Pragmatics found that latency can cause participants to perceive silence at points where talk should occur, to talk in overlap, and to struggle to return to one speaker at a time. The study concludes that participants act as if they share the same reality until latency becomes problematic. Theatre magnifies the problem because its pauses and overlaps are often scored, repeated and judged.

For directors and performers, the first practical conclusion is therefore not that online performance is impossible. It is that transparent, unmediated conversational timing cannot be assumed. A scene written for rapid interruption, a comic beat that depends on a half-second pause, or a cue that depends on simultaneous response may need to be re-blocked, re-paced or re-written for the network.

Three timing systems affected by delay

Conversation and dialogue

Stage dialogue is often described as alternating lines, but it depends on finer signals: breath, gaze, small listener responses, the slight overlap that shows agreement or disagreement. When these signals arrive late, an actor may read a delayed nod as indifference, or a delayed ‘mm-hm’ as an interruption. The speaker may already have continued, so the listener’s response lands in the wrong place. Video adds a second layer: the image of a face, the beginning of a gesture or the moment of eye contact may lag behind the voice. The video-consultation research is not theatre-specific, but it identifies the underlying interactional mechanisms clearly: delayed feedback creates false silences and false overlaps.

Cues and stage management

Scripted theatre often relies on a shared clock. A stage manager calls ‘standby’ and ‘go’; an actor waits for a door slam; a lighting change lands on a line. In a networked rehearsal, the call may arrive after the line it was meant to support, and the lighting change may appear to the remote audience at a different moment from the one the performers intended. Cue stacks become fragile when each link adds delay. Practical protocols include a single cue-giver, sequential rather than simultaneous cues, count-ins that allow a settling beat, and explicit confirmation that a cue has been received. These are not merely technical fixes. They change the rhythm of the scene and should be rehearsed as part of its dramaturgy.

Ensemble, music and movement

Telematic performance makes the problem visible because music and movement depend on synchronisation. Rofe and Reuben’s 2017 article on telematic performance addresses latency as a challenge for performance across networks. In music, the effects are measurable: a 2024 study of five duos of professional jazz musicians notes that network delay and jitter can cause significant deteriorations in synchrony and stability between performers, and found that the duos adopted diverse strategies when coordinating over a network. That diversity is important. There is no single natural adaptation; performers change how they listen, how they lead and how they follow. In theatre, chorus work, movement sequences and live music are especially exposed. Strict unison may become impossible, while canon, call-and-response, fixed media or a shared click can provide a workable structure.

Why the problem is easy to miss

Latency is often invisible in a clean video window. Participants compensate without noticing: they wait a little longer, repeat a word, or fill a gap with a gesture. The 2020 study describes latency as unnoticed until it becomes problematic. In rehearsal, a missed cue may be blamed on inattention when it is actually a delay. The problem is compounded because latency is not always symmetrical. One performer may hear another later than the reverse, and audio and video may have different delays. Schoenberg’s 2016 dissertation on mediated conversation under transmission delay treats delay as part of the perceived quality of mediated conversation, not simply as a network metric. For a company, this means that a single speed test is not enough. The working delay needs to be measured in the conditions and platforms the production will actually use.

Practical strategies for rehearsal and performance

No universal latency threshold applies to all online theatre. A slow monologue can tolerate delay that would destroy a fast farce, and a musical ensemble may have a tighter tolerance than a split-screen duologue. The practical task is to design for the delay you have, rather than assuming transparent real time.

1. Measure before directing. Run test calls between the actual locations and platforms, note the audio and video delay, and check whether it changes during the session. Record the conditions if the company keeps technical notes. Direct from evidence rather than from the impression of a single call.
2. Write cueing into the scene. Replace simultaneous cues with sequential ones. Use count-ins, standby-and-confirm routines, and a single cue-giver. Allow a beat after questions and before entrances. In multilingual work, make speaker changes explicit so that interpreters and audience members are not forced to guess who has the floor.
3. Rehearse the worst case. A rehearsal that only works on perfect connections is not a rehearsal for networked performance. Test headphones, bandwidth, camera positions and platform settings, then practise recovering from overlap, frozen images and dropped words. The aim is not to normalise failure but to make it survivable.
4. Choose forms that tolerate delay. Monologue, direct address, split-screen conversation, epistolary exchange, pre-recorded sequences and fixed-media elements can absorb delay more readily than naturalistic overlap. This is a dramaturgical choice, not an admission of defeat.
5. Decide whether delay is visible or hidden. If latency is part of the aesthetic, make it deliberate and safe. If it is meant to be hidden, avoid exposing it through tightly timed physical comedy or rapid cross-talk. Audiences forgive a visible convention more readily than an unexplained failure.
6. Protect access and language. Interpreters, captioners, Deaf performers and audience members may be affected differently by delay. Build pauses and explicit turn-taking into the performance, and check that access workers can hear the signal they need. Latency is an access issue as well as a timing issue.
7. Document the conditions. Keep a simple technical log of platforms, locations, delay observations and cue protocols. Documentation helps a production transfer between venues and makes the rehearsal process legible to collaborators who join later.

Latency as material, not only fault

Some telematic work treats delay as compositional material: performers play with echo, dislocation and the gap between action and response. The jazz study’s finding that duos adopt diverse strategies supports this view of adaptation as creative, not merely corrective. But there is a difference between latency chosen as a constraint and latency imposed as an unequal burden. In international collaborations, bandwidth, platform access and technical support may vary between participants. A company that asks performers to ‘just cope’ with unstable connections may be distributing the cost of latency unevenly. The artistic decision to use delay should be distinguished from the practical failure to plan for it.

Terms and evidence: what can be claimed

Latency is the delay; jitter is its variation; lag is a broader colloquial term. Video-mediated interaction research, telematic performance practice and networked music studies each describe different parts of the problem. The sources cited here include video-mediated healthcare consultations, mediated-conversation research, telematic music and networked jazz. They do not establish a theatre-specific maximum acceptable delay, and readers should be sceptical of any single figure presented as universal. What the evidence does show is that delayed feedback changes turn-taking, overlap, cue response and ensemble stability. Those mechanisms are enough to justify measuring delay, re-authoring timing and rehearsing under realistic network conditions.