Fingertip HRV: phone cameras, finger clips and rings

The pad of your finger is one of the easiest places on the body to read a pulse with light. A phone camera, a finger-clip pulse oximeter and a smart ring all use the finger, and at rest each can time your heartbeats closely enough to estimate HRV. For how light-based pulse sensing works in general, read PPG: how your watch reads your pulse.

Why the fingertip

A fingertip is packed with small blood vessels just under the skin. Each heartbeat pushes a little extra blood into them, and blood absorbs light, so the light coming back from the finger dips once per beat. PPG PPG PPG An optical method that shines light into the skin and detects the pulse of blood with each heartbeat. Most watches, rings and phone-camera apps use it to estimate HRV. More about PPG sensors read this ripple of blood volume in the fine vessels (Allen, 2007).

The finger gives a stronger pulse than most other places. A study of 36 healthy adults recorded PPG at six sites: the finger, the inner and outer wrist, the arm, the earlobe and the forehead. The finger had the largest pulse of the six and the highest share of waves clear enough to analyse (Hartmann and colleagues, 2019).

Fingertip sensors come in two layouts:

• Transmission. The light sits on one side of the finger and the sensor on the other, so the light passes through. The finger-clip pulse oximeter used in clinics and sold for home use works this way.
• Reflection. The light and the sensor sit side by side and the sensor catches light that scatters back out of the skin. A phone camera with its flash works this way, and so does a ring.

1. Flash: an LED lights the fingertip from below
2. Light scatters through the skin and its small blood vessels, and some comes back out
3. Camera lens: each frame’s brightness becomes one sample
4. Each heartbeat dims the light a little: one pulse wave per beat
5. Ring: light and sensor side by side at the base of the finger (reflection)
6. Finger clip: light shines through the fingertip to a sensor on the other side (transmission)

Two side-view drawings. The first shows a fingertip lying on the back of a phone, covering the camera lens and the flash. A curved arrow runs from the flash up into the fingertip, through a few small blood vessels, and back down into the lens. Above the finger, a small wave with three bumps shows the pulse the camera records. The second drawing shows a whole finger. A ring sits around its base, with a light and a sensor side by side on the inside of the band and a short curved arrow between them. A clip closes over the fingertip, with a light in the upper jaw and a sensor in the lower jaw, and a straight arrow runs from the light down through the finger to the sensor.

How a phone camera reads your pulse

When you cover the rear camera and flash with a fingertip, the flash lights the skin and the camera films the glow that comes back. The picture is a red blur, because red light travels furthest through skin. The app does not need the picture itself. For each video frame it averages the brightness over the middle of the image, usually in the red colour channel and sometimes the green, and keeps that single number.

String those numbers together and you get a pulse wave. The app smooths out slow drift, finds the peak of each pulse and measures the time from one peak to the next. Those pulse-to-pulse gaps stand in for the beat-to-beat gaps of an ECG ECG ECG A recording of the heart's electrical activity, taken with electrodes on the skin. ECG-based RR intervals are the reference standard for measuring HRV. More about ECG , and a formula such as RMSSD RMSSD RMSSD Root mean square of successive differences: an HRV metric built from how much each heartbeat interval differs from the next, in milliseconds. It mainly reflects vagal activity. More about RMSSD turns them into an HRV number. Strictly, a pulse sensor measures pulse rate variability pulse rate variability Pulse rate variability Variation in the time between pulse beats, measured optically by watches, rings and phone cameras. It closely tracks HRV at rest but drifts from it during movement. More about Pulse rate variability (PRV), which tracks HRV closely at rest (Schäfer and Vagedes, 2013). HRV proper is timed from the ECG, the clinical reference described in ECG and 24-hour Holter HRV.

The frame rate is the sampling rate

A camera gives one brightness sample per frame, so its frame rate sets how finely it can time each peak. Most phones film at 30 frames a second, which leaves 33 ms between samples; some film at 60, which halves the gap. RMSSD is built from differences of a few tens of milliseconds between neighbouring beats, so a 33 ms step is coarse. Software can do better than the raw frames by fitting a small curve through the samples around each peak and placing the peak between frames. The sampling rate section of the PPG guide shows how much this matters.

