The Complete Talking-Head Filming Setup: Phone, Light, Sound, Script and Eye Line, With No Studio

Most talking-head videos fail before the first word is spoken. The phone is too low, the window is behind the speaker, the microphone is on the other side of the room, and the script is on a laptop a metre to the left so every sentence ends with a glance away. None of that is fixed by a better phone. It is fixed by a setup you build once, write down, and rebuild the same way every time you record.

This is that setup, in the order you should build it: position, light, sound, framing, then script. It assumes an iPhone, a desk or table, and a budget of roughly nothing to a small tripod and a light. Earlier posts in this series went deep on single pieces, such as keeping eye contact with the camera and mirror mode and DIY rigs; this one is the whole picture, so you can get from nothing to a repeatable recording corner in an afternoon.

Step 1: Pick the spot by the light, not the background

Walk around your home in the daytime and find the place where a window would be in front of you while you talk, slightly to one side. That is your spot. Everything else, including the background, is secondary and can be fixed; bad light cannot. A window behind you turns you into a silhouette and no setting on the phone rescues it. A window directly beside you gives a half-lit face, which can look dramatic and usually just looks like a mistake.

If daylight is unreliable or you record at night, one affordable LED panel or ring light placed where the window would be does the same job. Put it slightly above eye level, angled down, and about an arm's length further away than feels natural; closer light is softer but shows every pore and makes you squint. Turn off the overhead room light, which casts shadows under the eyes and fights the colour of your main light.

Step 2: Put the lens at eye level, not where the tripod ends

The single most common amateur tell is a camera looking up at the speaker from desk height. Stack books, use a tall tripod, or mount the phone on a shelf; whatever it takes, the lens should be at your eye level or a finger's width above. If you are sitting, that is roughly the height of your eyebrows when you sit up straight, and you will slump, so aim a touch high.

Distance matters too. Arm's length is the minimum for a phone; closer and the wide lens stretches your features. A little further back, with the phone zoomed slightly or using the 2x lens if your model has one, gives a flatter, more flattering look and makes it easier to keep the script near the lens, which matters in step 5.

Step 3: Sound is half the video

Viewers forgive soft focus and odd colours; they do not forgive echo. Three things fix most audio problems in a normal room, in order of cost: record in the smallest, softest room you have (a bedroom with a bed and curtains beats a kitchen every time), get the microphone close (any wired or wireless lavalier clipped at the sternum will outperform the phone's built-in mic from a metre away), and put something soft behind the phone so your voice does not bounce straight back off a wall into the mic.

Do a ten-second test take and listen on headphones before every session. Fridges, laptop fans and the neighbour's lawnmower are all invisible until you are editing.

Step 4: Frame once for every platform

Record in portrait. Vertical video is what Reels, Shorts and TikTok want and a portrait master can be cropped to a square for feeds; a landscape master cannot be turned into a usable vertical without cutting off your shoulders. Place your eyes on the upper third line, leave a hand's width above your head, and keep the bottom quarter of the frame reasonably clear because captions and interface buttons will sit there on every platform.

Lock exposure and focus before you start (tap and hold on the subject in the camera app) so the image does not pulse when you move your hands. If you want the detail that lets you crop later, Teleprompter: Camera Overlay records in portrait 4K, which gives you room to punch in for a tighter shot in the edit without the result turning to mush.

• Portrait, eyes on the upper third, bottom quarter clear for captions.
• Exposure and focus locked; no auto-adjust pulsing.
• A background with depth (a room behind you) rather than a flat wall right behind your head.
• One or two objects in the background that are yours: a plant, a shelf, a lamp. Not a shrine, not a bare wall.

Step 5: Put the script where your eyes already are

This is the step that separates talking-head videos that feel like a conversation from ones that feel like a hostage statement. Any script that lives away from the lens, on a laptop, a second phone or a printout, pulls your eyes off camera every few seconds, and viewers notice even when they cannot say why. The fix is to put the words as close to the lens as physically possible.

On a phone you have two practical options. The first is a camera-overlay teleprompter: the script floats over the camera preview on the same screen, right next to the front lens, so reading and looking at the camera are the same movement. Teleprompter: Camera Overlay does exactly this, with the text positioned beside the lens and a choice of scroll modes. The second is a physical beam-splitter rig that reflects a second screen in front of the lens; it is better for long, formal pieces and worse for everything else because of the setup time, which is why the rig post recommends it only for people recording hours of material.

Whichever you use, set the text large enough that you read in phrases rather than words, and narrow the column so your eyes do not visibly track left to right. Use voice-follow scrolling if your app has it, so the script moves at your speaking pace and waits when you pause to think; the voice-follow vs auto-scroll post explains when the classic timed scroll is still the better choice. In the overlay app the script is only on the preview, never in the saved video, so there is nothing to crop out afterwards.

Step 6: The two-minute pre-flight checklist

Write this on a sticky note and put it on the tripod. It is the difference between a setup and a habit.

• Light on, overhead off, blinds adjusted if daylight has moved.
• Lens at eye level, phone level (check the horizon of a shelf behind you).
• Mic clipped, headphones test take done, phone on Do Not Disturb and airplane mode so a call does not kill the recording.
• Battery above 40 percent or plugged in; storage checked (4K eats space).
• Exposure and focus locked. Lens wiped with a cloth. This one fixes more soft videos than any setting.
• Script loaded, first line visible, text size and scroll mode set from the last session.
• Water within reach, out of frame.

When NOT to bother with all this

If you are replying to a comment, filming a quick reaction, or recording a behind-the-scenes clip where the point is that it is unpolished, do not build the set. Hold the phone, talk, post. Polish on the wrong content reads as trying too hard. The full setup is for the videos that carry your ideas, your course material, your sales pitch or your explanation of something you know well, where the viewer's attention is the thing you are spending and you want them to spend it on the words rather than on the shadow under your nose. Those videos are also the ones where a script earns its keep, so they are the ones where step 5 pays off most.

One last honest note: the setup makes a good speaker look professional, it does not make a nervous speaker relaxed. If the camera itself is the problem, the fixes are repetition and script craft, not equipment; the writing a script that doesn't sound written post is the better starting point, and a later post in this series will deal with camera anxiety directly.

Next in the series

With the room sorted, the next question people ask is about the app itself, specifically what it does with your voice when voice-follow scrolling is on. The next post looks at why on-device speech recognition matters for privacy, what actually happens to the audio when a teleprompter listens to you, and how to tell whether an app is processing speech on the phone or sending it somewhere else.