It works the same as with any other stimuli. Define your <video> element(s) for the visual clips, define your <sound> element(s) for the audio instructions (use <video> for audio instructions too if you are using any compressed audio format such as mp3). Have your <trial>(s) display both:
/ stimulusframes = [1=myvideoclip, myaudioinstructions]
If you want to pair certain video items with certain audio instruction items, see the "How to display stimulus pairs" topic in the Inquisit documentation. Hope this helps.