The camera follows whoever speaks.
Automatically, and without losing them. All 7 emirates.
Presenters move. Speakers change. A tracking camera keeps the right face framed without anyone touching a remote — if you buy the right one of the four technologies hiding behind the words "AI tracking" on every box. Vector Digital Systems stocks them, installs them, and tunes the tracking in your actual room before handover. Since 2009, from Dubai stock.
- The right technology of the four sold as "tracking"
- Optical zoom that reaches your back row
- Blocking zones so it never chases the doorway
- Ceiling-mic pairing for whoever speaks, anywhere
- Tracking tuned in your room, with your people
- Proven on a live call before we leave
Every box says "AI tracking" — four different technologies hide behind those words
Search for a tracking camera and every result promises the same thing. The camera follows the speaker. AI-powered. Automatic. The boxes look identical from the outside.
Inside those boxes are four completely different technologies. One follows a body with computer vision. One aims the lens wherever sound comes from. One watches the whole room with a second eye and frames with the first. And one — the cheapest and loudest-marketed — simply crops a wide webcam picture and calls the crop "tracking." Buy the wrong one and it fails in ways the datasheet never mentioned: the classroom camera that freezes when a student walks past the teacher, the boardroom camera that swings toward the air-conditioning vent, the "tracking" webcam whose zoomed face turns to fog past four metres.
This page is the decoder. The four technologies with their honest failure modes, every brand's tracking names translated into plain language, the tuning that separates professional installs from restless ones, and the one-photo shortcut to the right answer for your room.
The four technologies sold as “tracking” — and the one that will disappoint you
1. Vision body-lock tracking (AED 4,000–15,000). The camera's own AI detects a person and physically pans, tilts and zooms to keep them framed — no wearable, no lanyard, no operator. This is presenter tracking in its true form: the technology behind education and lecture cameras, where a teacher roams from board to bench and the lens follows. Its honest failure mode is crowds and crossings — a second body walking through the frame can steal or freeze the lock — which is exactly what professional-grade multi-presenter detection and the blocking zones explained below were built to solve.
2. Audio-localisation tracking (built into cameras, or added through microphone pairing). Beamforming microphones hear where a voice is coming from and aim the camera at that seat. This is speaker tracking in its true form: nobody moves much, but the talker keeps changing — the boardroom pattern. Its failure mode is sound that isn't speech: a cough, a closing door, a projector fan — all solved by thresholds and hold-times tuned on site, never by the out-of-box defaults.
3. Dual-lens detect-and-frame (AED 5,500–9,000). Two cameras in one body: a wide panoramic eye that watches everyone constantly, and a PTZ eye with real optical zoom that shoots the chosen frame. Because the wide eye never stops watching, the switch to a new speaker is fast and the room never disappears off-screen during a move. The Yealink UVC86 class — dual 4K eyes, 12x optical, six beamforming microphones — defines this tier, and it is the camera we install most in serious Dubai boardrooms.
4. EPTZ digital-crop "tracking" (AED 300–1,500) — the trap. A fixed wide webcam with no moving parts crops a rectangle around a face and moves the crop. On the box it reads "AI tracking, 32x zoom." In the room, the crop is digital magnification of the same small sensor — and a face at five metres becomes a smear of pixels. Honest use: one person at a desk, huddle spaces under three metres. Sold into classrooms and boardrooms, it is the technology behind every "we bought a tracking camera and it's useless" call we receive.
Ranges are supply-only for the camera class; installation, mounting and on-site tuning are quoted per room. The expensive mistake is not choosing between tiers one and three — it is being sold tier four for a tier-one job.
Yealink, Poly, Logitech, AVer, Neat — what each brand's tracking names actually mean
Every manufacturer invented its own vocabulary for the same four technologies, which is why comparing datasheets feels impossible. This table is the translation no brand will print.
