Buying a Webcam for Video Calls: What the Software Actually Sends, and Why Light Comes First

A webcam is sold on a resolution number, and a video call is delivered at whatever resolution the meeting software decides to send, which is usually lower and is decided by your plan, your CPU, your bandwidth and how many people have video on. That gap is the whole buying problem. This guide lays out what the three big meeting platforms document about the video they send, what three widely sold webcams document about themselves, and what that implies about the order in which to spend money.

Documentation checked September 22, 2026. This is a selection guide built from vendor documentation, not a review of a camera we used. An earlier version of this article was written as a six-month first-person review of one webcam; it had no supporting records and has been replaced.

Illustration only; not a photograph of a specific camera or call.

What the meeting software will actually send

Before comparing cameras, it is worth knowing the ceiling. Each platform documents its own, and none of them defaults to the number printed on a webcam box.

Platform	What it says it sends	Conditions attached	Source
Zoom	Enabling HD “will activate 720p video quality”. Full HD 1080p is a separate tier that “must be enabled by Zoom Support”.	720p needs a “Pro, Business, Education, or Enterprise account” and 2.6 Mbps up / 1.8 Mbps down for group calls. 1080p needs a Business, Education or Enterprise account (or Pro with a Zoom Events or Webinars Plus licence), an “i7 Quad Core (physical core) CPU or higher”, 3.8 Mbps to send, and “virtual backgrounds must be disabled”.	Enabling HD video for Zoom Meetings
Google Meet	Three send tiers: 360p, 720p and 1080p. 1080p “uses the most data, but your camera sends the highest quality picture”.	Sending 1080p needs a camera and device with enough processing power, a qualifying Workspace edition (Business Standard/Plus, Enterprise, Education Plus) or Google One with 2 TB or more, and Chrome or Edge 91+. “1080p isn’t supported in Firefox and Safari.”	Improve your video & audio experience
Microsoft Teams	“Up to 1080p video resolution, and up to 30 fps” where bandwidth is not limited; Teams “can deliver HD video quality in under 1.5 Mbps”.	The minimum bandwidth tier (150 kbps) supports “up to 240p”; the recommended tier is 1.5 Mbps for one-to-one and 2.5 Mbps up / 4 Mbps down for meetings. “Expect up to 1080p quality but depending on your network conditions, video resolution, and quality will be optimized accordingly.”	Prepare your organization’s network for Teams

Three things follow from that table. First, none of the three platforms sends 4K in a normal meeting, so a 4K sensor is buying pixels that will be scaled down before anyone sees them. Second, 1080p is conditional on all three: a paid plan, a browser, a CPU, or a bandwidth figure that a home connection on a busy evening may not clear. Third, the platform decides per call, silently. A camera cannot push its resolution through a pipe the software has narrowed.

The practical check is to open the video settings of the platform you actually use and read what it offers you. On Zoom that is Settings → Video, where “HD” is a checkbox and 1080p does not appear unless your account tier includes it. On Meet it is the send-resolution selector, which will only list 1080p on a supported browser and edition. If the option you would be paying a camera for does not appear there, the camera cannot add it.

What three webcams document about themselves

These are three Logitech models chosen because they span the price range and because Logitech publishes a spec page for each. They are not a recommendation list, and this is not a full market survey; the point is what the spec sheets do and do not tell you.

Model	Listed price (US, at check)	Resolution options	Sensor and lens	Field of view	Microphone	Low-light claim	Source
C920s Pro HD	$69.99	“1080p at 30fps” or 720p/30	“Five-element glass lens”, “premium autofocus”	78°	“Two microphones, one on either side of the camera”	“RightLight 2 technology” adjusts “even if you’re in a dim setting”	Logitech C920s page
Brio 4K	not listed on the product page at check	“4K/30 fps … 1080p/60 fps … or 720p/90 fps”	“13MP” sensor, autofocus	90°/78°/65°	“Dual omni-directional built-in microphone with range up to 1.2 m”	“RightLight 3 for clear image in various lighting environments ranging from low light to direct sunlight”	Logitech Brio 4K page
MX Brio	$199.99	“UHD 4k resolution at 30fps or 1080p at 60fps”	“Our largest webcam sensor yet” with “70% larger pixels”; megapixel count not stated	not stated on the page	“Integrated dual beamforming noise-reducing mics”	“2x finer image detail and face visibility” versus the Brio 4K in low light	Logitech MX Brio page

Read the MX Brio row against the Brio 4K row. Both are 4K cameras. The newer, dearer one is marketed not on more pixels but on larger pixels and a larger sensor, and the benefit it claims is specifically in low light. That is the vendor’s own framing, and it is the right one: the resolution column is the same, and the thing that moved is how much light each pixel collects. A pixel count tells you how finely the sensor divides the light it receives; it says nothing about how much light there is to divide. A small sensor split into many pixels gives each pixel very little light, and the camera compensates by amplifying the signal, which is where the grain in a dim room comes from.

That is why a laptop’s built-in camera looks acceptable by a window and poor after dark, and why swapping it for a camera with the same resolution but a bigger sensor helps more than swapping it for a camera with a higher resolution and a similar sensor. The number on the box is the least informative column in the table.

