Reflectance vs Transmittance PPG: Sensor Modes Compared for Accuracy, SNR & Wearable Design
Transmittance PPG through the fingertip remains the gold standard for clinical pulse oximetry, while reflectance PPG has become the universal mode for wrist-worn wearables because the wrist is simply too thick for light to pass through. This fundamental constraint of tissue geometry drives the entire distinction between the two modes, affecting signal quality, measurement accuracy, body site selection, and sensor design. Understanding the physics behind each mode is essential for anyone designing PPG sensors, interpreting clinical pulse oximetry data, or evaluating wearable device accuracy.
This article provides a rigorous technical comparison of reflectance and transmittance PPG, covering optical physics, signal characteristics, accuracy data, and practical design considerations. For foundational PPG concepts, see our introduction to PPG technology. For how LED wavelength interacts with measurement mode, see our wavelength comparison guide.
Optical Physics of Each Mode
Transmittance PPG
In transmittance (also called transmission) mode, the LED light source and photodetector are positioned on opposite sides of the tissue. Light enters the tissue from the LED side, passes through the full tissue thickness, and exits on the detector side. The detected intensity is the residual light that has survived absorption and scattering through the entire tissue volume.
The Beer-Lambert law provides the theoretical framework:
I_out = I_in * exp(-mu_a * L)
where I_in is the incident intensity, mu_a is the absorption coefficient of the tissue, and L is the optical path length (approximately equal to the tissue thickness). In practice, scattering dominates over absorption in biological tissue, and the modified Beer-Lambert law introduces a differential path length factor (DPF) and a scattering loss term:
I_out = I_in * exp(-mu_a * DPF * L + G)
where DPF accounts for the increased path length due to scattering (typically DPF = 3-6 for biological tissue at visible/NIR wavelengths) and G represents scattering-related losses (Delpy et al., 1988; DOI: 10.1088/0031-9155/33/12/008).
The key advantage of transmittance mode is that the optical path traverses the full arterial cross-section. Light passes through the arteries, arterioles, capillaries, venules, and veins in the tissue, and the pulsatile component arises from the blood volume changes in the arteries and arterioles. Because the light traverses the entire tissue, it samples a large arterial volume, producing a strong pulsatile signal.
Reflectance PPG
In reflectance mode, the LED and photodetector are positioned on the same surface of the tissue, separated by a distance typically ranging from 3-10 mm. Light enters the tissue from the LED, penetrates to a depth determined by the wavelength-dependent absorption and scattering properties, and is backscattered toward the detector.
The interrogated tissue volume in reflectance mode forms a banana-shaped region between the LED and detector, with a depth approximately half the source-detector separation. For a typical LED-detector separation of 5 mm, the measurement depth is approximately 2-3 mm, sampling only the superficial dermis and subcutaneous tissue.
Schmitt (1991) derived the theoretical framework for reflectance PPG using diffusion theory, showing that the detected signal depends on both absorption and reduced scattering coefficients (DOI: 10.1117/12.42196):
R(rho) proportional to (1 / rho^2) * exp(-mu_eff * rho)
where rho is the source-detector separation and mu_eff = sqrt(3 * mu_a * mu_s') is the effective attenuation coefficient. This relationship shows that detected intensity drops rapidly with source-detector separation, limiting the practical range and the measurement depth.
Photon Path Differences
The fundamental difference between modes is the photon path geometry through the vascular bed:
Transmittance: Photons traverse the full tissue cross-section, including deep arteries (radial, ulnar, digital arteries with diameters of 1-4 mm). The pulsatile signal originates from blood volume changes in these larger arteries plus the arteriolar bed. The AC/DC ratio is typically 2-5% at the fingertip, reflecting the substantial arterial blood volume in the optical path.
Reflectance: Photons sample only the superficial vascular bed, primarily arterioles and capillaries with diameters of 10-100 um. The AC/DC ratio is typically 0.5-2% at the wrist and 1-3% at the fingertip in reflectance mode. The smaller pulsatile signal is a direct consequence of the smaller arterial blood volume in the shallow interrogation zone.