| Brand | Name on the datasheet | What it actually is |
|---|---|---|
| Yealink | Speaker Tracking / Presenter Tracking / Auto Framing | Audio-localisation close-ups of whoever talks · vision body-lock on one mover · group framing (not tracking). The UVC86 dual-eye carries all three. |
| HP Poly | DirectorAI — Speaker Framing, Presenter Tracking, Group Framing | A production-style director inside the camera choosing between talker close-ups, a locked mover, and everyone-in-frame — the E70's twin lenses feed it. |
| Logitech | RightSight 2 — Speaker View & Group View | Dual-view intelligence on Rally systems: a talker close-up beside a whole-room view, composed automatically. |
| AVer | Auto Tracking / Wide Area, Stage & Segment Tracking / PTZ Link | True vision body-lock with multi-presenter detection and up to 8 blocking zones on the pro education line — plus PTZ Link, which lets ceiling microphones aim the camera. |
| Neat | Neat Symmetry | Framing intelligence: every in-room person presented clearly to remote viewers — closer to smart framing than to tracking, and honest about it. |
| Microsoft Teams | IntelliFrame | Platform-side AI composing in-room faces into tiles on certified cameras — it rides on top of the camera's own tracking, gated by the Teams Rooms Pro licence. |
Two honest observations from the table. First: "framing" appears everywhere because fitting everyone in frame is easier than following one person — know which one you are buying. Second: the names change every generation; the four underlying technologies have not. Learn the four, and every datasheet becomes readable.
Auto-framing is not tracking — here’s the difference you’re paying for
Framing fits everyone into the picture. Tracking follows one person through the room. They sound like cousins; they solve opposite problems.
Auto-framing counts the people at the table and crops the shot so nobody is wasted space — brilliant for a static meeting, and standard on every decent video bar we sell. But it has no idea your trainer exists the moment she leaves the table for the whiteboard: framing keeps the table perfectly composed while the person actually teaching walks out of relevance. Tracking is the opposite commitment — one subject, followed — and it is the difference between a meeting camera and a teaching camera.
The buying rule in one line: if the important person moves, you need tracking; if the important people sit, framing is enough — and framing is cheaper. A supplier who asks "who moves in your room?" before quoting is applying this rule for you. A supplier who doesn't is selling whatever is on the shelf.
Speaker tracking or presenter tracking? Pick by who moves in your room
The two names get swapped constantly, including by people selling the cameras. The distinction takes one paragraph and saves thousands of dirhams.
Speaker tracking follows the conversation. Ten people around a boardroom table, nobody walking, but the voice moving from chair to chair — the camera's job is to cut its attention to whoever talks. That is audio-localisation work: microphones decide, lens obeys. Presenter tracking follows a body. One instructor, moving constantly — board, bench, front row — while everyone else sits. That is vision body-lock work: the AI holds one subject and ignores the rest.
Rooms that need both exist, and that is what dual-lens cameras and mode switches are for: a lecture theatre runs presenter tracking during teaching and flips to speaker tracking for panel discussions. But most rooms are honestly one or the other, and naming yours correctly is half the specification done. The other half — reach — is next.
Why 32x digital zoom loses to 12x optical — the spec that decides back-row faces
Tracking gets the camera pointed at the right person. Zoom decides whether that person has a face when the camera arrives. The two specs are sold as one number, and they could not be more different.
Optical zoom magnifies through glass before the image is captured: a 12x optical lens at ten metres delivers a genuinely sharp head-and-shoulders frame. Digital zoom crops pixels that were already captured: a "32x" digital figure on a fixed webcam means the same small sensor stretched until faces fog. Our large meeting rooms page carries the full reach arithmetic; for tracking cameras the practical line is this — past four or five metres of subject distance, only optical zoom keeps a trackable face worth tracking. It is why the trap-tier webcams fail in real classrooms, and why every camera in our tiers one and three carries true optical glass.
Blocking zones: how professionals stop the camera chasing doorways
Out of the box, a tracking camera is an enthusiastic puppy. Someone enters through the back door — it looks. A silhouette crosses the corridor glass — it follows. The projector screen flickers — it considers the screen a person. None of this is a faulty camera; it is an untuned one.
Professional tracking cameras solve it with blocking zones — regions of the image the tracking engine is told to ignore. The pro education class supports up to eight per camera, and our commissioning draws them deliberately: the doorway, the corridor glazing, the display itself (so the camera never tracks the presenter's own projected face — a genuinely common failure), and the seating block when the mode is presenter-only. Alongside zones sit the timing values from our multi-camera page — hold-time so the frame doesn't twitch, sensitivity so a passing student doesn't steal the lock. These numbers do not exist on datasheets. They are set in your room, with your people walking the actual paths, during staging — which is why "installation" from us includes an afternoon of tuning, and why a camera ordered online and screwed to a wall behaves like a puppy for its whole life.
