Emagic Zap

Paul White tries Emagic's new audio file compression utility while balancing the implications of time and space.

Digital audio recorded to hard disk is an established reality — it's just that it's an established reality that also happens to take up an enormous amount of disk space! Emagic — of Logic fame — have just launched a simple utility for the Macintosh (a Windows version is in the pipeline, but is not yet available) called ZAP (Zero‑loss Audio Packer), and while it won't solve the problem completely, it does enable audio files to be compressed to around half their original size in a totally loss‑free manner. This is obviously a major saving in disk space — so what's the catch?

The answer is that there is no catch in ZAP's compression process — this isn't like the data‑changing compression systems used in DCC or MiniDisc, where less important information is discarded; it's more like the audio equivalent of ZipIt, Disk Doubler or Stuffit, programs used to compress computer files or applications. What you get back when you un‑ZAP (for want of a better term) is exactly the same file you started with. Though ZAP is intended mainly for users of Logic Audio, it can compress and decompress files in Sound Designer II format (mono or interleaved stereo), 16‑bit, uncompressed AIFF and both mono and stereo 16‑bit WAV files.

Around 1.5Mb of free RAM is needed to run ZAP, which will work on any 68020 Mac or faster, provided the operating system is at least System 7.0. A Power Mac native version of the code is included in the program. Installation is from floppy, and this being the music business, the software is protected by a disk install system — you can install up to two times from the master disk, and you can also deinstall and restore one installer to the disk if you need to reformat or change drives. Once installed, the software is authorised to run from one specific hard drive.

Using ZAP

ZAP's user preference box lets you determine whether the original files should be trashed after compression or not, and whether or not you want your compressed files to be self‑expanding — which means that you don't then need ZAP to restore them to their former glory. So, why not make all files self‑extracting? The answer is that a self‑extracting file contains a mini‑program to unravel it, so the file size is around 30K larger than it would be if it were not self‑extracting. On large files, this isn't much of a penalty, but if you're into saving hundreds of short files, it might be significant.

Self‑extracting files made using the Mac version of ZAP will only be able to self‑extract when used on a Mac, though you can decode them if you un‑ZAP the file from within ZAP Windows when this becomes available (as opposed to just double‑clicking on the file). The easiest way to compress a file, according to the manual, is to drag the file icon onto the ZAP icon — compression should then follow automatically. You can also go via the menu within ZAP if that makes you feel better. In fact, I couldn't get 'drag and drop' to work at all, though working from the menu presented no problems. ZAP can also compress multiple files, either individually or as a single archive, depending on the preference settings. A neat feature is that ZAP works in the background, so you don't have to stop using the computer to let it do its thing. If you do compress a bunch of files as a single archive, the downside is that you have to uncompress the whole archive to get them back again, whereas if you choose to have the files compressed on an individual basis, they can also be retrieved individually. If you select the audible warning to let you know ZAP has finished unpacking a file, you are greeted by a very Wayne's World‑ish 'Excellent!'.

Inside ZAP

ZAP works using a principle called 'non‑linear redundancy elimination', which involves a statistical analysis of the audio material to locate redundant material. The redundant material isn't discarded, but rather is digitally 'described' in more economic terms than simply storing the number for each sample. Totally uncorrelated material, such as white noise, does not tend to get any smaller if ZAP‑ped, but more coherent waveforms can be reduced inside by up to 50% or so. Speech with pauses may be often be compressed beyond a 60% saving. Because the process depends on the nature of the sound file data, it isn't possible to say exactly how much space you'll save until you've done the compression, but then this is also true of the computer data compression systems used to compress graphics or text files. If a ZAP file becomes corrupted in any way, ZAP will announce the fact when you try to decompress it — you won't be left working with a dodgy file without being informed.

Reasons To ZAP

One of the main reasons to ZAP files is so that they can be archived without taking up so much space, but it's also beneficial to ZAP audio files destined for transmission over the Internet. With the proviso that self‑extracting Mac‑created archives can't self‑extract under Windows, this could save a lot of time when transmitting audio files that must not be compromised in quality. It's useful to note that the self‑extraction code is written for 68xxx machines to ensure maximum compatibility, while the native ZAP application works rather faster on Power Macs. For this reason, Power Mac owners are better off opening files from within ZAP, rather than relying on the 'double‑click to self extract' operation.

Impressions Of ZAP

One of the disconcerting things about ZAP is that it estimates the time the operation will take, and then seems to keep amending it — upwards. On my admittedly old 68040 Centris 650, a four‑minute stereo SDII file came up with an initial estimate of around 15 minutes, but after five minutes or so had elapsed, this had increased to around 23 minutes, and showed little sign of going down again! Eventually, I went into the lounge to watch Drop the Dead Donkey — well, you have to, don't you — and when I came back, the job was done. I must confess that I didn't have the patience to do it again to get an idea of exactly how long it took — suffice it to say that it's long enough to be boring. UnZAP‑ping is slightly faster, but still takes significantly longer than simply copying a file of the same size — around 20 to 25 minutes in this case. On a stereo file, I saved around 35% on the size of the original file.

There's no doubt that, other than the puzzling reluctance to 'drag and drop', ZAP does as it claims, and the reconstituted file is an exact duplicate of the original — but unless you have one of the latest light‑speed Macs, I feel the process is simply too slow for most situations. Batch processing for long‑term archiving is feasible, assuming you have something else to do in the meantime (background processing slows ZAP down even further), but for use during sessions, a super‑fast Mac is a necessity, not a luxury. I accept that the amount of processing taking place means ZAP‑ping isn't a trivial task, but in my own situation, where I'm working on a Mac that's starting to show its age, and where my files are often of album length, I think I prefer to waste disk space rather than time.