← How to benchmarkPeripheral guide
How to check a used webcam
Webcams are sold on a resolution number that tells you almost nothing. What decides whether the picture looks good is the sensor size, the frame rate it can actually sustain at that resolution, and how it behaves when the light is less than ideal — which is every room most of the time.
The short answer
Test it in the light you will actually use it in, at the resolution and frame rate you will actually use. A webcam that looks acceptable by a window at noon and falls apart under a ceiling light in the evening is the normal case, not the exception.
Get these first
Test it the way you will use it
- Confirm the frame rate it sustains at the resolution it claims. Many cameras advertise 1080p and deliver it at 30fps, then drop into the teens as soon as the light dims, because the sensor needs a longer exposure. Open the camera in software that shows the negotiated rate and check that it holds while you are actually lit the way you normally are.
- Test in your evening lighting, not by a window. This is the single most informative test and almost nobody does it. Sensor size, not megapixels, decides low-light performance, and a small sensor in a room lit by one overhead bulb produces a noisy, smeared, low-frame-rate image no matter what the box says.
- Watch the autofocus with a still subject. Sit still and look for the picture breathing in and out of focus. Autofocus hunting is a common failure as the focus motor wears, and it is far more distracting on a call than being slightly soft. Confirm whether focus can be locked manually, because on many models it cannot.
- Check the exposure behaviour when something bright is behind you. Sit with a window or lamp behind you and see whether the camera exposes for your face or turns you into a silhouette. Some cameras handle this well, many do not, and it is not something firmware updates fix.
- Test the built-in microphone separately. Record yourself and listen on headphones. Webcam microphones are frequently the actual reason people upgrade, and a failed or intermittent one is easy to miss when you are testing the picture.
- Try it on a different USB port, ideally a different controller. Uncompressed high-resolution video uses real bandwidth, and a camera that works alone may drop frames when it shares a controller with other devices. If you use a hub or a dock, test through that specifically.
- Work the mount through its full range. The clip is the part that breaks. Open and close it, check the hinge holds an angle without drifting, and confirm the tripod thread underneath is not stripped — that thread is often the only alternative once the clip gives up.
- Check whether the privacy shutter or cover is intact. These snap off, and on models where the cover is part of the housing, losing it is not repairable. Minor, but it is the kind of thing that annoys someone daily.
The one nobody checks
Low light. Every webcam looks decent facing a window at midday, which is exactly when people test them and exactly when nobody takes calls. Sensor size does the work that the resolution number takes credit for, and a small sensor under one ceiling bulb produces noise, smearing and a frame rate in the teens. Test it in the evening, in the room you actually work in.
Note
Resolution is the least useful number in this category. A 1080p camera with a larger sensor will look better in a real room than a 4K camera with a small one, because at typical call bandwidth the 4K stream is being compressed back down anyway. Judge the picture, not the badge.
What to measure
Higher is better▲Higher is betterLower is better▼Lower is betterMust be zero∅Must be zeroMust match the spec=Must match the spec
What actually determines picture quality
| Metric | Which way | What it tells you | Healthy looks like |
|---|---|---|---|
| Sustained fps at your resolution | Higher is better | The number behind "it looks laggy". | Holding 30fps at 1080p in your normal lighting, not just in daylight. |
| Low-light noise | Lower is better | Sensor size doing the work megapixels get credit for. | Usable under a single ceiling light. Test at the time of day you take calls. |
| Autofocus stability | Must be zero | Wear in the focus motor. | Locks and stays locked on a still subject. No breathing in and out. |
| Backlight handling | Must match the spec | Whether a window behind you erases your face. | Exposes for the subject. Not fixable in firmware if it does not. |
| Microphone | Must be zero | Often the real reason a camera gets replaced. | Clear and consistent on a recorded playback through headphones. |
| Behaviour on a shared controller | Must be zero | Bandwidth contention that only appears in real setups. | No dropped frames through your actual hub or dock. |
| Clip and hinge | Must be zero | The part that physically breaks. | Holds an angle without drifting; tripod thread not stripped. |
| Privacy cover present | Must be zero | Small, and often not replaceable. | Present and still attached. |
Walk away if
Walk away if
- A frame rate that collapses as soon as the light drops.
- Autofocus that hunts continuously on a still subject.
- A silhouette whenever there is a window behind you.
- A microphone that cuts in and out.
- Dropped frames through the hub or dock you actually use.
- A stripped tripod thread on a camera with a broken clip.
- A seller who only demonstrated it in daylight.
What each fault is worth
Bring the measurement, not the opinion. Name the fault, name the fix, name what the fix costs — that is a solvable problem with two acceptable answers, which is what makes it work.
| What you found | What it actually costs | What to ask for |
|---|---|---|
| Poor low-light performance | Unusable for evening calls. | Price it as a daylight-only camera, because that is what it is. |
| Autofocus hunting | A distracting picture with no fix. | A large discount, unless focus can be locked manually. |
| Dead microphone | A separate purchase. | The cost of the microphone you will now need. |
| Broken clip, good thread | Needs a stand or tripod. | The cost of a small tripod. |
| Missing privacy cover | Minor, but permanent on some models. | A token amount — worth raising, not worth arguing over. |
Ready to go looking?
Take the checklist with you, then browse what is listed. Everything on this page works just as well the other way round when you are the one selling.
Read alongside