...

For more in-depth information see →  NielsenNorman

...

General

Formulate a clear question before the experiment. This could be about the following things:

• Verification of the (affective) effect of sounds
• Verification of usability / functionality
  Verification of affordances in comparison with design hypotheses
• What are the relationships between the effect, the psychoacoustic dimensions and the physical/acoustic dimensions of a sound?

Measurements / Quantifying:

Simple Measurements

Threshold Measurements (indirect)

Here a certain parameter is continuously increased until subjects express a (provoked) perception (e.g. "now I hear the sound" or "now we hear the sound aggressively").

Ratio Scaling

Participants are asked to evaluate defined parameters of certain sounds according to a scale of strength.

Ranking Experiment

In the Ranking Experiment different sound variants are arranged according to one criterion. The "Random Access" method should be used, i.e. the test persons should be able to listen to sounds repeatedly and then gradually arrange them until the arrangement is right for them.

...

(An attempt to measure isolated dimensions continuously over a certain period of time is the "Continuous Response Digital Interface (CRDI)")

More complex (multidimensional) measurements of sound properties and effects

Semantic differential (uni- oder bipolar).

Bsp: Bipolar, as a list (from: H. Fastl (2006): Psychoacoustic Basis of Sound Quality Evaluation and Sound Engineering) > see LINK


...

Displayed as (monopolares) Network diagram (from P.
Susini,
 G.
Lemaitre, 
S.
McAdams: Psychological measurement for sound description and evaluation)

Requirements for meaningful results:

• The measured parameters / attributes must be understood by all (equally). "Ugly - Beautiful", for example, is not a very suitable dimension. Better is e.g. "Pleasant - Unpleasant".
• The parameters / attributes must be clearly defined.
• Do not use too many attributes in Semantic Ratings.
• If necessary, have attributes determined by test persons.
• The sounds to be compared must be comparable in terms of duration, pitch and volume.

Emotionsassessment

Emotions are often "measured" based on the  "Circumplex Model of Affect" by Russel. A common tool in the field of Human Computer Interaction is the "Self Assessment Mannequin" (SAM). This is particularly interesting because it also captures the feeling of "control" or "dominance". The design is deliberately kept very "neutral":

Interpreting, Observing

Observation (videographed), possibly with debriefing: Example of observation protocol: Observation Sheet Sonic Interaction_KP-DH_engl.pdf 

...