Verdict: Among under-display fingerprint readers, 3D Ultrasonic sensors (used in the Samsung Galaxy S26/S25 series, Google Pixel 9/10 series, OnePlus 13/15, and Xiaomi 15/16 series) beat 2D Optical sensors in speed, dark-room comfort, and wet-finger reliability because they bounce high-frequency sound waves off the 3D depth of your skin ridges rather than shining a blinding OLED spotlight to photograph your thumb. Optical sensors hold only one practical advantage: their light-based cameras tolerate cheap, thick tempered-glass screen protectors that leave microscopic air gaps. Meanwhile, Apple’s 3D TrueDepth Face ID delivers the highest mathematical spoof resistance (a less than 1-in-1,000,000 false-accept rate versus Android’s 1-in-50,000 Class 3 threshold) and never fails when your hands are wet, gloved, or peeling—whereas standard 2D camera facial recognition on most Android phones is only a Class 1 convenience feature that cannot authorize banking apps or FIDO2 passkeys.
Smartphone manufacturers rarely explain the hardware trade-offs behind the biometric sensor listed on a spec sheet until you install a tempered-glass screen protector and suddenly can’t sign into your phone, or try to approve a mobile payment in the rain. Below is how 3D Ultrasonic fingerprint readers, 2D Optical fingerprint readers, and 3D infrared Face ID compare across security metrics, real-world failure conditions, and screen protector compatibility.
At-a-Glance Comparison: Ultrasonic vs. Optical Fingerprint vs. 3D Face ID vs. 2D Face Recognition
| Feature / Metric | 3D Ultrasonic Fingerprint (Under-Display) | 2D Optical Fingerprint (Under-Display) | 3D Structured-Light Face ID (Infrared Depth) | 2D Camera Face Recognition (Standard Android) |
|---|---|---|---|---|
| Sensing Physics | High-frequency acoustic pulse-echo maps 3D epidermal ridges, valleys, and pores | OLED panel lights up at max brightness so a CMOS macro camera photographs a 2D print | Infrared flood illuminator + dot projector map a 3D mesh of thousands of facial points | Standard RGB front selfie camera matches a flat 2D image of your face |
| Cryptographic Security Tier | Android Class 3 (Strong) | Android Class 3 (Strong) | Apple Secure Enclave / Exclave (or Android Class 3 on 3D ToF phones) | Android Class 1 (Convenience) on most phones (Class 3 only on Pixel 8–10 dual-pixel ML) |
| False Accept Rate (FAR) | At least 1 in 50,000 (lower on wide-area 3D Sonic Gen 2 / Max) | At least 1 in 50,000 | Less than 1 in 1,000,000 (with single enrolled appearance) | Varies widely; Spoof Acceptance Rate (SAR) exceeds 20% against printed/screen photos |
| Spoof Acceptance Rate (SAR) | 0% – 7% (Resistant to 2D photos and thin latex casts) | 0% – 7% (Relies on optical liveness / blood-flow heuristics) | < 1% (Attention-aware IR depth rejects flat photos and most masks) | > 20% (Not trusted for cryptographic key binding) |
| Performance With Wet or Oily Fingers | Excellent (Sound waves penetrate thin water and sweat films) | Poor (Water droplets refract OLED light and blur the 2D reflection) | Unaffected (Hands-free) | Unaffected (Hands-free) |
| Nighttime / Dark Room Comfort | Invisible (Zero screen flash required to read print) | Bright Glare (Must flash a 600–1,000 nit white/green circle to illuminate finger) | Invisible (940nm infrared flood illuminator works in total darkness) | Poor (Requires screen fill-flash or ambient room light) |
| Tempered Glass Screen Protector Tolerance | Strict (Fails if there is any air gap or glass thicker than ~0.25–0.30 mm) | Forgiving (Light passes across small air gaps and thicker 0.33 mm glass) | High (As long as the TrueDepth IR notch/cutout is uncovered or uses clear glass) | High |
| Authorizes Banking Apps, Wallet & Passkeys? | Yes (Full Hardware Keystore & FIDO2 access) | Yes (Full Hardware Keystore & FIDO2 access) | Yes (Apple Pay, Passkeys, App Store, Password Manager) | No (Except on Google Pixel 8/9/10 series with Class 3 ML verification) |
| Representative Phones (2026) | Samsung Galaxy S26/S25 series, Pixel 9/10 series, OnePlus 13/15, Xiaomi 15/16, iQOO 13/15 | Mid-range Galaxy A-series, Nothing Phone series, Motorola Edge series, Older OnePlus/Pixel 6–8 | iPhone 15, 16, 17 & 18 series, iPad Pro, Honor Magic Pro series | Secondary lock-screen convenience option on most Samsung, Xiaomi, Motorola & OnePlus phones |
