Goniometer vs. Motion Capture Measurement in Physical Therapy



The observation and measurement of motion is the base of physical therapy. Without seeing and interpreting movement, appropriate assessment can become challenging. In this blog post, we will examine the different ways motion is traditionally observed or measured. Being a tech startup, we will examine the role of technology in observing motion.

For the purposes of the blog, the process of physical therapy is oversimplified and illustrated in the below schematic.

{else=abandon ship}

Figure 1: The (oversimplified) physical therapy process. Created by EuMotus.

The analog goniometer

The goniometer is a classic device for measuring angles. Cheap, accurate, and compact, it is a common tool in the physical therapy clinic, as well as in other fields.

Apparently the goniometer was used for geodetic triangulation almost 5 centuries ago. Birkhäuser, Basel (2014).

Figure 2: The classic goniometer [5].

The goniometer quickly helps provide a static angle measurement. Within the context of physical therapy, these type of measurements are well suited to establishing a patient’s passive range of motion. The speed of measurement is limited to the clinician’s dexterity and the stop-and-go process of passive range of motion procedure. Single measurements are easy and quick. However, performing a holistic physical exam consists of measuring ankle, knee, and hip flexion, and the shoulder ranges of motion in each plane. Multiple measurements become a more complex task.In the literature, there are widely ranging reports of the reliability of goniometry.For goniometry of shoulder range of motion, Hayes et. al. found interrater reliability ranges from 0.64 – 0.69, and intrarater: from 0.53 to – 0.65 [4]. Examining knee flexion, Watkins et. al. found near-perfect goniometer intratester reliability 0.99 for flexion and 0.98 for extension, and 0.90, 0.86 for intertest reliability, respectively [1]. Pandya et. al. found intertester reliability from 0.25 to 0.91 for goniometry for patients with a specific condition [2].


  • Easy to use for single measurements.
  • Repeatable results if the measurement manner is consistent (more reliable when one PT takes measurements [1]).
  • Compact.
  • Commoditized, therefore easy to procure and affordable.


  • Not practical for measuring an athlete’s active range of motion. Getting the subject to hold certain positions in their maximal extension/flexion may not be possible.
  • Problematic interrater reliability.
  • Limited to single joint measurements.
  • Effort and time-consuming process for holistic analysis.
  • Impossible to measure torsion.

Digital goniometer

  • Accuracy 2.5/5
  • Affordability 5/5
  • Ease of use 5/5 PROM; 1/5 AROM

Digital goniometers in the context of rehab therapy largely carry the same positives and negatives as analog goniometers. However, the attempt of angle measurement from a distance becomes a risk. Digital goniometers can manifest as software, simply relaying 2D cell phone camera captured still-frames. Other goniometers may rely on accelerometers and gyroscopes.

For video-based goniometers, if the angle of the camera is not set up appropriately (e.g. looking at a surface at an angle different than 0 degrees), there is added risk of measurement error.

For accelerometer and gyroscope based measurement tools, there is added risk of ‘measurement drift’.

Hayes et. al. found that for still photography (i.e. digital goniometer) interrater reliability ranged from 0.62 – 0.73 and intrarater reliability of 0.56 – 0.61 [4], worse than traditional handheld goniometers.

Benefits (incremental to analog goniometer): -Incorporated in your phone – no need carry extra stuff or plan ahead.

Drawbacks (incremental to analog goniometer): -Subject to additional measurement error risk.

Markerless motion capture

  • Accuracy 4/5 (compared to a max of 5/5 in traditional multi-camera marker-based systems)
  • Affordability 4/5
  • Ease of use 4/5

At first glance, one could ask the question – why measure active range of motion using markerless motion capture? It may seem like overkill compared to a basic and quick goniometer measurement.


Figure 3: Commercially available 3D motion capture sensor [6].But it can also create value and free up time for higher level work.

While markerless motion capture software is certainly less practical for one-off passive range of motion measurements, holistic movement analysis becomes an attractive proposition. Mocap software automatically detects the biomechanics of multiple body joints. The clinician no longer spends his or her energy and time recording motion measurements. Rather, in a couple of minutes, the PT runs a movement screen on their patient. Instead of spending his or her time and energy to manually measure and log measurements, he or she focuses on observing the entire kinetic chain and performing analysis and treatment.

Studies comparing gold standard traditional motion capture systems (Vicon system and force plates) to markerless motion capture resulted in intraclass correlation coefficient (ICC) of 0.84 for initial-contact of a drop-jump test, and 0.95 for peak flexion [7].

