ST Sound Digitizer
1043 Stierlin Road
Mountain View, CA 94040
Medium or High Resolution
by Andy Eddy
With the memory capacity, speed and power of the ST, many tasks that were previously delegated strictly to large-scale computer systems are crossing over to this "side of the tracks." Software for CAD/CAM/CAE (Computer Aided Design/Modeling/Engineering), high-quality graphics and animation, and digitizing (both video and audio) are appearing in great numbers for Atari users.
Digitized sound samples—such as those that appear at key moments in Firebird Software's Starglider, as well as many new entertainment releases for the ST—involve reading the analog input signal and converting the voltage level to numbers for storage, enhancement and playback by the computer. This is the same technology behind compact disks, though CDs are a much better medium for large storage of data. That's the problem with digitizing: For a good quality reproduction, you have to have a lot of memory for your sample. The more frequently your sound is sampled, the better the replication.
Navarone has licensed the Hippo Sound Digitizer from the now-defunct Hippopotamus Software, and they've got a pretty good product on their hands. It allows a sampling frequency between 1 KHz and 64 KHz (between 1,000 and 64,000 samples per second), the high number being the highest quality "recording." Of course, as we said earlier, on a 1040ST (with 1 megabyte of RAM) you are limited to an approximately 12-second sample at 64 KHz, but that figure will increase with a lowered sample rate; size is a trade-off for quality.
The cartridge that plugs into the ST cartridge port is exactly the same as Navarone's video digitizer and clock cart in size; they use the same casing for all their hardware add-ons to trim production costs. To suit its electronic requirements it contains two 1/8-inch phone jacks for Line In and Line Out, as well as two thumb wheel potentiometers for adjusting those levels. These are best set-up by watching the waveform on the real-time O-Scope (Oscilloscope) display, selected from the menu bar at the screen bottom.
On top of that, you'll have to equip yourself with a playback source (like an amplifier or stereo system), a microphone (if you want to record sounds or voices) and cables (one cable from your input source and one for your output to your playback device), because the package doesn't include them.
Conveniently, the manual briefly describes various microphones, adapters and cables that are available, along with their Radio Shack part numbers.
The ST Sound Digitizer software—which runs in either medium or high resolution—has two screens other than the above-mentioned O-Scope display: the Rack screen (which we'll get to later) and the Command screen, which is where the majority of the work is accomplished. Here you set the size of the sample (in seconds), the rate of sampling (in KHz), initiate the digitizing process and manipulate the end result in a variety of ways. They've even set you up with a folder full of sounds, like drum types and basic waveform types.
To find out where you stand at any given moment, you get a bar chart that displays the amount of available RAM broken down in three ways: space taken up by the sample, the space used by the copy buffer and the RAM left to use. Hitting the Stats button gives you a readout of the vital figures: the length of the sample (or, as you'll see later, a marked section) in seconds and number of samples (directly translated to RAM used), the sample rate and what percentage your high and low peaks are.
At the grass roots you can adjust the volume and overall level of the sample. You can mark a block and cut, copy, insert and replace it within the main body of the sound, in addition to saving and loading blocks of entire sound to and from disk. They've even slipped in a couple of commands for zooming in or out on the marked section of the sample to make your edits more precise. Overall, this cut-and-paste capability allows you to edit sounds together for whatever purpose you have, be it commercial applications or just for fun, giving you a low-cost recording and editing studio.
You can also alter the sound itself with the Reverse, Squeeze, and Stretch functions. The Reverse command is easy enough to understand: It will take the entire sound (or a marked block) and re-display it backwards. All those folks complaining of "backward masking" (those inverted messages inserted in musical passages that some claim are really evil, subconscious suggestions) will have an easier time of proving their claims with this tool.
Selecting Squeeze and Stretch shortens or lengthens a sound and changes its pitch equal to one musical note each time you enact them. If you double-click on either of these buttons, you can enter directly the number of times you want the effect to be processed. Unfortunately, there's no provision for changing the length of the sound without altering the pitch, so voice samples that are stretched or squeezed will tend to sound unnatural with regards to pitch, like the effect that spinning a turntable faster or slower will have on its playback.