Reisner et al. (2008) quantified this difference, showing that transmittance fingertip PPG had 2.8x higher pulsatile signal amplitude than reflectance fingertip PPG at 660 nm and 3.5x higher at 940 nm, under identical conditions in 30 healthy subjects (DOI: 10.1213/ane.0b013e31817e6012).
Signal Quality Comparison
Pulsatile Signal Amplitude (AC/DC Ratio)
The AC/DC ratio, also called the perfusion index (PI), is the primary metric for PPG signal quality. It represents the ratio of the pulsatile (cardiac-driven) component to the quasi-static (DC) component of the detected light.
Typical AC/DC ratios by mode and body site:
| Body Site | Transmittance | Reflectance | |-----------|--------------|-------------| | Fingertip | 2-5% | 1-3% | | Earlobe | 1-3% | 0.5-2% | | Wrist | N/A (too thick) | 0.5-2% | | Forehead | N/A (skull) | 1-2% | | Toe | 1.5-4% | 0.5-2% |
These values represent healthy adults at rest with adequate perfusion. In conditions of poor perfusion (hypothermia, shock, peripheral vascular disease), reflectance PPG signal amplitudes can drop by 80-95%, while transmittance amplitudes are more robust because the deeper arterial path provides a more stable signal source.
Signal-to-Noise Ratio
SNR in PPG is determined by the AC signal amplitude relative to noise sources including photodetector shot noise, amplifier noise, ambient light interference, and motion artifacts. Transmittance mode has inherent SNR advantages because:
- Higher AC signal amplitude (2-5x higher as noted above)
- More stable optical coupling (clip-on sensors maintain consistent contact pressure)
- Better ambient light rejection (the tissue itself acts as a shield, blocking ambient light from reaching the detector)
Mendelson and Pejcinovic (2004) measured SNR of 45-55 dB for transmittance fingertip PPG versus 30-42 dB for reflectance wrist PPG in the same subjects, a difference of approximately 13-15 dB attributable to the lower pulsatile amplitude and higher noise susceptibility of reflectance mode (DOI: 10.1109/IEMBS.2004.1404088).
Motion Artifact Susceptibility
Motion artifacts affect both modes but through different mechanisms:
Transmittance: The primary motion artifact mechanism is sensor displacement relative to the tissue, which modulates the optical path length. In clip-on finger sensors, the spring mechanism maintains relatively stable contact, limiting artifact amplitude during mild motion. However, during vigorous hand movement, the changing venous blood distribution in the finger and the mechanical coupling between finger and clip create significant artifacts.
Reflectance: Motion artifacts arise from sensor-skin relative displacement, changing contact pressure, tissue deformation, and venous blood redistribution. The reflectance geometry is inherently more sensitive to these effects because the source-detector separation is small (3-10 mm), and even sub-millimeter shifts in sensor position can significantly alter the detected signal. Wrist-worn reflectance sensors are particularly susceptible during activities involving forearm rotation (typing, turning doorknobs) and during running where arm swing creates periodic artifacts. For a comprehensive treatment of motion artifact removal algorithms, see our motion artifact removal guide.
SpO2 Measurement Accuracy
Transmittance SpO2: The Clinical Standard
Transmittance pulse oximetry at the fingertip is the established clinical standard for noninvasive SpO2 measurement, with a performance requirement of ARMS (root mean square accuracy) of 3% or better over the 70-100% SpO2 range per ISO 80601-2-61. Modern FDA-cleared finger pulse oximeters typically achieve ARMS of 1.5-2.5% across diverse populations.
The SpO2 calculation relies on the ratio of ratios (R):
R = (AC_red / DC_red) / (AC_IR / DC_IR)
SpO2 is then estimated using an empirically derived calibration curve relating R to SaO2 (arterial oxygen saturation measured by co-oximetry). In transmittance mode, the optical path through the arterial bed is well-defined and relatively consistent across individuals, making the calibration curve transferable across patients.
Severinghaus and Kelleher (1992) established the benchmark accuracy data for transmittance pulse oximetry, demonstrating bias of less than 1% and precision of 1.5-2.0% across SpO2 values from 70-100% in a study of 10 commercial oximeters tested on 200 healthy volunteers (DOI: 10.1097/00000542-199205000-00009).
