Teachers have acknowledged the richer learning environment and interactivity of multimedia teaching, its flexibility to different learning styles, and learner control that allows the learner to fully engage in the learning process. However, they still have problems in courseware design because their work is mainly centered on exercises and not on what the machine can do best. This is why a new audio segmenting device, called Virtual Recorder, has been derived from the LAVAC (Laboratoire Audio-Visuel Actif-Comparatif) toolkit to allow them to use videos. The video sequencer can complete real-time automatic segmenting of sound and images and automatically insert an answering time span after each sequence. Coupled to IBM ViaVoice, the teacher can speak during this time span to create the transcript from which the necessary textual help will be derived for easier aural comprehension. A 5-minute video requires no more than a few minutes' work from the teacher to produce a 2-hour student session. (Author/MES)