In audio-for-video work, there's a lot of discussion about matching audio to video. This is because in general, the video is produced first, then the soundtrack. As a result, various methods have been developed over the years to make sure that audio "hit points" line up with video, from using ACID style time stretching, to varying tempo subtly.

However, few people consider the reverse: Matching video to audio. For example, with a music video, the audio happens first, and the video needs to follow it. Or, consider an industrial video with narration. You might end up inserting stills, or small video segments, that tie to particular words in the narration. I've also edited videos based around demos of musical instruments that contain music, and again, the video has to tie to the audio—not the other way around.

Most of time, you can develop a skill for "reading the waveform" of audio signals. For example, Fig. 1 shows an audio track that starts off with four snare drum hits. The hits show up as clear, high-amplitude "spikes" compared to the rest of the waveform. You can zoom way in on the audio, place the time cursor at the precise beginning of the waveform, then "snap" your video still or segment to the cursor. In this particular example, the snare drum hits happen on eighth notes; a still is placed to coincide with each hit.

With narration, tying video to audio can be even easier. With a good narrator, there are often spaces between words that make it effortless to identify the start of a word; just line up the video with the beginning of the word.

But sometimes, life isn't simple. The music may not show well-defined beats, and the narration might be combined with background sounds that makes it hard to pick out individual words…so here's a workaround that can really help.

Vegas' marker function is intended primarily for navigating to, and identifying, various sections of a project; but it can also provide a sort of "tap tempo" function. As you listen to a piece of music, tap the keyboard's "M" key (which deposits a marker wherever you tap) on the beat, or for that matter, tap wherever you want an accent to occur. Then, you can simply line up the cursor with a marker, and snap the video to the cursor.

Fig. 2 shows an example of using markers to indicate the beat, along with stills that have been cut to the markers. Generally, I wouldn't advise trying to tap markers with any finer resolution than one beat (a quarter note)—it's very hard to hit the "M" key accurately enough to create something like 16th notes. You're better off just hitting on every beat, then "eyeballing" any sub-divisions. As the old saying goes, what you end up with will be "close enough for rock and roll."

Bio: Craig Anderton is Editor in Chief of www.harmony-central.com, and Executive Editor of EQ magazine. Check out his podcast at www.cyberears.com/index.php/show/audio/130.