Magazine Archive

Home -> Magazines -> Issues -> Articles in this issue -> View

Plasmec Systems ADAS

Hard Disk Recording System

Direct-to-disk digital recording comes of age on the Atari ST computer. Vic Lennard checks out the system which may change your life.

Like synthesis and sampling before it, direct-to-hard disk recording is coming out of professionals' studios and into those of lesser musos.

From sound effects to pop music, sampling has become an indispensable part of the '90s recording studio. Yet in spite of its revolutionary nature, the sampler is not without its shortcomings. With hardware samplers, long sampling times equate to large amounts of memory (RAM) which is relatively expensive - relative to running directly to a computer hard drive, that is. A megabyte of RAM costs around £40, depending on the computer, while a megabyte of hard drive space is less than a quarter of this cost.

Memory cost is not the only reason to consider a direct-to-hard-drive sampling system. As anyone who has switched to a computer sequencer from one of the hardware variety is aware, you should never underestimate the power of having a monitor to see clearly what you are doing. Post-production work with cue sheets and the like profits from this visual approach, and this is the very area in which direct-to-disk recording provides most benefit. Add to this the fact that many hard drives allow back-up to tape streamer or DAT and the reasons to invest in such a system are obvious.

The only problem is a financial one. Direct-to-disk systems such as Digidesign's Sound Tools on the Atari and Mac, and Hybrid Arts' ADAP on the Atari are expensive to buy and generally require a separate computer additional to the one running your sequencer. Plasmec have decided that such a unit can be manufactured at an affordable price and ADAS, designed by Mike Lynch of D2D, is the result of their research.


ADAS IS A stereo, 16-bit, direct-to-disk recording system running on the Atari ST. It's capable of digitising analogue audio and transferring it directly to hard drive; the ST's RAM isn't used for recording purposes, so the same computer may be used to run a sequencing program concurrently. All you need is an Atari ST, a hard drive and the ADAS hardware. What makes this interesting is that the ADAS package retails at £849 inclusive of VAT.

ADAS software comes in two forms; the main ADAS program, which allows you to record, edit and work with a cue list, and the ADAS desk accessory for recording and playing back samples while using the same Atari for a sequencer program.

ADAS itself is a plastic box measuring about 9" square. The front panel has four LEDs; Power (red), Disk active (green), Recording (red), and Digital board in use (yellow). The rear panel offers four sets of connections; a short lead connects ADAS to the hard drive (DMA) port on the Atari while a separate lead then continues the signal path to the hard drive itself. Power is via an external PSU and a six-pin DIN socket. A pair of stereo quarter-inch jacks for the audio in/out and a pair of phono sockets for the SPDIF digital in/out complete the guided tour. It's fair to say that you couldn't fit anything else on the rear panel.

Which make of hard drive you use is up to you, but the time taken to read/write to the drive needs to be reasonably short. For example, trying an Atari Megafile 60 was unsuccessful until the DMA lead connecting it to the ADAS hardware was significantly shortened. On the other hand, the DAC 44Meg removable media system worked fine. Whichever drive you use, each 10Meg gives you about one minute of stereo recording time.

As ADAS sits in-line between the Atari and the hard drive, you can use your main ST hard drive for recording but this can lead to problems. If the same disk space is being used for writing to and reading from on a daily basis, the data on the disk will become fragmented - data cannot be saved on continuous disk sectors and so are written at different places on the disk. Writing to such a disk entails movement of the read/write heads which slows down the access time, causing a situation in which ADAS will refuse to work. If the same hard drive is going to be used, it would be better to have a separate partition specifically for ADAS.

The other problem which may occur is incompatibility between the program and some STs. Atari have gone through at least five different operating systems in their computers and some are flakier than others. Plasmec have spoken with Atari and are compiling a list of problem serial numbers and possible solutions.


ON LOADING THE program (review v1.3 - v1.4 current at time of publication) you are presented with a vertical strip of icons on the left-hand side of the screen and the usual Atari menu bar along the top. Some of the icons duplicate menu selections, and various functions have keyboard equivalents to make life easier.

To make a recording, open the Transport Control window (the cassette-style icon). This has the usual transport commands as well as the sampling rate selector (48kHz, 44.1kHz and 32kHz), start time and length of sample. A MIDI Time Code stamp is recorded with a sample as standard so that the correct position within a sample can be chased when playing back synchronised to MTC. A click on the File box brings up the standard Atari file selector, at which point you type in a new file name, set the recording length and monitor the incoming audio, which passes through the A/D and D/A converters. If you hear distortion, then the incoming signal has to be reduced - there's no setting of gain option within ADAS. Also, there are no "meters" on-screen to show the approximate input level, which means that you could go into distortion part way through a recording and have to retake. Once the settings are satisfactory, click on the record button. The main Counter ticks over as the recording takes place and, at the end of the recording, the punch in and out counter shows the start and end points, ready for you to retake if necessary.

