Sound First: Shooting Video Without a Studio

Most small businesses that start making video buy the wrong thing first. They buy a camera, or a light ring the size of a tractor tyre, or a set of coloured LED bars. Then they shoot in a room with a ceiling fan and a tiled floor, and the result is unwatchable — not because it looks bad, but because it sounds like a phone call from a stairwell.

The order that works is the opposite of the order things are sold in. Fix sound. Then fix light. Then worry about the camera, which in 2025 is almost certainly already good enough in your pocket.

Why sound decides whether anyone watches

People forgive a soft, slightly grainy picture. They will watch a shaky clip of something interesting all the way to the end. What they will not do is stay with audio that makes them work — echo, hiss, a fan hum under every word, a voice that fades when the speaker turns their head.

There is a physical reason bad audio is so common, and it is worth understanding because it tells you what to fix. Sound pressure falls off with distance in a predictable way: double the distance between mouth and microphone and you receive roughly a quarter of the sound energy — about a 6 decibel drop. Meanwhile the room noise, the fan, the traffic outside, the compressor of the split AC in the next room, stays exactly the same. So the ratio of voice to noise collapses as you back away.

Your phone held at arm's length puts the microphone perhaps 60 centimetres from your mouth. A clip-on microphone on your collar sits at around 20 centimetres. That is roughly three times closer, and it changes the recording completely — not because the clip-on microphone is a better microphone, but because of where it sits.

The second enemy is the room itself. Hard surfaces — tile, marble, glass partitions, a steel shutter, an empty conference table — reflect sound back at the microphone slightly late. Those reflections are what makes a recording sound "boxy" or "like a bathroom". A close microphone helps here too, because the direct voice arrives loud and the reflections stay quiet.

Do the clap test before you shoot anywhere. Stand in the room, clap once, hard, and listen. If you hear a short ring or tail after the clap, that ring will be in your video. Move to a smaller room with soft things in it — curtains, a sofa, a bookshelf, a cupboard full of files. A bedroom almost always sounds better than a boardroom.

The audio checklist that actually matters

• Microphone within 20–30 cm of the mouth. Clipped to the collar, just below the chest line, pointing up.
• Kill the fan and the AC for the take. Yes, it will get warm. Shoot in shorter blocks. A constant hum can be reduced in editing but never removed cleanly — you always hear the voice thin out with it.
• Record a few seconds of silence in the room before you start speaking. Editors use that to sample and subtract the room noise.
• Watch your levels. If your recording app shows a meter, aim for peaks around −12 to −6 dBFS (decibels relative to full scale — the top of the meter is 0). Anything touching 0 is clipped and permanently distorted. Too quiet is recoverable; clipped is not.
• Headphones on the person monitoring. Not to be fancy. Half of all ruined shoots are a cable brushing a kurta, and you cannot see that on a screen.

What to buy

A simple wired lavalier that plugs into your phone costs very little and solves ninety per cent of the problem. The main annoyances are the cable and the adaptor you will need if your phone has no headphone socket. A wireless clip-on pair — a small transmitter and a receiver that plugs into the phone — costs more but removes the cable and lets two people speak. Street prices move constantly, so check before you buy; the point is that this is the cheapest part of the whole exercise and the one that changes the result most.

What I would not buy: a shotgun microphone mounted on top of the camera. It is sold as the "professional" option and it is the wrong tool for a talking-head video shot in a small Indian office. Mounted on the camera it is still a metre away from the speaker, so it picks up the same room and the same fan — just more directionally. Shotgun microphones earn their keep on a boom pole, held just out of frame above the speaker's head, by a person whose job is to hold it. If nobody on your team is going to do that, skip it.

Also skip, for now, the large studio condenser microphone on a desk arm. It sounds wonderful for a podcast where the microphone is allowed to be in shot. For video where you want a clean frame, it is in the way.

Light second, and mostly free

Light is the second thing because it is the second most noticeable, and because the best light source available to an Indian office is free and already installed: a window.

The rule is simple. Face the window. Sit so the daylight falls on your face, with the camera between you and the window, or just beside it. Do not sit with the window behind you — the camera will expose for the bright background and turn you into a silhouette. If you must have a window behind, pull the curtain.

Avoid direct midday sun through glass; it is harsh and it moves. Softer, indirect daylight from a north-facing window or a shaded side of the building is more flattering and more stable across a two-hour shoot. If the light is too hard, tape a plain white bedsheet over the window — that is a diffuser, and it is the same thing a rented softbox does.