1. How Each Biometric Sensor Actually Works Under the Glass
3D Ultrasonic Fingerprint Sensors (Acoustic Pulse-Echo)
Bonded directly beneath the flexible AMOLED display panel, an ultrasonic transducer—most commonly from Qualcomm’s 3D Sonic Gen 2 or 3D Sonic Max family—emits a brief, high-frequency ultrasonic pulse upward through the display stack when your thumb touches the target zone. Because epidermal fingerprint ridges are pressed firmly against the Gorilla Glass while the microscopic valleys between ridges contain tiny pockets of air, the acoustic wave reflects back at different intensities and travel times. The sensor captures this 3D acoustic topography—including sweat pores and ridge depth—in roughly 150 to 250 milliseconds without needing a single photon of light from the display.
2D Optical Under-Display Fingerprint Sensors (OLED Illumination + CMOS Camera)
An optical under-display sensor is fundamentally a miniature digital camera mounted underneath the microscopic gaps between the red, green, and blue subpixels of an OLED screen. Because the underside of your finger is completely dark when pressed against glass, the phone must briefly drive a circular patch of OLED pixels to maximum brightness (typically bright white, cyan, or green). That light reflects off your fingerprint ridges, travels back down through the pinholes in the OLED backplane, and hits a tiny CMOS image sensor that compares the flat 2D contrast pattern against your stored template.
3D Structured-Light Face ID (Apple TrueDepth)
Rather than capturing a standard color photo, Apple’s TrueDepth camera system—housed inside the Dynamic Island on iPhone 15, 16, 17, and 18 models—fires an invisible Flood Illuminator alongside a laser Dot Projector that projects thousands of invisible infrared dots onto your face. An infrared camera reads the distortion of that dot grid to compute an exact 3D depth map of your eye sockets, nose bridge, and contours, which is then processed by the Neural Engine inside the hardware-isolated Secure Enclave.
2. Security & Cryptographic Trust: Android Biometric Class 3 vs. Apple Secure Enclave
A common misconception is that an optical fingerprint sensor is treated as “less secure” by banking apps than an ultrasonic sensor. In reality, both Android and iOS govern biometric access through strict mathematical thresholds rather than marketing labels.
How Android Classifies Biometrics (Class 3 vs. Class 2 vs. Class 1)
According to the Android Open Source Project (AOSP) Biometric Security specification, every biometric sensor on an Android phone is tested against imposter attacks and assigned to one of three architectural classes based on three metrics: Spoof Acceptance Rate (SAR) (how often a physical replica like a printed photo, video, or latex mold tricks the sensor), False Accept Rate (FAR) (how often a random stranger’s real finger or face matches), and False Reject Rate (FRR):
- Class 3 (formerly “Strong”): Requires a Spoof Acceptance Rate (SAR) between 0% and 7%, a False Accept Rate (FAR) no higher than 1 in 50,000, and a False Reject Rate (FRR) at or below 10%. Only Class 3 sensors are allowed to access hardware-backed
Keystorecryptographic keys, authorize Google Wallet tap-to-pay, sign FIDO2 passkeys and password managers, and authenticate banking apps via theBiometricPromptAPI. Class 3 sensors allow a maximum of 72 hours before forcing a primary PIN, pattern, or password fallback. Both 3D Ultrasonic and modern 2D Optical fingerprint sensors on certified Android phones qualify as Class 3. - Class 2 (formerly “Weak”): Permits a Spoof Acceptance Rate (SAR) between 7% and 20% alongside a 1-in-50,000 FAR. Class 2 biometrics can show the
BiometricPromptdialog for basic app sign-in, but cannot bind cryptographicKeystorekeys or authorize financial transactions, and enforce a stricter 24-hour fallback timeout (or 4-hour idle timeout). - Class 1 (formerly “Convenience”): Has a Spoof Acceptance Rate exceeding 20%. Standard 2D front-camera facial recognition on Samsung Galaxy, OnePlus, Xiaomi, and Motorola phones sits in Class 1: it can dismiss your lock screen for convenience, but banking apps and Google Wallet ignore it completely. (Notably, Google’s Pixel 8, Pixel 9, and Pixel 10 series use dual-pixel autofocus depth data and Tensor machine-learning liveness verification to pass Class 3 certification with a single front camera.)