Two camera habits add error. Frames do not always arrive evenly spaced, and a busy phone may delay or drop one, so an app that assumes perfect spacing misplaces beats. And cameras adjust their exposure and colour automatically, which can shift the brightness in steps that have nothing to do with your pulse. A 2015 study of an Android app that read frames at 30 a second found that, with people lying down, the phone model made a significant difference to the beat-to-beat error (Guede-Fernandez and colleagues, 2015).

This site’s camera test asks the browser for 30 frames a second, uses each frame’s own time stamp rather than assuming even spacing, and fits a curve through the top of each pulse to place it between frames.

What the studies show

Almost all the evidence comes from small groups of healthy adults at rest.

• Heart rate. A 2018 meta-analysis pooled 14 studies of camera apps that measure heart rate. In adults at rest with a normal rhythm, the apps agreed with a validated method, with correlations of 0.90 or higher; in children, the apps were not validated (De Ridder and colleagues, 2018). A 2024 review of 10 studies reached the same conclusion for healthy people, and pointed out that all of them ran under tightly controlled conditions (Mather and colleagues, 2024).
• HRV in 30 adults. Peng and colleagues compared 16 HRV measures from a phone camera with an ECG. Fourteen, including RMSSD, were strongly correlated with the ECG values, but only seven agreed closely enough to be swapped for them, and RMSSD was not among the seven (Peng and colleagues, 2015).
• RMSSD in 29 adults. Plews and colleagues recorded 5 minutes at rest with a phone camera app, a Polar H7 chest strap and an ECG at the same time. Compared with the ECG, differences in RMSSD were trivial for both the camera and the strap (Plews and colleagues, 2017).
• Slow breathing in 57 adults. During five 5-minute slow-breathing exercises, camera and ECG values of lnRMSSD lnRMSSD lnRMSSD The natural logarithm of RMSSD, with no unit. Taking the log evens out the spread of RMSSD values, which makes day-to-day changes easier to compare and analyse. More about lnRMSSD and three other measures agreed near perfectly, with intraclass correlations above 0.90 (van Dijk and colleagues, 2023).
• One app, two results. In a 2021 test of seven consumer products in five young adults, the CameraHRV app was off by 112% on average for RMSSD, while the most accurate products were within about 4% to 7% (Stone and colleagues, 2021). In a 2026 study of 37 athletes, the same app was off by 17.5% on average, against 2.2% for a Polar H10 chest strap, and its repeat readings were consistent (Johansson and colleagues, 2026).

Taken together, a camera can come close to an ECG at rest when the app processes the signal well, but a good result for one app or phone does not carry over to another. Fingertip PPG itself has a long record: in a 2008 study, fingertip PPG and an ECG gave matching HRV in 10 resting adults over 5 minutes (Selvaraj and colleagues, 2008). Every study above tested people at rest. During physical activity, pulse-based and ECG-based HRV often disagree to an unacceptable degree (Schäfer and Vagedes, 2013). What PPG does to HRV numbers explains where the timing error comes from and which measures it changes most.

Practical problems and how to avoid them

Pressing too hard

Pressure squeezes blood out of the small vessels the camera is watching. In 15 people with a reflection sensor on the finger, the pulse first grew and then shrank as the contact force rose from 0.2 to 1.8 newtons; most had their largest pulse at a light touch of 0.2 to 0.4 newtons (Teng and Zhang, 2004), about the weight of 20 to 40 grams. Pressing hard into a bright flash can also saturate the red channel, and a saturated channel shows no pulse at all. Rest your finger on the lens without pushing.

Cold fingers

Cold narrows the vessels in the skin, which shrinks the pulse. In 16 people, local cooling and warming of the skin changed the PPG signal (Jeong and colleagues, 2014). Warm your hands first, for example around a warm mug.

Movement

A small shift of the finger or the phone changes how light enters the skin and can swamp the pulse. Rest your hand and the phone on a table or your leg, breathe normally and do not talk.

The flash warms up

The flash stays on for the whole reading, and the phone can get warm around the camera. If the lens feels hot, stop and try again later; a hot spot makes you fidget, and fidgeting spoils the signal.

Which finger

Most people use an index finger. What matters is that the pad covers both the lens and the flash completely, that you can hold it still without strain, and that you use the same finger each time.