Pair the camera with your ceiling microphones — it tracks whoever speaks, anywhere
Here is the capability almost nobody selling cameras mentions: the ceiling microphones you already own can aim the camera. Beamforming ceiling arrays know where in the room a voice is coming from. Integration software — AVer's PTZ Link is the reference, pairing its cameras with microphones from Shure, Sennheiser, Nureva and Yamaha — turns that location into a camera command: a voice from the third row, and the lens is already framing the third row.
This is the technology behind the audience Q&A experience on our multi-camera rooms page: the questioner captured by the ceiling and framed by the camera, no handheld passed down the row, no operator. In single-camera rooms it turns one good PTZ into a speaker-tracking system for the whole space; in lecture halls it lets a body-lock camera own the presenter while a second, mic-aimed camera owns the audience. If your room already carries Shure ceiling audio, the tracking upgrade is smaller than you think — the intelligence is largely in the pairing, and the pairing is configuration, cabling and tuning we do in a day.
Classroom, boardroom, townhall, auditorium — the right tracking type for each
The classroom and training room (vision body-lock). One instructor, constant movement, recording as important as the live call. The AVer education tracking line built its reputation here — no wearable, segment modes that capture the board as the teacher crosses it, and streaming outputs for the learning library. Our AVer DL30 auto tracking camera has served exactly this brief in UAE classrooms for years, and the current education line carries the same logic further.
The boardroom (dual-lens speaker tracking). Nobody walks; the voice moves. The UVC86 class — panoramic eye watching, optical eye framing, beamforming ears deciding — delivers the talker close-up without the room ever vanishing during a switch. This is the default serious answer for Dubai boardrooms and the camera behind most of our tier-three installs.
The townhall (tracking plus directing). A roaming presenter needs body-lock; audience Q&A needs mic-aimed framing; and once there are two cameras, something must choose between them — which is the multi-camera directing story, one page over. Tracking cameras are the actors; that page is the director.
The auditorium, worship hall and recorded venue (tracking plus NDI). Long throw distances demand serious optical zoom, and the output feeds streams and recordings as much as calls — the NDI-equipped tracking PTZ class, exported from our Dubai stock to venues across Africa and Asia far more often than most buyers expect.
Will it work with Teams and Zoom? Tracking lives in the camera — so yes, with one licence note
The question every buyer asks first has the friendliest answer on the page. Tracking happens inside the camera, before the video ever reaches the software — so to Teams, Zoom, Google Meet or a guest's laptop, a tracking camera is simply a camera whose picture happens to keep finding the speaker. No plugin, no special mode, no compatibility matrix. The one note worth knowing sits on the Microsoft side: Teams Rooms can layer its own IntelliFrame intelligence on certified cameras, composing in-room faces into individual tiles — a genuinely nice addition that requires the Teams Rooms Pro licence, exactly as our multi-camera page documents from Microsoft's own 2026 guidance. The camera's tracking works either way; the platform AI is a bonus tier, not a dependency.
Why a 2,000-dirham "tracking" camera costs you the meeting — and what honest quotes include
The price gap in this category is brutal: three hundred dirhams to fifteen thousand, all wearing the same two words. The gap is the four technologies — and one more thing quotes hide.
The hidden line is tuning. A tracking camera that has never had its zones drawn, its sensitivity set and its hold-time tuned in the actual room will chase doorways, freeze on crossings and twitch between targets — and the buyer will blame the camera. Our quotations therefore state five things in writing: the technology tier by name (so you know which of the four you are buying), the optical zoom figure and the room distance it must serve, mounting and cabling included, an on-site tuning session with your real presenters walking their real paths, and a live-call demonstration before sign-off. A quote missing the tuning line is not cheaper — it is unfinished, and the missing afternoon returns as a camera nobody trusts.
Send one photo and one sentence — get the right tracking camera the same day
This category does not need a long survey to reach a correct shortlist. It needs two facts. A photo of the room taken from where the camera would mount, looking at the space it must cover. And one sentence answering the only question that matters: who moves — the presenter, the speakers, or both? Add the room depth in metres if you know it. That is enough for a technology tier, a specific camera recommendation and a budget the same day, 8am–6pm Mon–Sat. Rooms heading into multi-camera or mic-paired territory get the full survey from there — camera positions, zone maps and cable routes drawn before anything is ordered.