Bonnechere et. al. found intraday ICC of 0.94 posture assessment compared to 0.98 for traditional motion capture. For interday measurements, ICC for markerless mocap was 0.88, compared to 0.87 for traditional motion capture, suggesting that for some use cases, markerless motion capture is as accurate as traditional motion capture [8].


  • Greater ease of use and time-savings compared to measuring tens of measurements using manual, static goniometry.
  • Automatic data capture and logging. No need to manually record measurements. Can track progress.
  • As opposed to spending time and effort in manually making and recording measurements, the clinician’s time is now freed up to synthesize and analyze physical health from a mix of measurements.
  • Capable of measuring metrics that require 3 dimensional data, such as torsion.


  • Overqualified for certain cases, such as one-off passive range of motion measurements.
  • Portable but to a lesser degree than goniometers.
  • More expensive than commodity goniometers.

Unaided visual observation

  • Accuracy 1/5
  • Affordability 5/5
  • Ease of use 5/5 for PROM 1/5 for AROM

Unaided visual observation is the de facto standard of physical therapy examination. It is quick, and does not cost anything. Eyeballing a single measurement is an easy task. However, similar to goniometry, simultaneously monitoring multiple body joints can be a challenging task.

While visually observing a measurement is easy to perform, researchers have found inferior accuracy relative to the use of a goniometer or motion capture tools.

For shoulder measurements, Riddle et. al. found intertester reliability between 0.26 to 0.55 [3].

Hayes et. al. found that in assessing shoulder ROM visual estimation interrater reliability ranged from 0.57 to 0.7, an intrarater from from 0.59 to 0.67 [4].

Watkins et. al. found intertester reliability of visual estimation at 0.83 for flexion and 0.82 for extension [1].

Watkins et. al. recommends the use of a goniometer to reduce error of unaided visual measurement of passive range of motion [1].


  • Free
  • Easy to use for single body joint measurements.


  • Greatest degree of error.


Measurement is the initial step in complex biomechanical analysis.

Unaided observation is an invaluable and irreplaceable part of the physical therapy process. However, observation can carry relatively more error than other measurement methods.

Traditional goniometers are well suited to one-off passive range of motion measurements and are an upgrade to unaided visual observation, yet carry their own drawbacks. It is not feasible to reliably measure active range of motion using goniometers, nor is it practical to make multiple measurements at the same time (isolated movement vs full kinetic chain examination). Error also increases if multiple persons measure a subject using a goniometer due to inconsistent measurement methodology.

In a holistic biomechanical analysis, motion capture systems can help speed up the data capture, so the clinician’s time is freed up to perform assessments and work with their clients. Unlike goniometry, motion capture allows for active range of motion measurements. Repeatability, concurrent measurement of multiple body joints, and higher accuracy of mocap mean that the data generated is typically more reliable than data obtained using goniometry or observation alone.

For single joint passive ROM measurement, motion capture can be overkill. That being said, it should certainly be considered in the case of holistic movement screens, which have become more and more prevalent in standard patient assessments.



[1] Watkins M., Riddle D., Lamb R., Personius W. Reliability of goniometric measurements and visual estimates of knee range of motion obtained in a clinical setting. Physical Therapy Volume 71, Number 2, 90-6; discussion 96-7. February 1991.

[2] Pandya S., Florence J., King W., Robison J., Oxman M., Province M.. Reliability of goniometric measurements in patients with Duchenne muscular dystrophy. Phys Ther. 1985 Sep;65(9):1339-42.

[3] Riddle D., Rothstein J., Lamb R. Goniometric reliability in a clinical setting. Shoulder measurements. Phys Ther. 1987 May;67(5):668-73.

[4] Hayes K., Walton J., Szomor W, Murrell G. Reliability of five methods for assessing shoulder range of motion. Australian Journal of Physiotherapy. Volume 47, Issue 4, 2001, Pages 289-294.

[5] Goniometer. Wikimedia. https://upload.wikimedia.org/wikipedia/commons/2/20/Goniometer.svg. Accessed January, 2018.

[6] Microsoft Kinect. Flickr. https://c2.staticflickr.com/8/7457/14175572202_64b7e17e59_b.jpg. Accessed January, 2018.

[7] Gray A., Willis B., Skubic M., Huo Z, Razu S., Sherman S., Guess T. Development and Validation of a Portable and Inexpensive Tool to Measure the Drop Vertical Jump Using the Microsoft Kinect V2. Sports Health. 2017 Nov/Dec;9(6):537-544.

[8] Bonnechère B., Sholukha V., Jansen B., Omelina L., Rooze M., Van Sint Jan S. Telemedicine and e-Health. May 2014, 20(5): 451-453.

Leave a Reply

Your email address will not be published. Required fields are marked *