How Apple’s Face ID Enforces Hardware Security
According to Apple’s official Face ID Advanced Technology documentation, the probability that a random person in the population could look at your iPhone or iPad Pro and authenticate via Face ID is less than 1 in 1,000,000 with a single enrolled appearance—twenty times more selective than the 1-in-50,000 baseline used by Touch ID and Android Class 3 fingerprint standards.
Apple also enforces strict anti-coercion and timeout boundaries inside the Secure Enclave (and the Secure Exclave on iPhone 18 Pro models or later, which isolates supplemental camera processing):
- Face ID allows a maximum of 5 unsuccessful match attempts before requiring your passcode.
- Your passcode is mandatory if the device has just restarted, has not been opened for more than 48 hours, or if the passcode has not been used to authenticate the device in the last 6.5 days and Face ID has not authenticated it in the last 4 hours.
- With Require Attention for Face ID enabled, the TrueDepth infrared camera verifies that your eyes are open and actively directed at the screen, preventing someone from opening your iPhone while you are asleep.
3. The Screen Protector Problem: Why Tempered Glass Breaks Ultrasonic Sensors
If you have ever applied a 9H tempered-glass screen protector to a Samsung Galaxy Ultra or Google Pixel Pro and watched your fingerprint success rate drop to zero, you ran into fundamental physics rather than a defective phone.
- Why Ultrasonic Sensors Hate Air Gaps: Sound waves travel efficiently through dense solids (like Gorilla Glass and skin) but reflect almost 100% at a solid-to-air boundary due to acoustic impedance mismatch. Inexpensive tempered-glass protectors that only apply adhesive around the black outer borders leave a microscopic 0.05 mm air cushion over the center of the screen. That thin layer of air acts like a soundproof wall to an ultrasonic sensor. Even full-glue tempered glass thicker than 0.25–0.30 mm dampens the return echo enough to trigger “Press harder” errors.
- Why Optical Sensors Tolerate Thicker Glass: Light waves travel through a microscopic air gap with minimal distortion, provided the glass is clear and does not have a black privacy-filter louver grid. As a result, 2D optical fingerprint readers continue working through most standard $10 tempered-glass protectors.
If your under-display sensor is failing after applying a screen protector: Delete your old fingerprint templates, enable Touch Sensitivity (under Settings > Display), and re-enroll the same thumb twice while pressing at normal angles. For step-by-step calibration fixes across Samsung, Pixel, and OnePlus, see our complete guide on fixing a slow or unresponsive phone fingerprint sensor, and verify your display digitizer isn’t dropping touches in the target zone using our Free Online Touch Screen Test.
4. Daily Usability Trade-Offs: Wet Thumbs, Night Glare, Desk Angles & Dry Skin
Beyond raw security numbers, the three biometric systems fail under completely different everyday conditions:
- After washing hands, swimming, or working out (Winner: 3D Face ID, followed by Ultrasonic): Water droplets on a fingertip scatter the bright light of a 2D optical sensor, causing repeated failures until you wipe both your thumb and the screen dry. Ultrasonic sensors bounce acoustic pulses right through thin moisture and sweat layers, while 3D Face ID never touches your hands at all.
