In general, we remember best if we see take in information presented in a multisensory fashion. In fact, we may not even have to be aware that we're taking multisensory information in - but we'll still show that we remember it better.
If you're interested in reading up more on the research basis of memory, check out the papers below. This is the general trend of what things will be remembered accurately: pictures + congruent sound > pictures > colored text > black and white text.
It tells us why computer-based learning holds so much progress. Moving images also improve memory, but there are also saturation effects if too much information is presented at one time.
The importance of multisensory learning for 'automatic' learning is why even minor processing problems in visual, auditory, or touch / balance modalities can cause havoc in children. We negotiate and learn from our environment with an integration of all the senses - perturbation in one sense affects the others, and in fact, an unreliable sense (like vision or hearing) is worse that if that sense were completely eliminated. For those who are familiar with neurology, it's why a stroke patient's auditory neglect often improves if they are blindfolded. Irregular or out of place signals are harder to understand than complete deprivation.
Multisensory Memory without Awareness
Synergy Between Kinetic Vision & Touch
Multisensory Integration of Faces and Voices in the Monkey