You can immediately play back the sample from its start, or from a particular point by fast forwarding or rewinding to that specific time on the main counter. Punch in and out are available; the punch times are set and then recording takes place as usual. The only problem here is that you can't hear the material before and after the punch points without recording first and then playing back - there isn't the equivalent of changing the monitoring from tape to input as on a tape recorder. That said, punching in and out is achieved very smoothly - without glitches or clicks.


IF YOU ENTER the Edit window and load the recording from disk, the waveform appears in a new window on screen. The stereo source is displayed as a pair of mono, block waveforms, one above the other, with the start and end times of the waveform displayed. To hear the entire sample, click on the loudspeaker icon; to stop at any time, press both shift keys on the keyboard.

Zooming in and out by a factor of two can take place in both horizontal (X-axis) and vertical (Y-axis) directions by using the "lollipop"-style icons. Taking the latter first, the only reason to want to zoom vertically is to see the waveform more clearly when its amplitude is small. However, if its amplitude is that small, it would be better to re-record to get a better signal-to-noise ratio, especially as ADAS doesn't have the ability to normalise a sample - this is where the sample's greatest amplitude is found, increased to the highest amplitude possible and the rest of the waveform scaled accordingly. It would have been better for this window to be resizeable to display the vertical aspect of the waveform more clearly, because although there are inches of spare space around the waveform window, you can't make use of it.

Horizontal zooming is essential for finding zero-crossing points (where the waveform cuts through X axis). These are used to ensure that splicing together different samples doesn't give audible glitches. To zoom in on an area of a sample, click on the start or end point and drag out the area you want to hear. If you were editing on a tape-based system, you could hear the audio source being edited as it passed the playback head. With a digital system, you have the option of "scrubbing" the sample. This lets you hear a rough representation of the digital data with an option of either hearing a single shot of each bit as you drag across, or continuously repeating about half a dozen times a second while the mouse button is being held down. On the review version playback involved hacking backwards and forwards to the loudspeaker icon, but v1.4 facilitates playback from the Atari's Help key - a significant improvement.

The dragged-out area is shown by reversed graphics; the left-hand end of the area is the Start marker and the right-hand end the End marker. If you now zoom in, the inverted area on screen stays in the same place on screen as opposed to time-wise, which is a bit pointless - even the times attributed to the start and end markers on-screen change. On top of this, there are no vertical and horizontal scales, leaving you uncertain of where you are within the sample.

What you can do is to drag out an area and then click on either the "M1" or "M2" icons to zoom in on the start or end markers respectively, which then appear in the centre of the edit window. Continued clicking on either of these zooms in by a factor of two, and the inverted area then stays static time-wise. Let's say that we want to fine-tune the start point. Zoom in on M1 until we can see where the start point should be, hold down the shift key and drag the start marker to that point. What happens if the point you want to go to is beyond the edge of the screen - highly possible if you have zoomed in many times? We can use the Remember Markers and Recall Markers icons; if the markers can be shown within the current zoom factor, they are; alternatively, if we've zoomed in beyond the point where this is possible, the start and end markers are at either end of the edit window, in which case the current resolution of the waveform on screen may not be high enough to see the point that you want to set for the start or end. Either way, we have to scroll with the horizontal scroll bar beneath the window, because the waveform doesn't automatically scroll when you drag to the edge of it. It would be better if it did. Having selected a part of the recording, the markers can be saved as a Segment to a disk file for use on the Cue List screen, and such a file only takes up 120 bytes as long as a different file name from the original recording is used. If the same name is used, the area outside the markers is permanently erased. Fortunately, you're given a prompt warning of this.

Currently no editing facilities, such as reversing of regions, maximising of waveform, fade ins and outs, are available. It's also not possible to treat the left and right sides of the stereo image separately and there are no cut, copy and paste options or the ability to splice samples together, although this latter facility can effectively take place in the cue list.

The lack of processing power in an Atari ST is self-evident when it comes to rapidly re-drawing and scrolling screens. Consequently, ADAS offers you the option to view the waveform as a horizontal piece of tape, so speeding up re-draws dramatically.


A CUE SHEET is a list of events triggered (or "cued") at specific times. In the case of ADAS, the cue sheet is the playing back of recordings, or segments of recordings, at allocated times.

