How to Automate Your Presentation Livestream

Presentation livestreams look simple until you operate one for 3 hours.

A “boring” talk typically needs:

• a clean speaker shot (ideally tracked)
• slides when they matter
• a wide stage shot as safety
• audience cutaways (optional but improves production value)
• PiP layouts (speaker + slides)

Doing this manually is repetitive and error-prone, especially with small crews or long programs.

This guide shows how to build an automated, operator-friendly workflow using PTZ cameras and switchers like Blackmagic ATEM or software mixers like vMix, with an emphasis on production-quality rules: clean cuts, no live PTZ moves, and predictable behavior.

What “automation” should achieve in presentation workflows

Baseline automation:

• Track the presenter (framing stays correct)
• Switch between speaker / slides / wide / PiP at the right moments
• Use audio to detect “who is speaking” (single or multi-speaker)

Advanced goals:

• Detect slide changes and react
• Detect applause/laughter and cut to audience
• Multi-speaker panel automation using per-speaker mics + presets

Step-by-step: Automate a presentation livestream

Step 1 - Define your shots

A reliable automation workflow starts with a stable shot plan:

Recommended minimal set

• Stage (wide): static safe shot over the whole scene (required)
• Speaker: PTZ that follows the presenter
• Presentation: direct feed of slides (HDMI/SDI capture or NDI®)
• Optional: Audience camera for reactions
• Optional: PiP input (speaker + slides combined)

Select your cameras: Depending on your budget constraings and requirements, you can use a wide range of PTZ cameras by Telycam, Panasonic, Sony, Canon, Marshall, BirdDog or Obsbot.

Step 2 - Build low-latency video paths

Tracking and good switching depend on latency.

Common input paths:

• Capture card (SDI/HDMI) → usually the lowest latency
• NDI → moderate latency (often acceptable) - many PTZ cameras support NDI natively
• MJPEG streams → moderate latency (often acceptable) - almost all PTZ cameras support this
• SRT and RTSP → high latency - useful for sending video feeds over the internet but not for tracking

Step 3 - Slide capture that doesn’t break your show

Slides are often the source of last-minute chaos.

Options:

• HDMI out of the presentation machine → capture → switcher
• NDI® output from the presentation machine → receiver → switcher/automation

Rules:

• Use a stable resolution
• Avoid OS popups / notifications

Step 4 - Audio routing (single vs multi-speaker)

Single-speaker talk

• One clean mic feed + optional room mic

Multi-speaker panel / town hall

• You’ll get better automation with:
  • one mic channel per speaker (or at least per speaking position)
  • plus optional room/audience mic

Why it matters:

• Active-speaker logic requires usually isolated channels.

Step 5 - Choose your switcher control strategy

You need a switcher or software that can be driven by automation.

Typical choices:

• Blackmagic ATEM (hardware)
• vMix (software)
• Ross / Roland / OBS

Your automation layer must know:

• input mapping
• preview/program states

Step 6 - Decide your automation logic

There are three common models:

Model A: Simple schedule + manual override

• Start with wide
• Show slides at fixed segments
• Return to speaker
• Manual override for special moments

Good for:

• scripted / rehearsed talks

This can be easily implemented with most switchers using macros or plugins (for software mixers). However, it might not be flexible enough for dynamic presentations.

Model B: Audio-driven + rules

• Speaker mic active → show speaker
• Silence → wide
• Audience mic active → audience
• Optional: slide change triggers slide shots

Good for:

• most real-world talks

When implemented well, this gets you 80% of the way there, although this will require extensive programming skills to get right.

Model C: AI-driven auto-cut (best for long “boring” sessions)

• Uses audio + video AI analysis to decide:
  • when slides matter
  • when to show speaker
  • when to show wide / audience
• Coordinates PTZ moves with switcher states

Good for:

• presentations, panel discussion, long conferences, universities, town halls

This is the class of workflow MiruSuite is designed for: automated, high-quality production with minimal operator intervention. MiruSuite is a ready-made solution that can be integrated into existing AV workflows, without the need for custom programming.

FAQ

“How do I automate lecture streaming without camera operators?”

You need:

• at least one tracking-capable speaker camera (PTZ + tracking)
• a slide feed
• a switcher that can be controlled
• an automation layer like MiruSuite that handles switching and prevents live camera moves

Start with: wide + speaker + slides.

“How do I auto-switch between slides and speaker?”

You either:

• drive switching with rules (audio activity + slide change detection)
• or use an AI auto-cut layer that analyses the slide feed and decides when it’s “interesting” to show slides.

“Can I automate multi-speaker panels?”

Yes, if you have:

• mic channels per speaker (or per position)
• presets for each speaker from at least one camera (two cameras is better)
• logic that avoids flip-flops during crosstalk

Where MiruSuite fits

If you’re producing talks that are:

• long
• repetitive
• high volume (universities, conferences, internal town halls)

MiruSuite is designed for exactly these formats:

• automated person tracking
• automated live cutting (“AutoCut”) for presentation scenarios
• slide-aware decisions
• switcher integration so cameras don’t move on program
• manual override when needed

MiruSuite supports common PTZ ecosystems (Panasonic, Sony, Canon, Marshall, BirdDog, Telycam) and typical switching stacks (Blackmagic ATEM, Ross Video, vMix, OBS, Roland AV).