Using a teleprompter with a DSLR or mirrorless camera means one specific arrangement: a beam-splitter hood sized to your lens, with a phone or tablet beneath the glass running the script, and the camera shooting through it. The rig is the most involved on the ladder and produces the best-looking footage — and the part that catches people out is not the glass, it's the workflow, because recording and prompting become two separate devices with two separate jobs.
This guide covers choosing a hood for a proper camera, mounting and balancing it, the microphone question that the two-device split creates, and how to actually run takes when the record button and the script live on different machines.
Why this rig exists
At the point where you're shooting on a mirrorless or DSLR body, you've already decided image quality is part of your channel's offer — the lens choice, the depth of field, the low-light behaviour. What the beam-splitter adds is eye contact to match: the script appears directly over the lens, so reading and looking at the viewer become the same act, and the camera's quality is no longer undermined by eyes that visibly track a script beside it.
If you're still weighing this against simpler setups, the YouTube teleprompter guide walks the whole ladder; this piece assumes you've landed here.
Choosing a hood for a proper camera
Camera hoods differ from phone rigs in three ways that matter at purchase time.
Lens compatibility first. The hood mounts around or in front of your lens, so check the maximum lens diameter it accepts and the depth it allows — a long zoom may simply not fit a compact hood. Most camera hoods attach via the tripod thread on a baseplate, with the camera sliding in behind the glass; check your lens sits close enough to the glass to avoid shooting the hood's own edges at wide angles.
Glass size against reading distance. The screen behind the glass is your script surface, and its usable size is capped by the hood. A phone-sized window works close in; if you present from two to three metres, favour a hood that takes a tablet, because screen area buys font size and font size is what survives distance. The glass itself takes its usual tax — roughly a third to a half of the light in each direction — which costs your sensor up to about a stop and, more importantly for you, dims the script. Set text bigger and bolder than the bare screen suggests, and keep room light off the back of the glass. The full physics and lighting treatment is in the beam-splitter setup guide.
Weight and balance. A hood, glass, phone and camera together are front-heavy. Use a tripod head rated for the total, and expect to rebalance the plate the first time — a rig that slowly nods during a take ruins it invisibly until the edit.
Mounting order that saves faff: set the prompter software's mirror mode and font size before the phone goes in the hood, then mount, then fine-tune from your mark. Fishing a device out from under glass to reach a setting gets old immediately.
The two-device workflow
Here is the structural difference from a phone rig: your camera records, and a separate device prompts. Each job gets the best tool, and the camera is free to be a camera — any lens, any distance, proper audio inputs. The costs are a little more setup, and two machines to think about at the top of every take.
The shape that works, take by take:
- Frame and set the camera first, completely — focus, exposure, audio level — before the prompter enters your attention. The camera's settings shouldn't change again all session.
- Start the camera rolling and leave it rolling across retakes if your card and battery allow. Cutting the camera between every take doubles your trips behind the rig; one long clip with several takes inside it is cheaper to make and easy to cut later.
- Run takes from your mark, by voice. With the phone under glass and your hands in frame, the prompter needs to be operable without touching — this is where voice control stops being a convenience and becomes the workflow. Saying "Telly, stop" ends a take from where you stand; "Telly, again" starts the retake; "Telly, marker" drops a note at a stumble so you know what to check without scrubbing footage. The camera keeps rolling through all of it.
- Clap once at the top of each take. With picture on one device and your working audio on another, a single sharp clap in frame gives the edit an alignment spike. Thirty years of film practice, still the cheapest sync tool there is.
The microphone question
The two-device split creates a subtlety worth getting right on day one: there are now two listeners in the room with different jobs.
Your camera's microphone — usually an external shotgun or a lav feeding the camera — captures the audio the audience will hear. Position it for the recording, as you always would.
The prompting device's microphone is what the voice-following listens to. The script follows you because the device under the glass hears you read; a beautifully positioned shotgun feeding the camera does nothing for the prompter. In practice this is rarely a problem — you're speaking at performance volume a metre or two from the hood, which is exactly the situation device microphones handle — but it explains the one failure mode people hit: recording audio on the camera, phone buried and muted in a case, and wondering why the script won't follow. The prompting device needs to be able to hear the room.
Worth saying plainly while we're on microphones: the listening is local. Speech recognition runs on the device and the audio is never uploaded — the device under the glass is matching your words against the script, not transcribing your takes to a server. And the honest limit travels with the feature: recognition is the browser's, so quality varies by platform and language, and Firefox doesn't implement the Web Speech API at all — there, the constant and sound-level scroll modes still work, and a camera rig runs on them the same way, just with a speed to set beforehand.
Settings that earn their keep on this rig
Mirror mode, verified with a real sentence in the glass — the whole display flips, controls included.
Font and contrast for the glass, set from your mark, not the desk. Light text, near-black background; the dark background also keeps the hood's interior from glowing and washing the image.
A key light brighter than the screen's glow on your face. The prompter is a light source aimed at you; give the camera a better one.
Rotation, if your hood mounts the phone sideways and the phone's own rotation lock is on — the software's rotation control exists for this case.
A first-session checklist
- Balance the rig on the head; confirm no slow nod with everything mounted.
- Camera: frame, focus, exposure, audio level. Lock it.
- Prompter: mirror on, rotation if needed, font sized from your mark, dark background.
- Hood: no room light striking the back of the glass; wide shot checked for hood edges.
- Sound check both listeners — a sentence at performance volume; the highlight should follow it.
- Roll camera, clap, and run half the script as a rehearsal take before the one that counts.
The promise of this rig is worth its setup: your best camera, unbroken eye contact, and takes you run entirely from your mark. When the balance is set, both microphones can hear what they need to and the retake is one spoken word away, the machinery stops being the story — which leaves the footage, and you.