In June 2007 Julian McGlashan, Cathrine Sadolin and Henrik Kjelin performed an endoscopy study at CVI in Copenhagen.

Twenty-one singers (ten males and ten females) trained in the Complete Vocal Technique were recruited and asked to produce a sustained vowel in each mode. Each singer was instructed to produce the mode as perfect (near to the centre of the mode) as possible with minimal additional effects, twang and extremes of sound colour.  The larynx was imaged using a videonasoendoscopic camera system (OTVS7 camera and ENFV2 videoscope from Olympus/Keymed) and the Laryngostrobe system (from Laryngograph Ltd) used for image capture.

Three subjects were excluded as they could either not tolerate the examination (1) or the image quality was not adequate for interpretation (2).

When looking at these videos certain patterns were seen in the laryngeal gestures within each mode. A still image for each of the modes was produced from the videos. The still images were chosen by listening to the sound without seeing the images and extracting the stills when the sound was as close to the centre of the mode as possible.

Each set of five (Neutral with and without air were viewed separately) video images were analysed by Julian McGlashan and Cathrine Sadolin and the appearance and relationship between the key anatomical features was agreed by consensus and documented. The consistent features were identified and schematic drawings produced with descriptive and explanatory text to aid pattern recognition.

Then these laryngeal gestures were schematized by the most obvious gestures (the ’first glance’) and then further examined and categorized into what could be seen at a ‘second glance’.

We gave the various parts in the vocal tract levels in order to identify and specify on which levels the various changes take place. The levels also make it easier to communicate where the changes take place. On the stills from the footage only the first 3-4 levels are seen.

To indentify the structures on the endoscopic images see the anatomy of the larynx.


The ‘first glance’ laryngeal gestures in Edge on the levels are:

Illu E 1












Level 1

Level 2

  • Width of the vocal folds is encroached on by false folds even more
  • False folds are less retracted (straighter) giving a more square appearance

Level 3

  • There is less of a gap and more contact between arytenoids and posterior pharyngeal wall. The larynx is generally elevated.

In everyday language: 

The opening of the larynx is even more narrowed, and on top the cuneiform cartilages are rolled in even more. This makes it hard to see the vocal folds.


The ‘second glance’ laryngeal gestures in Edge on the levels are:

Illu E 2












Level 3

  • Forward rotation of arytenoid apices (apexes)
  • Petiole of epiglottis more posterior
  • Angle between Aryepiglottic folds and arytenoid apex is flatter.

Level 4

In everyday language: 

The narrowing creates an angle between the arytenoids and the aryepiglottic fold, and also an even more acute angle between the aryepiglottic fold and epiglottis.

The area between the back wall and the larynx becomes very small and often it disappears altogether.

The larynx is raised to a higher position (you can see that it gets closer to the camera)


The laryngeal gestures in Edge without drawings

Illu E 3












This information comes from a study Visual Vocal Mode Test Study on stills, with the title ‘Laryngeal gestures and Laryngograph data associated with the four vocal modes as described in the Complete Vocal Technique method of singing teaching’. This study was presented by Cathrine Sadolin and Julian McGlashan at BVA ‘Choice for voice’ conference in London, England, 2010.