Light: the lever every vendor quietly ships a workaround for

Notice how many rows in both tables are about lighting. Zoom has an “Adjust for low light” setting, documented as something that “improves video brightness and quality in low-light surroundings” with Auto and Manual modes (Zoom settings reference). Google Meet documents automatic lighting adjustment that “can automatically enhance brightness to improve your visibility”. Logitech’s RightLight, in three generations, is a low-light correction. Every one of these is software brightening an image that did not get enough light in the first place. Brightening in software raises the noise along with the signal; adding light raises only the signal.

Illustration only.

So the cheapest upgrade to any camera, including the one already in the laptop, is a light source in front of the face rather than behind it. Three placement rules cover most desks, and they cost nothing to test:

1. Face the window, do not sit with your back to it. A bright background makes the camera expose for the background and leave the face dark.
2. Put the lamp above eye level and slightly to one side, pointed at the face. Straight-on light flattens; light from below is unflattering; light from above and to the side is what a phone’s portrait mode is imitating.
3. Test at the time of day you actually take calls. A room that is fine at ten in the morning is a different room at six in the evening, and the camera you are evaluating will be judged in the second one.

If those three rules fix the picture, the camera was never the problem. If they do not, the next paragraph is where a purchase starts to make sense.

Autofocus: an upgrade with a failure mode

Most laptop cameras are fixed-focus: everything from a short distance to infinity is roughly, never exactly, sharp. Most external webcams in the table have autofocus, which is why a face in front of one looks crisper. The documented trade-off is that continuous autofocus re-evaluates focus as things move, and a hand raised near the lens, a lean forward, or a change in light can make it hunt: drift soft and snap back over a second or so.

Illustration only.

Two things are worth checking before buying rather than after. First, whether the vendor’s camera software lets you lock focus or slow the autofocus; the spec pages above name autofocus but do not document a manual override, so look for it in the vendor’s desktop app for the specific model, or in the platform’s own camera controls. Second, whether you actually move on calls. A person who sits still in front of a fixed camera gets all of autofocus’s benefit and none of its cost; a person who gestures, stands, or holds objects up to the lens gets the opposite.

Audio: the column the box says least about

Look again at the microphone column. The most specific claim across all three cameras is a pickup range of “up to 1.2 m”, and the rest is adjectives. A webcam’s microphone sits at monitor distance, above the screen, in front of whatever else is in the room. It will capture the voice, and it will also capture the keyboard, the fan and the room’s echo, because those are also within 1.2 metres of it.

Whether that matters depends on what the other side of the call experiences, which is the opposite of how cameras are chosen. A soft or grainy picture is a mild inconvenience for the listener; a voice that keeps dropping under room noise is a reason to ask you to repeat yourself. The upgrade that addresses that is a microphone close to the mouth, either a headset or a small desk microphone, and it is sold separately because webcam marketing is about the picture. If headphones are judged over a full working day rather than a demo, microphones should be judged by whoever is listening, and the simplest test is to record a minute and play it back.

A proposed order of spending

Putting the three sections together gives an order that is the reverse of what the word “webcam” suggests, and it is offered as a rule to test rather than a report of results:

Step	What it costs	What it fixes	How to test before spending more
1. Light	A lamp you probably own	Grain, dark face, the camera exposing for the window	Record 30 seconds in the platform you use, at your usual call time, with and without the lamp
2. Audio	A headset or small microphone	Being asked to repeat yourself	Record a minute with the laptop mic and with the headset; play both back on a phone speaker
3. Camera	$70 to $200 in the table above	Fixed focus, small sensor, narrow or too-wide framing	Only after 1 and 2, and only at the resolution the platform will actually send you

The reason light comes first is not that cameras do not matter; it is that every camera in the table, and every platform, already spends effort compensating for missing light, and the compensation is worse than the light. The reason audio comes second is that its failure is felt by other people, and theirs is the opinion that decides whether you look professional. The camera is third because its headline number is the one the software will most likely discard.

Illustration only.

Checklist before paying for a camera

1. Open your platform’s video settings and confirm the resolution it will send on your plan. If it caps at 720p, a 4K camera is buying nothing you will transmit.
2. Check the bandwidth figures in the table against your actual upload speed at call time. Zoom’s 1080p send needs 3.8 Mbps; Teams’ recommended meeting tier is 2.5 Mbps up.
3. Fix the light and re-record. Judge the recording, not the live preview, because the preview is not compressed and the call is.
4. Decide whether you move during calls. If yes, confirm the model has a focus lock or slow-focus option in its software before buying.
5. Decide who else has to see the frame. A 90° field of view shows the room; 65° shows a face. Of the three cameras above, one lets you choose, one is fixed at 78°, and one does not state its angle.
6. Compare sensor claims, not pixel claims, between two candidates at the same resolution.

Who this is not for

If video is on for a few minutes a week, or the room has a window and the calls happen in daylight, a built-in camera is adequate and the money is better spent elsewhere. The case for an external camera is a person on video for a large part of most days, in a room that cannot be lit by daylight at call time, who has already done steps one and two. Even then, the same rule that applies to a keyboard applies here: the property on the box is the one you notice in the first week, and the property that decides whether you keep it, here the way it behaves in your light over an hour of sitting still, is one you can only find by recording a call in your own room. As with screen position checked before a new panel, the cheapest change to the setup is the one to test first, and a lamp is cheaper than any camera in the table.