- Checking notifications in a dark bedroom at 2:00 AM (Winner: Ultrasonic & 3D Face ID): Because a 2D optical fingerprint sensor must fire an localized 800-nit light spot to illuminate your thumb ridges, any slight misplacement of your finger blasts bright white or green glare into your eyes in a dark room. Both 3D Ultrasonic sensors and 940nm infrared Face ID operate invisibly in pitch darkness.
- Opening the phone while it sits flat on a desk (Winner: Ultrasonic & Optical Fingerprint): Even though modern TrueDepth cameras support wide angles and vertical or horizontal orientation, authenticating a phone with Face ID while it rests flat on a conference table still forces you to lean forward over the top bezel. An under-display fingerprint sensor lets you tap the glass with your index finger without moving your head.
- Winter gloves, rock climbing, or peeling fingertips (Winner: 3D Face ID): Severe winter skin dryness, superglue residue, guitars/weightlifting calluses, or wearing non-conductive gloves stops every fingerprint reader cold. Face ID works normally with winter gloves and scarves (and supports Face ID with a Mask by scanning unique periocular features around your eyes). If your iPhone’s TrueDepth sensor ever stops recognizing you or displays a sensor warning, follow our iPhone Face ID Not Working / Not Available troubleshooting guide.
Which Biometric System Should You Choose?
- Choose a phone with a 3D Ultrasonic Fingerprint Sensor (plus Class 3 Face Recognition where available, such as Pixel 9/10 Pro, or Galaxy S26/S25 series) if: You prefer authenticating your phone by feel as you pull it out of your pocket, want zero bright screen flash in dark rooms, and frequently use your phone with damp hands at the gym or in the kitchen. Pair it with a thin TPU film or an ultrasonic-certified OCA tempered-glass protector.
- Choose 3D Structured-Light Face ID (iPhone 15–18 series or Honor Magic Pro) if: You want effortless hands-free authentication for Apple Pay and passkeys, live in a cold climate where you wear gloves or suffer from dry winter fingertips, or prefer thick tempered-glass screen protectors that would interfere with ultrasonic readers.
- A 2D Optical Fingerprint Sensor is still a solid choice if: You are buying a value flagship or mid-range Android phone and always apply thick tempered-glass protectors—just remember to keep a passcode handy when your hands are wet.
Frequently Asked Questions
Can someone open an ultrasonic or optical fingerprint sensor while I am sleeping?
Unlike Apple’s Face ID (which checks that your eyes are open and looking at the screen via Attention Awareness), both ultrasonic and optical fingerprint sensors only verify that a live finger matching your stored template is pressed against the glass—they cannot tell whether you are awake. If you are entering a high-risk environment or want to temporarily disable all biometrics before sleeping, enable Lockdown Mode on Android (hold the Power button and tap Lockdown) or press and hold the Side button + either Volume button for 2 seconds on iPhone to force passcode-only entry.
Why do polarizing sunglasses sometimes block Face ID or make optical fingerprint targets hard to see?
Certain heavily polarized sunglasses integrate infrared-blocking coatings on the lenses that prevent the TrueDepth 940nm infrared camera from seeing whether your eyes are open. If Face ID fails only when you wear sunglasses, you can either look at the screen briefly while tilting your glasses or toggle off Require Attention for Face ID under Settings > Face ID & Passcode (noting that turning off attention checks slightly lowers protection against unauthorized eyes-closed scans).
Are dual-sensor phones (3D Ultrasonic Fingerprint + 3D ToF Face Recognition) available in 2026?
Yes, though they remain rare outside select flagship lines such as Honor’s Magic Pro series, which pairs a pill-shaped 3D Time-of-Flight (ToF) front depth camera with an under-display ultrasonic fingerprint reader. Google’s Pixel 8, Pixel 9, and Pixel 10 series achieve a similar dual Class 3 biometric experience without a wide pill cutout by combining an under-display fingerprint sensor with Dual-Pixel PDAF + Tensor ML Class 3 Face Recognition (with the only caveat that camera-based Pixel facial recognition still needs ambient light, unlike infrared Face ID).