How long to record

The traditional short recording is 5 minutes. For RMSSD at rest, much less works: in 3,387 adults, a single 10-second ECG already gave a valid RMSSD, and 2 minutes agreed with 5 minutes almost perfectly (Munoz and colleagues, 2015). That study used an ECG. A camera times each beat less precisely and loses one now and then, so give it at least 1 minute and preferably 2. Readings of different lengths should not be compared freely (Shaffer and Ginsberg, 2017), so pick one length and keep to it.

Lying or sitting

Lie down or sit, and use the same position every time, because HRV differs between the two. Many people take the reading in bed soon after waking; a study of 28,175 app users based its results on short readings taken with a phone app on waking (Altini and Plews, 2021).

Finger clips and rings

A finger-clip pulse oximeter closes over the fingertip. A light in one jaw shines red and infrared light through the finger, and a sensor in the other jaw measures what gets through. It is built to show blood oxygen and pulse rate, and PPG of this kind is used in many medical devices for oxygen measurement (Allen, 2007). Most home oximeters show only those numbers, not the beat-by-beat timing, so they are rarely used for HRV. Research setups that record the fingertip pulse wave do get HRV close to an ECG at rest (Selvaraj and colleagues, 2008).

A smart ring reads reflected light from the base of the finger, with its lights and sensors on the inside of the band. Rings tend to do well at night for three reasons: a ring that fits well stays in close contact with the skin, the finger has a good blood supply, and during sleep the hand stays still for long stretches. In 49 adults, the Oura ring’s nightly average HRV agreed very closely with an ECG, with a mean difference of 1.2 ms (Kinnunen and colleagues, 2020). In 35 people monitored at home, the same ring measured nightly RMSSD accurately, both in 5-minute segments and as an average for the night (Cao and colleagues, 2022). In a study of 13 adults over 536 nights, two Oura models agreed with an ECG more closely for nightly HRV than WHOOP, Garmin and Polar devices (Dial and colleagues, 2025). These are small studies, and they cannot separate the ring’s position from the maker’s software. By day, with busy hands, a ring loses much of that advantage.

For how a chest strap times each beat from the heart’s electrical signal instead, see how chest straps measure HRV. For device-by-device evidence, see HRV on Oura, Whoop, Garmin and Apple Watch. For how wrist sensors compare with an ECG strap, see is wrist HRV as accurate as a chest strap?

Fingertip or wrist

A watch has a harder job. The wrist gives a smaller pulse than the finger (Hartmann and colleagues, 2019), the watch sits over tendons that move whenever your fingers do, and a strap can slide. That is why watches measure HRV mostly during sleep.

A camera suits a quick reading with no extra device; a ring or watch suits effortless nightly data. Compare any of them only with itself.

How to take a good camera reading

1. Measure at the same time each day, ideally soon after waking and before coffee or exercise.
2. Warm your hands if they are cold.
3. Sit at a table or lie down, in the position you will use every time.
4. Open the camera test on your phone and allow camera access. The flash turns on if the browser allows it; if not, sit near a bright lamp.
5. Rest the pad of one fingertip over the rear camera and the flash. Cover both completely and do not press. The preview should glow red.
6. Support your hand and phone on the table or your leg. Keep still, breathe normally and do not talk.
7. Record for 1 minute, or 2 if you can. The test waits 10 seconds after it finds your finger before it starts recording.
8. If the test reports a weak or noisy signal, warm your hand, ease the pressure and try again.

The test runs in your browser and the data stays on your device: each frame is reduced to its average colour on your phone, and the picture is never stored or sent. Nothing leaves the device unless you sign in and choose to save a result, and then only its RMSSD, mean heart rate mean heart rate Mean RR The average time between heartbeats in a recording, in milliseconds. Heart rate is the same number upside down: 60,000 divided by the mean RR. More about Mean RR , time and a note of which test and recording length you used are sent.

When to talk to a doctor

A fingertip reading is a wellness estimate. It cannot tell you why your pulse is uneven, and it cannot rule a heart rhythm problem in or out. If you often feel a racing, fluttering or uneven heartbeat, or you have chest pain, dizziness, fainting or shortness of breath, talk to a doctor.