Since the 1990’s the principal investigative tool for articulatory speech research has been the electromagnetic articulograph EMA. While this is an accurate, high resolution instrument, it is expensive and not affordable on a lab budget.  It is also limited in where sensors can be placed, how much the wires interfere with natural speech, and what can be measured. EMA allows sensor x/y positions to be measured along with derivatives of velocity and acceleration. Sensor displacement and tangential velocity (aka speed) of sensors are typical measures used in articulatory phonetic research.

The cost of the instrument has meant that he field of articulatory phonetic research has only been accessible to a few phonetics laboratories that can afford the equipment and the technicians to run it.  Partly as a result of this, theories such as articulatory phonology have dominated the field. Such theories are closely tied to the measurements provided by EMA. For example, Tongue body constriction degree TBCD and location TBCL and tongue tip constriction degree TTCD and location TTCL are tightly linked to the EMA sensors placed on the tongue body and blade. The theory has remained largely unchallenged even though there are no physiological sensors to directly detect closure distance. The theory has a highly interdependent relationship with the limitations of the instrument used to investigate it.

Articulate Instruments is a research vehicle that allows new ideas to be investigated without the need to apply for funding and recruit and run an experiment. The research is then implemented in the AAA app at no additional cost to users. The company also provides support for users at no cost.  To date there are over 200 phonetics labs with access to this research and the means to carry out independent studies on a lab budget.

In 2019 Articulate Instruments applied pose estimation to ultrasound image analysis with promising results. Since then, this strand of research has matured. The most recent studies show keypoints (virtual EMA sensors), estimated from ultrasound images now have accuracy and resolution comparable with EMA. These keypoints are not limited in there positions. Keypoints can be placed from the vallecula to the tongue tip, on the mandible, the origin of the short tendon and the hyoid.  This opens up a whole new view on lingual articulation.  For example, the contribution of the hyoid to speech articulation can be investigated. The distance between the short tendon and points along the tongue directly reflect the contractions of genioglossus neuro muscular compartments.

Research into lingual anatomy funded by Articulate Instruments shows that the tongue body position and shape is controlled by 5 independently innervated compartments.  Study of the timing and extent of these contractions promises a deeper understanding of speech motor control.  In 2022, Articulate Instruments developed these measures and termed them glossometric measures.  See the Research and development section of this website for more information on Glossometry and Glossograms.  

Articulate instruments has also made open access a co-registered EMA/ultrasound dataset available for anyone to study.  See EMA Ultrasound co-registered dataset in the Open access resources menu.

These are tools provided to the research community to allow articulatory phonetic research to be carried out by any laboratory anywhere in the world.