The Cue screen is divided into three parts. The bulk of the screen displays the cue number, name, L/R levels, start and end times. Above this are five boxes giving the file name, start time for cue list, main clock counter, clock source (internal or MTC) and start mode. The latter lets you choose whether the cue list starts at the cue list start time (Absolute), so leaving a gap of silence before the first entry, or starts immediately with the first entry (File). Because the start time for a segment is unlikely to be at, testing a cue list takes place in File mode, while working with the final product, perhaps locked to a sequencer, would take place in Absolute mode because the gap before the first cue would correctly position that first cue. Finally, the right-hand side of the screen lets you add, delete, edit and play the entries in the cue sheet and load/save to disk.

Creating a cue list is simple. A click on the Add box brings up the usual Atari file selector from which you select the first cue. The information box which then appears tells you the length of this file and the start time, which can be altered if you wish. The left and right levels can be altered, and the Insert mode set, for which there are three options. Delete places the entry in the cue list at its start time, and deletes any existing cues whose start times fall before the end time of the entry. Shift again places the entry at its start time, but then shifts all successive cues which would otherwise be written over. Close up sets the start time of the entry as that of the nearest end time of an existing cue and then moves successive cues back to close any gaps. If an existing cue is highlighted in the cue list, then the next entry you make automatically has the start time set to the end time of this cue. Editing cues again brings up the same information box so you get a second bite at the cherry. Deleting a cue gives you the choice of simply leaving a space where that cue was, or closing up the gap by moving successive cues back. It would be useful to be able to operate the Add and Delete functions via the Insert and Delete keys on the ST, to be able to choose to edit an entry by double clicking on it or to move up and down the cue list by using the cursor keys. Obviously, key equivalents of this nature save the time taken to move monotonously back and forth across the screen.

The ability to change the levels of recordings is limited, due to the fact that you have to leave a gap of a few frames between the end of one cue and the start of the next. This is because ADAS changes the digital data as opposed to using a VCA. Consequently, level changes can be used with, say, discrete sound effects but can't be used in the context of slicing up a song and reconstructing it.

To move several cues by the same, relative, time, you can edit the start time of the first cue and then use the Shift option to make the rest follow suit. Unfortunately, you can't do multiple deletes, as you can only select one cue sheet entry at a time. Why would you want to do this? Well, let's say that you've set the markers for a segment at 04.00 and 08.00 seconds and want this to repeat four times. Save the segment, go to the cue sheet and enter it four times. Fine. Now let's say that on playback you decide that the segment is slightly too long. Back to the edit screen, change the end marker to 07.24 (assuming that 25 frames per second is being used) and save the segment with the same name. On entering the cue sheet, the end marker times haven't changed, which means that you have to delete the four cues and re-enter them. This is a definite failing of the cue screen and makes it a little user-unfriendly.

"When you're looking at a direct-to-disk system which operates with the ST and costs under 900 quid, what do you compare it with?"

If you're running ADAS on a separate computer from your sequencer, you can lock the cue sheet to your sequencer using MTC, as long as the software you are using supports this - Steinberg and Passport software does on the ST. When you record a sample, the timing is "stamped" onto the sample along its length which means that you can begin playback at a place other than the start of the sample - similar to using MIDI Song Position Pointer within a sequencer. Unfortunately, it takes around two seconds for an entry in the cue sheet to start playing upon receipt of MTC, which makes it very difficult to start at a specific position. Also, 24, 25 and 30fps formats are supported, but what happened to 30 Drop Frame? It's true that this format is rarely used, but it should still be there.

Another feature that should be included in a cue list is the ability to crossfade from one cue to another. This is where the level of an outgoing cue is reduced as that cue finishes and the level of the incoming cue increases at the same time, giving a smooth changeover of cues. While it is possible to achieve glitch-free changeovers without this function, zero-crossing points invariably have to be used when the segments are saved and this takes a lot more effort to accomplish.


AS PREVIOUSLY NOTED, you can run ADAS and a sequencer on the same ST. To help in this area, there is a desk accessory which lets you access the ADAS tracks recorded to hard drive and even to record and play back while working with the sequencer.

Both Steinberg and C-Lab have collaborated with Plasmec to establish an internal MIDI link between digital recorder and sequencer. This means that you can trigger samples from MIDI data saved within any of these manufacturer's sequencers. With C-Lab, you assign note information to Port D, while with Steinberg you use the ADAS driver which they have created. But this is running before we can walk...

Taking a look at the accessory, you're initially presented with the first of six Control Cards, each of which have six slots for samples. Unfortunately, you have to type in the path of each sample, something like this: C:\AUDIO\TRYITOUT.TRK.

