| |
|
|
FACS describes appearance changes both in terms of movement of the face and configuration at the point when the movement reaches its apex (the point of maximum action). FACS was designed for measuring facial action where there is a motion record, film or video. It can also be applied to still photographs if there is a baseline photograph, showing no observable action, for comparison. The scoring of still photographs, however, is more difficult and vulnerable to more error than the scoring of movement records.
FACS should not be used without a record of behavior to allow repeated and, if it is a movement record, slow motion observation. Repeated viewing is necessary in order to reliably observe differences on the two sides of the face. It is also necessary to attend to changes in the brows and eyes, as well as in the lower part of the face. Even within a facial area repeated viewing is often needed if the action is complex, in order to determine how many different Action Units were involved. Slowed motion viewing is often necessary to observe certain changes in order to decide if an action occurred at all, or whether the changes are due to one or another Action Unit. Often small actions must be viewed repeatedly at real time, because in slowed motion the action may not be visible. We have not done a study to prove that reliability is poor when FACS is used at real time, but we have a wealth of experience which overwhelmingly suggests that FACS requires too many distinctions to be used for real time scoring without substantial modification.
FACS is easiest to apply if the camera provides a head-on view of the face being scored. It is possible to use FACS with a three-quarter view, and even with a profile view, but then the scoring is only of the visible side of the face. Shooting from above or below the subject may obscure some Action Units but our exploration of head and eye position does not suggest that many become undetectable. Chapter 9 in the Manual provides a simple system for scoring changes in head and eye position if the baseline position is head-on. It also describes the particular Action Units which become unscorable if there is a major deviation in up or down position of the head.
The more detail the picture of the face provides the less the strain in using FACS. It is hard to specify what the minimum size of the picture should be, since the lighting and quality of the lens, and other technical factors affect the resolution of the picture obtained. If one utilizes inexpensive 1/2 inch videotape equipment and cameras with normal room lighting, a very acceptable picture of the face can be obtained with one camera if the picture is shot from the waist up. If two cameras are multiplexed onto one videotape recorder, half the screen can be used for just the face, and the other half can show the full face and body. If 16mm film is used, and the picture shows the full face and body from toes to the top of the head, it will be a strain to score certain of the Action Units: 5, 7, 11, 23, 24 and 28, and slight level of activity for the others. These guidelines for video developed from experiences with 1970's era half-inch videotape. Today's higher quality analog and digital imaging equipment afford much greater flexibility and expanded possibilities.
If you have a doubt about whether your record of facial behavior is suitable for FACS scoring, or for scoring all of the Action Units in FACS, you can make a test. Make a recording of yourself in exactly the situation in which you intend to gather records. Be certain that the camera, lens, lighting, distance, etc., are exactly as they will be for your research records. Voluntarily perform each of the Action Units, starting with the smallest possible movement you can do, and slowly increasing it until you reach a very strong action. (In learning FACS you will be instructed how to make each Action Unit voluntarily, and over time you will become adept at performing these actions - this is further explained in Chapter 4 of this Guide.) At the moment you start each Action Unit performance you should make a visual signal (moving a hand) which will be easy to see. Inspection of this test videotape will show if all the Action Units are visible under your circumstances. Or, it may show that you need to eliminate certain Action Units from consideration in the scoring of your records.
In some situations where you intend to apply FACS to measure facial behavior it may be possible to obtain the cooperation of the subjects after the records are collected to obtain examples of the subjects' voluntary performance of certain AUs. These could be helpful in providing anchor points for scoring intensity of an action, including extreme versions of an AU. If your subject has difficulty performing an action you may want to show him the FACS film illustrating the AU, asking him to imitate it, perhaps observing his own performance in a mirror. Obviously you would do this after the records have been collected, and in many non experimental situations such performances would not be obtainable. FACS has been designed assuming that you do not have the benefit of such information on each person whose face is scored.
Note: Since this chapter was originally written, more detailed specifications for recording behaviors have been compiled, for example, see Harald G. Wallbott. Technical Appendix. Audiovisual recording: procedures, equipment, and troubleshooting. In K. R. Scherer & P. Ekman (Eds.) Handbook of Methods in Nonverbal Behavior Research. Cambridge: Cambridge University Press, 1982.
| |
|
|