Tracking cameras, exported from Dubai
Vector Digital Systems exports speaker and presenter tracking cameras — education body-lock models, dual-lens boardroom units, NDI streaming PTZs and full mic-paired kits with tuning documentation — from Dubai stock to 45 countries across the Middle East, Africa, South Asia, Central Asia and the Caucasus. Universities, worship venues, training networks and integrators receive pre-configured cameras with zone and threshold notes that turn on-site commissioning into hours; air freight typically lands in 3–7 days. For export enquiries, WhatsApp our export desk.
What tracking-camera clients say
"Our first 'tracking' camera was a cheap webcam that turned our trainer into pixels past the second row. Vector explained the four types in ten minutes, put a proper body-lock camera in, and spent an afternoon tuning zones so it ignores the corridor door. It has not lost her since."
"The dual-lens camera they recommended cuts to whoever is speaking around our table without the room ever disappearing. What sold us was the demo in our own boardroom before we paid — they tuned the hold time live until it felt natural."
"We ordered six NDI tracking cameras for our lecture halls in Accra through their export desk. They arrived with the tracking zones pre-drawn from our floor plans and notes for our AV team. Commissioning took a morning per hall instead of the week we budgeted."
Speaker tracking cameras — FAQ
Do tracking cameras work with Microsoft Teams and Zoom?
Yes — tracking happens inside the camera, before the video reaches any software, so Teams, Zoom, Google Meet and a guest's laptop all simply receive a picture that keeps finding the speaker. No plugin or special mode is required. On Microsoft Teams Rooms, certified cameras can additionally use IntelliFrame to present in-room faces as individual tiles, which requires the Teams Rooms Pro licence — a bonus layer on top of the camera's own tracking, not a dependency.
What is the difference between auto-framing and tracking?
Framing fits everyone into the picture; tracking follows one person through the room. Auto-framing counts the people at the table and composes the shot — perfect for static meetings — but it has no concept of your trainer once she leaves the table. Tracking commits to a subject and moves with them. The buying rule: if the important person moves, you need tracking; if the important people sit, framing is enough and costs less.
What is the difference between speaker tracking and presenter tracking?
Speaker tracking follows the conversation: nobody walks, but the voice moves from chair to chair, and microphones aim the camera at whoever talks — the boardroom pattern. Presenter tracking follows a body: one instructor moving constantly while everyone else sits, held by the camera's vision AI — the classroom pattern. Dual-lens cameras and mode switches cover rooms that need both, but naming which one your room truly is completes half the specification.
Does the presenter need to wear a lanyard, badge or tracker?
No — current professional tracking is entirely camera-side. Vision body-lock cameras detect and follow a person with no wearable at all, and multi-presenter detection handles a second person entering the teaching area. Wearable-based systems existed a generation ago and still appear in old tender templates; if a specification in front of you demands a lanyard transmitter, it is describing obsolete technology and worth updating before you buy.
Why does our tracking camera keep losing the presenter at the whiteboard?
Three usual causes, all fixable. Backlight: a bright window or the projected screen behind the presenter starves the AI of a clear subject — solved by camera position or exposure tuning. Zone confusion: the camera is tracking the presenter's own projected face on the display — solved by drawing a blocking zone over the screen. Crossings: students or colleagues passing between camera and presenter steal the lock — solved by sensitivity and multi-presenter settings. All three are commissioning work, not hardware faults, and an afternoon of tuning usually retires the complaint permanently.
What are blocking zones and why do they matter?
Blocking zones are regions of the image the tracking engine is told to ignore — doorways, corridor glazing, the display itself, the audience block in presenter-only mode. Professional education cameras support up to eight per camera. Without them, tracking chases every person who enters and every silhouette that passes; with them, the camera holds its subject through a busy room. Zones do not exist on datasheets — they are drawn in your room during commissioning, which is why tuning belongs on every honest quotation.
Can our ceiling microphones control the camera?
Very often, yes — and it is the most under-sold capability in the category. Beamforming ceiling arrays know where a voice comes from, and pairing software such as AVer's PTZ Link converts that position into camera aiming, with supported microphones from Shure, Sennheiser, Nureva and Yamaha. The result: whoever speaks, anywhere in the room, is framed — no handheld passed down the row. If your room already carries a ceiling array, the upgrade is mostly configuration, cabling and a day of tuning rather than new infrastructure.
Is 20x or 32x digital zoom better than 12x optical zoom?