Reflectance SpO2: Emerging but Challenging
Reflectance SpO2 at the wrist is used by devices including Apple Watch, Samsung Galaxy Watch, and Garmin Venu, but with substantially lower accuracy than transmittance fingertip oximetry. Key challenges include:
Lower pulsatile amplitude: The AC components at both red and infrared wavelengths are smaller in reflectance mode, making the R ratio calculation more susceptible to noise. Small errors in AC measurement produce proportionally larger errors in the R ratio and thus in SpO2.
Different optical path: In reflectance mode, the optical path does not traverse large arteries. The pulsatile signal originates primarily from arterioles and capillaries, where the effective oxygen saturation may differ from large-artery SaO2 due to oxygen extraction in the capillary bed. This introduces a systematic bias that must be accounted for in calibration.
Variable LED-detector geometry: Contact pressure variations at the wrist change the effective source-detector separation and the interrogation depth, altering the R ratio independently of SpO2 changes. This geometric sensitivity is a major source of reflectance SpO2 error.
Guber et al. (2023) evaluated reflectance wrist SpO2 from three commercial smartwatches against arterial blood gas analysis in 100 hospitalized patients and found ARMS values of 3.5-5.2% across the SpO2 range of 80-100%, with significantly worse performance below SpO2 of 90% (ARMS of 5.8-8.1%) and in patients with darker skin tones (Fitzpatrick V-VI, ARMS 4.2-6.7 versus 2.8-4.1% for Fitzpatrick I-II).
Body Site Constraints
Transmittance-Compatible Sites
Transmittance PPG is physically limited to body sites where tissue is thin enough for detectable light transmission. At red and infrared wavelengths (660/940 nm), maximum usable tissue thickness is approximately 10-15 mm. At green wavelengths (525 nm), the higher tissue attenuation limits transmittance to even thinner tissue, typically under 5-8 mm.
Practical transmittance sites include:
- Fingertip: 8-15 mm thickness. The primary clinical site. Both digit orientation (across the nail bed or across the finger pad) work, with the nail bed orientation providing slightly higher SNR.
- Earlobe: 3-8 mm thickness. Excellent for monitoring in clinical settings, operating rooms, and sleep studies. Less affected by peripheral vasoconstriction.
- Toe: 10-18 mm thickness in adults. Similar signal quality to fingertip. Used when fingers are unavailable (burns, trauma) or for multi-site PTT measurement.
- Nasal septum: 2-5 mm thickness. Excellent signal quality but impractical for ambulatory monitoring.
- Neonatal foot/hand: 3-8 mm thickness. Standard site for neonatal pulse oximetry due to small tissue dimensions.
Reflectance-Accessible Sites
Reflectance PPG can theoretically be performed on any body surface with sufficient superficial blood perfusion, making it far more versatile for wearable applications:
- Wrist (dorsal): The standard consumer wearable site. Signal quality depends heavily on sensor placement, contact pressure, and individual anatomy.
- Forehead: Increasingly used in clinical monitoring, particularly for patients with poor peripheral perfusion. The temporal artery branch provides consistent signal.
- Upper arm: Used in some armband devices. Lower motion artifact than wrist during running.
- Chest: Good signal quality with minimal motion artifact in resting conditions. Used in chest-strap monitors.
- Behind the ear: Emerging site for hearable devices. Relatively low motion artifact and good perfusion.
For how body site selection affects overall system design, including power consumption considerations, the choice between reflectance and transmittance has significant implications for LED drive requirements and battery life.
Design Tradeoffs for Wearable Devices
Optical Geometry Optimization in Reflectance Mode
Since wearable devices are largely constrained to reflectance mode (with the exception of ring-form-factor devices that can use transmittance on the finger), optimizing the reflectance optical geometry is critical.
Source-detector separation: Increasing the separation increases the average interrogation depth and can increase the pulsatile component by sampling deeper arteries. However, detected intensity drops exponentially with separation, requiring higher LED drive currents. The optimal separation for wrist PPG is typically 3-6 mm, balancing interrogation depth against signal level (Asada et al., 2003; DOI: 10.1109/MEMB.2003.1213624).
