You’re a filmmaker. You’ve got to deal with Terabytes of information, and you have to be able to deliver it at sometimes over 150 Megabytes a second. Ouch.
5 years ago that was a tall and expensive order for the independent filmmaker, but now, not only is it nearly possible with a single drive, but the price has plummeted to pennies per gigabyte.
You’re using video files. And while video wrappers have audio built in, the file architecture is set up to drive video without interruption. And the audio data footprint? A small inconsequential fraction of the overall data flow. But what happens when you need to play (or record) 50 stereo tracks at a time in a mix session? 100? 1,000? What changes have to be made in your storage system in order to keep everything running smoothly? Let’s debunk the myths about audio storage and get to the truth in this 5-part series.
“Audio Files Are Small. They Don’t Need Any Special Attention.”
“Why do we even care about audio files? They’re teeny little things which don’t really matter. Aren’t they
embedded in the movie files anyway?”
Well, I understand your feelings. So, consider this: a stereo audio file at 48kHz/24 bit needs 16.88MB/minute. I know. Video can do that a second. Sure. A stereo file is weenie. And, yes, it can be embedded in the video file without incident. Not even a blip. Not to mention what usually happens: the stereo file is compressed into something like AAC at 320 Kb/s which is 2.34 MB/minute. Yeah. Teenie. But since we NEVER, EVER use AAC/mp3 compression when we’re mixing, and since a reasonable mix has 50 tracks minimum – sometimes playing at the same time – that should have us sit up and take notice for a couple of reasons. 1. If, say, half of those tracks are stereo tracks, then you’re running 75 channels of audio at 48/24 and that comes out to 21.1 MB/s and 2. it means that this data is no longer coming from a single file, it means it’s coming from 75 different files (non-interleaved).
“Big Video Files Are Harder to Deliver Than Small Files.”
Without getting deeply into the theory of how hard drives work, it will suffice to say that reading many small files a second is much harder than reading 1 large file a second. It gets worse when multiple takes are used across these 50 channels, and the hard drive has to randomly access on top of it all. For you musicians, sample players are 100 fold worse, because each note you play may be triggering 30 different audio files at the same time, and those files are very small and require sequential reading. In other words, the quickest way to kill a hard drive is use it for musical sample playback. Not too many years ago I used to use GigaStudio. I’d kill a hard drive a year with it.
“Mark, are you telling me that audio is harder on a system than video? C’mon!”
Not exactly. Certainly not in any capacity outside of the final mix. But once there, you ARE going to want to consider a storage and delivery system which has a different configuration entirely than your video system.
“Mark, you’re crazy, but I like you anyway. And BTW, I’m using SSDs. So there!”
“SSDs Solve Everything”
Well, it IS nice to be rich and have all Solid State Drives. SSDs are pretty much the solution for most of the issues I’ll be talking about in this thread, as long as you’re running computers with SATA III or better. If you have a PC or Mac which is only SATA II, then your SSD is probably only going to be able to crank out 250 MB/s. Sure. Not bad. But nowhere near the capability of that drive, and if you think you’re going to be playing that uncompressed 6k RED footage just because it’s an SSD? Think again.
In the mix, audio file delivery and storage is no trifling matter, and in the next parts of this article, we’ll be discussing how to best get audio backed up, and used in a price-efficient manner.
Let us know your thoughts or Tweet about them!