Three specific problems worth knowing about:

1. Overhead tubelights make raccoon eyes. A tubelight or downlight directly above the head throws shadows into the eye sockets and under the nose. Either switch them off and use the window, or add a light from the front at roughly face height to fill those shadows. A cheap LED panel on a stand, or even a second phone with its torch bounced off a white wall, works.
2. Mixed colour temperature looks wrong and is hard to fix. Daylight is bluish, around 5600 kelvin. An old warm tube or bulb is around 2700–3000 K. When both hit your face, one side goes orange and the other goes blue, and no white-balance setting can correct both. Pick one source and switch the other off.
3. Cheap LEDs flicker on 50 Hz mains. India runs at 50 hertz, and many inexpensive LED panels and tubes pulse with it. If you shoot at a mismatched frame rate or shutter speed you get rolling dark bands across the picture. Shooting at 25 or 50 frames per second, with the shutter at 1/50 where your app lets you set it, avoids most of this. Always shoot ten test seconds and play it back at full screen before you commit to a take.

What I would not buy: a full three-point lighting kit with stands, for a team that will shoot once a fortnight. Setting it up honestly takes twenty minutes, and the kit that takes twenty minutes is the kit that stays in the box. One soft key light you can switch on in ten seconds, next to a window, beats three lights you never assemble.

Camera last, and probably already in your pocket

A recent mid-range phone shoots video that is entirely good enough for Instagram, YouTube, a LinkedIn post or a page on your own site. The things that make phone footage look amateur are almost never the sensor.

• Use the rear camera, not the selfie camera. On most phones the rear lens and sensor are substantially better. Prop the phone up and use a timer or a second person, or record a preview on a laptop screen behind the phone.
• Clean the lens. A fingerprint smear is the single most common cause of soft, hazy, low-contrast phone video. Shirt corner, ten seconds.
• Lock focus and exposure before you start — press and hold on the face until the app locks. Otherwise the camera hunts every time you move, and you get a visible brightness pump mid-sentence.
• Do not use digital zoom. Move the phone closer. Digital zoom is cropping, and it throws away detail.
• Put the phone on something. A ₹500 tripod, a stack of books, a chair. Handheld looks fine on a reel of someone walking around a factory and terrible on a two-minute explainer.
• Shoot 1080p unless you have a reason. 4K eats storage and editing time, and most of it will be viewed on a phone screen at a fraction of that resolution. The exception is when you plan to crop into the frame in the edit.
• Frame for where it will be posted. Vertical 9:16 for reels and status, or shoot 4:5 which survives both a feed and a modest vertical crop.

What I would not buy: a mirrorless camera, before the team has shot thirty videos on a phone. A proper camera does give you nicer depth of field and better low-light performance. It also gives you battery doors, memory cards, a separate audio recorder to sync, and a learning curve. Buy it when the phone is genuinely the limiting factor and someone in the building is excited about learning it. That is a real moment, and most businesses reach it later than they expect.

Captions, and the thing nobody mentions

There is an apparent contradiction here. Social feeds often autoplay video silently, so your video needs burnt-in captions — text on the screen — for the first few seconds to have any chance. If people watch with sound off, why obsess about sound?

Two reasons. First, captions win the first three seconds; sound wins the remaining ninety. The viewer who is interested turns the sound on, and that is the moment a hollow, echoing recording loses them. Second, automatic captioning tools transcribe clean audio far better than muddy audio. Accuracy on Indian English, and on the natural Hindi-English mixing most of us actually speak, is inconsistent at best — you will be correcting captions by hand either way, but you will correct far fewer lines if the recording is clean.

Be honest about how much correction that is. Budget the editing time, or the captions go out with your product name spelled three different wrong ways.

What I would not pay an agency for

The standard offer in this category is a "content day" — a crew arrives, shoots for six hours against a backdrop, and delivers twelve clips. It is good value on paper and it usually fails for a specific reason: all twelve clips were shot on one day, in one shirt, about whatever you happened to think of that morning. They go stale together, and there is nothing to say in month two.

The setup described above costs less than one such day and lets you shoot on a Tuesday because something interesting happened on Tuesday. That is worth more than production value. Where outside help genuinely earns its fee is in deciding what to shoot, editing consistently, and getting the clips in front of the right people — which is planning and distribution work, not camera work. If that is the part you want to hand over, our social media marketing work covers the planning and posting side, and there are more practical pieces on making things yourself in the blog.

What to do this week

Walk your office with your phone. Clap in three rooms and pick the deadest one. Note which window gives soft light in the late morning. Order one clip-on microphone and one small tripod. Then shoot a ninety-second answer to the question a customer asks you most often — with the microphone on your collar, facing the window, phone on the tripod, fan off.

Watch it back with headphones. If you can listen to the whole thing without wincing, you have a working setup, and everything after that is practice. If you want a second opinion on whether the clips you are making are worth promoting — or where they should sit on your site — get in touch and describe what you are trying to shoot.

Questions we get asked