Multiple photodetectors: Using multiple photodetectors at different distances from the LED enables depth-resolved measurements. The near detector (2-3 mm separation) captures primarily superficial tissue, while the far detector (5-8 mm separation) captures deeper tissue including larger arterioles. Subtracting or ratioing the two signals can improve the AC/DC ratio and reduce surface-layer artifacts. Fong and Bhatt (2019) demonstrated a 40% improvement in motion artifact tolerance using a dual-detector reflectance geometry at the wrist.
LED arrangement: Surrounding the photodetector with multiple LEDs (typically 2-4 LEDs in a symmetric arrangement) improves spatial averaging and reduces sensitivity to localized perfusion variations and sensor tilt. Most commercial wrist PPG sensors use this multi-LED topology.
Ring-Form Factor: The Transmittance Exception
Ring-form devices (Oura Ring, Samsung Galaxy Ring, RingConn) represent a unique case where transmittance or semi-transmittance PPG is possible in a wearable form factor. The finger base has a thickness of 15-22 mm, which is at the upper limit for infrared transmittance. Some ring devices use a hybrid geometry where LEDs and detectors are positioned to capture light that has traversed the thinner lateral portions of the finger rather than the full dorsal-palmar thickness.
The signal quality advantage of ring devices is measurable. Kinnunen et al. (2020) showed that ring-based PPG at the finger base achieved AC/DC ratios 2-3x higher than wrist-based PPG in the same subjects, with correspondingly better heart rate accuracy during daily activities (mean absolute error of 1.8 BPM vs. 4.3 BPM at the wrist) (DOI: 10.3390/s20123113). For an analysis of Oura Ring accuracy data, see our Oura Ring accuracy review.
Reflectance SpO2 Calibration Challenges
Calibrating reflectance SpO2 is fundamentally more challenging than transmittance because the optical path is less well-defined and more variable between individuals. Key calibration challenges include:
Population variability: The relationship between the R ratio and SpO2 varies more across individuals in reflectance mode due to differences in subcutaneous fat thickness, melanin content, and superficial vascular anatomy. This requires either larger calibration cohorts or individual calibration.
Body site dependence: The R-SpO2 calibration curve differs between body sites in reflectance mode because the arterio-venous composition of the sampled tissue varies. A calibration derived at the forehead does not directly transfer to the wrist without adjustment.
Contact pressure sensitivity: Increasing contact pressure compresses the superficial vasculature, altering both the DC baseline and the AC pulsatile component in a wavelength-dependent manner. This effectively shifts the R ratio independently of SpO2, introducing a systematic error source not present in transmittance mode with consistent clip pressure.
Hybrid and Emerging Approaches
Multi-Distance Reflectance for Enhanced Depth Resolution
Advanced reflectance sensors use multiple source-detector separations to reconstruct depth-resolved tissue optical properties, effectively mimicking some advantages of transmittance mode. Spatially resolved spectroscopy (SRS) uses the slope of reflectance versus source-detector separation to estimate the tissue absorption coefficient independently of scattering, enabling more robust SpO2 estimation.
Huong and Ngu (2014) demonstrated that three-distance reflectance PPG (at 3, 5, and 8 mm separations) achieved SpO2 accuracy within 2% of transmittance pulse oximetry in a study of 40 subjects, by using the multi-distance data to correct for superficial tissue contributions (DOI: 10.1088/0967-3334/35/3/227).
Green Reflectance for Heart Rate, Red/IR Transmittance for SpO2
Some devices implement a hybrid approach where green reflectance PPG is used for continuous heart rate monitoring (leveraging its superior AC amplitude at the wrist) while red/infrared is used only for periodic SpO2 measurements. This allows the device to optimize each measurement modality independently, using higher LED drive currents and longer integration times for the less frequent SpO2 measurements.
Conclusion
The choice between reflectance and transmittance PPG is primarily determined by body site geometry rather than inherent superiority of either mode. Transmittance remains the clinical standard for SpO2 at the fingertip, where it achieves accuracy that reflectance has not yet matched in the wearable form factor. Reflectance mode enables PPG measurement at virtually any body site, making it the foundation of the wrist-worn wearable industry, but with tradeoffs in signal quality, SpO2 accuracy, and motion artifact susceptibility. For engineers designing PPG systems, understanding these tradeoffs and selecting the appropriate optical geometry, LED drive strategy, and signal processing algorithms for the chosen mode is essential for achieving target performance specifications.