No — the numbers measure different things. Optical zoom magnifies through the lens before capture and keeps faces sharp at ten metres and beyond. Digital zoom crops already-captured pixels, so a large digital figure on a fixed webcam delivers fog past four or five metres. For any room deeper than a huddle space, treat the optical figure as the real specification and the digital figure as marketing.
Can a tracking camera record and live-stream lessons as well as run calls?
Yes — and for education it should. The professional tracking class outputs USB for calls plus HDMI and network streams, and NDI-equipped models feed recording and streaming systems natively, so the same camera that follows the teacher fills the learning library. Segment-style modes capture board content as the presenter crosses it. If recording matters, say so in the first message — it moves the recommendation toward the education line and its streaming outputs.
How many people can one tracking camera follow?
Body-lock tracking follows one subject at a time by design — that is its job. Professional models add multi-presenter detection so a second person entering the teaching area is handled gracefully rather than breaking the lock, and dual-lens cameras keep the whole room visible while the optical eye frames the chosen subject. When several people must each be framed as they speak, that is speaker tracking through microphones — or a multi-camera design, which is its own page.
What does a proper tracking camera cost in Dubai?
Supply pricing by technology: education vision body-lock cameras run roughly AED 4,000–15,000 depending on zoom and streaming outputs; dual-lens boardroom cameras roughly AED 5,500–9,000; EPTZ digital-crop webcams cost AED 300–1,500 and honestly serve only desks and sub-three-metre huddles. Installation, mounting and the on-site tuning session are quoted per room and belong on every serious quotation — a tracking camera without tuning behaves like an untrained puppy regardless of price.
Can we add tracking to the video bar we already own?
Frequently, yes. Most current video bars already include auto-framing and some include speaker view; what they lack is presenter tracking for a moving instructor. The common retrofit keeps the bar as the room's framing camera and adds one body-lock tracking camera for the presenter — either switched by the platform, or directed as a two-camera setup as our multi-camera page explains. Send a photo of the existing bar and the room; compatibility is usually answerable the same day.
What should we send on WhatsApp for a same-day recommendation?
Two things: a photo of the room taken from where the camera would mount, looking at the area it must cover, and one sentence answering who moves — the presenter, the speakers, or both. Add room depth in metres if known, and mention recording if it matters. That is enough for a technology tier, a specific model and a budget the same day, 8am–6pm Mon–Sat.
Which areas of the UAE do you cover for tracking camera installations?
Vector Digital Systems serves businesses across all seven UAE emirates — Dubai, Abu Dhabi, Sharjah, Ajman, Ras Al Khaimah, Fujairah and Umm Al Quwain — supplying, installing and tuning speaker and presenter tracking cameras. Within Dubai, active client sites include Business Bay, DIFC, Jumeirah Lake Towers (DMCC), Barsha Heights (TECOM), Dubai Internet City, Dubai Media City, Dubai Silicon Oasis, Downtown Dubai, Sheikh Zayed Road, Jebel Ali Free Zone (JAFZA), Dubai Investment Park (DIP), Dubai World Trade Centre, Dubai Healthcare City, Al Quoz, Deira, Bur Dubai, Al Barsha, Dubai Marina and Dubai South. Across the Northern Emirates and Abu Dhabi: Musaffah Industrial City (ICAD), Khalifa Industrial Zone (KIZAD), Masdar City, Al Ain, SAIF Zone (Sharjah), Hamriyah Free Zone, Sharjah Industrial Area, RAKEZ, Fujairah Free Zone and Ajman Free Zone. On-site installation, configuration and fault response are available across all emirates; remote support is available UAE-wide.
Do you export tracking cameras from Dubai?
Yes. We export speaker and presenter tracking cameras from Dubai stock to 45 countries: Saudi Arabia, Oman, Qatar, Kuwait, Bahrain, Iraq, Jordan, Kenya, Ethiopia, Tanzania, Uganda, Rwanda, Burundi, Djibouti, South Africa, Mozambique, Zimbabwe, Zambia, Mauritius, Seychelles, Nigeria, Ghana, Senegal, Côte d'Ivoire, Cameroon, Angola, Togo, Benin, Congo DR, Gabon, Egypt, Morocco, Algeria, Tunisia, Pakistan, Bangladesh, Sri Lanka, Kazakhstan, Uzbekistan, Turkmenistan, Kyrgyzstan, Tajikistan, Georgia, Azerbaijan and Armenia — pre-configured with zone and threshold notes, full export documentation, air freight in 3–7 days.