Some hi-tech solutions have been devised for "compressing" a sound—in other words, changing the time of a sound without affecting the pitch—and some commercial producers, among others, are using it to cram more information into a limited period. I'm not sure what effect adding this procedure would have had on the cost of Navarone's unit, but it would have made the package that much more powerful by letting the user change the length of a sample without an audible difference.
There are a few more sound-processing tools available to you. The first involves a Mix command which will take a block and overlay it onto another area of the sample. Like "over-dubbing" in a recording studio, this procedure combines the two signals into one. It'll make one voice become many voices or make various digitized sounds appear simultaneously created.
The other enhancements are on the Rack screen: namely, Echo and Reverb. Echo is a repeating of the initial sound (like yelling in a valley and hearing the sound come back) and reverb is a very fast echo (similar to singing in a tiled room and hearing the sound reverberate rapidly in the small area). You can add these effects to an existing sound in memory or use it to enhance a sound in memory or use it to enhance a sound as it is digitized.
The rates of the effects can be controlled as well. I found that these two enchancements, as well as the Mix command, severely alter the original sound's clarity, bringing about a fuzzier quality not heard at the start. For this reason, you should save the original sound, as you may find the result from these effects to be detrimental enough to be undesirable for your application.
Aside from digitizing a sound from an audio source, you can create your own by drawing the waveform and envelope on the screen. I'm not sure what use this is as it's ten times easier to record a sound with a mike or other source. If you choose to do things this way, you only have to click on the Draw Wave/Draw Env button on the Command screen and use the onscreen cursor to create the desired waveform and enveloping.
Another application that the ST Sound Digitizer comes equipped with—though, as you'll see, it's some-what limited and I can't really see where you'd put it to use—allows you to plug a MIDI keyboard into the ST's MIDI ports for playback of sounds. I thought it strange that the manual would have you hooking a MIDI cable from the ST's MIDI IN port to the keyboard's MIDI OUT port because I figured that the digitizer software would send the sound to the keyboard. Reading further informs you that the software "does not provide any kind of sequencing or downloading function for your keyboard."
This is where the limitation comes in. Your MIDI keyboard will only trigger the ST, which holds the sample in its memory. I was hoping that it would be playable through the keyboard's electronics, like you have with a MIDI "patch" program (a utility that lets you move synthesizer sounds from the computer to a MIDI instrument and vice versa, as well as store them on a computer disk). The sound varies in pitch with the note you play (middle C is the key that plays the sound at its normal rate and is the keyboard's "pivot point"), so you can "play" on the ST, but the output is only monophonic (playing one note simultaneously), regardless of the capability of the keyboard. I doubt that this option would be usable in a performance. And again, you have to provide the cable.
Outside of the actual digitizer and software, the manual was found to be lacking in many respects. Granted, this package doesn't require much in the way of instruction; it's reasonably simple to use. What irked me is that some of the diagrams don't match their on-screen counterparts. An example of this is the Rack screen and its corresponding manual illustration on page 10 of the booklet. Similarly, on page 11, the Command screen buttons aren't labeled in the manual, and the overall layout varies a bit from the actual graphic. On top of that, much is left to the user's discovery, as the booklet is a bit thin on some points.
On the other hand, they've gone to quite a lot of trouble to include a good deal of technical data for those who'd like to utilize the digitizer in their programming efforts or those who'd like to know how it ticks. They've detailed how it works, the file format, even how to read the various voltages that are output by the cartridge.
The ST Sound Digitizer is none-too-high-priced and a neat piece of work on the whole. Complex works will require some other means as the length and quality of the sample are very restricted by memory usage, and mixing of multiple sounds is not the cleanest it could be. But if you'd like to build a library of sound effects for use in programs, or record and edit advertisements for your local radio station, this product fits the bill.