divergent media

Understanding Audio

Until recently, audio in the digital video world was about as simple as can be. Whether you were shooting DV or DigiBeta, DVCProHD or Digital8, you were recording your audio as something called Linear PCM. As with everything else in this mixed up, modern world, things have gotten a bit more complicated.

First let’s back up and talk PCM. It stands for Pulse Code Modulation, and it’s just about the simplest way to store digital audio. If you’re familiar with digital audio, you probably know that there are two main values that get discussed – bit depth and sample rate. An audio CD is 44.1khz at 16bits. That means that there are 44,100 samples per second, and each sample contains a 16 bit value. PCM audio is essentially just a string of these samples, played at the proscribed sample rate. You can even do a little math to confirm it for yourself. 44100 times 16, divided by 8 gives you 88200 bytes per second. Double that for stereo, divide by 1024 and you’ve got 172 kilobytes per second for stereo CD audio. DV audio runs at 48khz, so the math comes to 1500 kilobits per second, or 187 kilobytes per second. Math is fun!

Linear PCM, or LPCM, is still the preferred format for digital audio. You can seek to any point in an audio stream easily, it takes almost no computational power to decode, and the quality is limited only by the sample rate and bitrate. So why not use it? Well, just like the transition from intraframe to interframe codecs, storage efficiency meant that raw audio had to go.

HDV was the first major format to use something other than LPCM audio to the digital workflow. HDV uses an audio format called MPEG1-Layer2 (MP2). Similar to the better known MP3 format, MP2 uses acoustic modeling to remove audio that is likely to be imperceptible. Just like with MP3 files, the quality of MP2 audio is largely based on the bitrate it is encoded at. HDV uses a generous 384 kilobits per second for a stereo recording, resulting in audio that most listeners cannot distinguish from an uncompressed source. Indeed, with most HDV camcorders, the analog portions of the audio signal chain will be the limiting factor. But, that 384 kilobits per second audio is four times less than the equivalent linear PCM recording. When you’re trying to cram 1080i60 into 25mbps, every extra bit helps.

AVCHD introduced a second compressed format, AC3. Many home theater aficionados know AC3 as the format that delivers 5.1 channel surround sound on DVDs. AC3 is an efficient way of encoding and compressing surround sound data using substantially less data than 6 discrete linear PCM channels would require. In addition to acoustic modeling, AC3 is able to recognize redundant data between channels, to achieve increased bit savings. AVCHD uses AC3 in two different configurations, depending on the camera. Some cameras record full 5.1 surround sound, using an array of microphones. Other cameras record in a configuration called 2/0, which essentially means stereo recording.

So what about post production? Well, most nonlinear editing applications, including Final Cut Pro, deal best when presented with LPCM audio. While some can deal with other formats, you’re likely to face audio renders or glitches. For this reason, ClipWrap defaults to converting all audio to LPCM. If your workflow requires retaining the original audio, untouched, we also provide a preference for passing through the audio from your source. However, we strongly recommend avoiding this option unless you have an explicit reason.

When it comes to converting AC3 to linear PCM, ClipWrap defaults to doing a multichannel mixdown to stereo. We do this because most developmental editors aren’t looking to work with multichannel audio at this point, and because most AVCHD cameras don’t record meaningful data in their surround channels. If you’d like to play with extracting six discrete channels of linear PCM audio from your AC3 soundtrack, you can toggle that setting in the audio tab of the ClipWrap preferences.

So that’s audio. It can sometimes feel like a bit of an afterthought in the world of digital video. But, as with most things, a little understanding goes a long ways.

ClipWrap