First you have to remember the partition, folder and file name. Secondly, if you select a segment, you're told that the file isn't sample data. It would be better to be able to click on the file space and be presented with the usual ST file selector from which you could select your file, although this would make it awkward to access samples in different partitions. Also, segment and sample files should have different names if you can't access the main file information by selecting the segment file.

For each slot, you can set a MIDI channel, MIDI note number and whether ADAS is to play back a sample or to record one. As there are six Control Cards available, each can be allocated a MIDI Program Change number, on different MIDI channels.

Assuming you're using either C-Lab or Steinberg software, the simplest way to use the accessory is to play back existing samples - set the path, assign the MIDI channel and note number, enter the note into the sequencer editing grid and that's it. The sample will play back when the MIDI note is reached. However, this only touches on the possibilities, because you can also use the desk accessory to trigger the recording of a sample, which is marvellous if you're sync'd to a tape machine and wish to sample multi-layered backing vocals which are then to be replayed at different places in the song. To do this, you set the slot up to record but immediately hit a problem - you can only use file names which relate to existing sample files, which means that you have to record a sample that you don't want via the ADAS program and then use that name.

If you're recording via the desk accessory, there are two ways to proceed: set the accessory up to record, record what you want to with your system in sync and then go to the accessory and change from record to play. The recording should then play back at the same time in the song that it was recorded. Alternatively, you can use two different MIDI note numbers for record and playback respectively and adjust the trigger note in the sequencer from one to the other.

A common situation in which ADAS might be used is the setting up of an audio loop over which MIDI equipment is played. The intention is that the accessory is used for the recording of the audio in situ but you can take a pre-recorded sample from the program, which you could do if you need to trim the start and end points precisely. However, there appear to be problems when you do this. A recorded piece of exactly eight seconds was placed into a slot on a Control Card. At a tempo of 120bpm on Cubase (each bar is two seconds long), the trigger note was written to a part of four bars in length and then copied a number of times.

Taking into consideration the fact that there is a time lapse between calling up the sample from hard drive and the audio actually starting (which Plasmec estimate at being around 200 milliseconds), this delay should be constant - the first sample plays late, and so do all others. Consequently you can set a negative track delay to compensate for this. In the situation I tried, there was an additional gap of around 200 milliseconds between the end of one sample and the start of the next. This makes it impossible to loop a continuous sample which has been taken from the ADAS program. Also, a second sample cannot start until the first has finished - surely it should be the other way round, with a second sample cutting off the first? The way things are currently, you have to be spot-on with the track delays; there's a fine dividing line between getting a gap and no second sample at all.

There are also problems in attempting to use the desk accessory to the same end. Using Cubase and the method of switching the Control Card entry from record to playback, the last 200mS or so that you hear being recorded is missing on playback, giving the same kind of gap. I can't see how this system can currently be used to provide seamless playback of a loop unless you use a cue sheet which cannot be triggered via the sequencer. One final problem is that of Cubase slowing down when handling a fair amount of MIDI data and then sending out a patch change to the Control Card; I'm talking about audible slowing down here. Apparently, this is the only situation where ADAS uses the ST's processor for playback. Perhaps the compromises necessary to run ADAS and a sequencer are too great. Bearing in mind that the desk accessory requires over 330 Kbytes of RAM to run, you can't operate either Cubase or Notator with ADAS on a 1Meg ST.

In the imminent software update (v1.4) one significant improvement is the provision of ten markers instead of two. This will mean that you can accurately slice up an audio piece, and in conjunction with a space bar playback/loop playback feature, ensure that playback from the cue sheet is a lot more accurate. Other features include real-time input level meters (monitoring only), DAT back-up software, and provision for saving personal configuration preferences (see sidebar for update list).

Plasmec intend to sort out the problem with the cue sheet edit updates, and the gap for level changes in the cue sheet (by using a small crossfade to get rid of the audible click). More interestingly, they're working with Steinberg to have ADAS included as a module within Avalon, which could then handle all of the editing which ADAS can't.


WHEN YOU'RE LOOKING at a direct-to-disk system which operates with the ST and costs under 900 quid, what do you compare it with? The obvious comparison is with Digidesign's Sound Tools (for the Mac), which costs over £2000 and has been down the difficult path of testing, bug-fixing and feature-adding for a couple of years, but the reality is that there's no competition. Even the Atari version of Sound Tools required a Mega4 ST (a lesser ST with a memory upgrade wouldn't do).


Atari keyboard playback function (Help key)
Repeat playback function (Control key)
Real-time input level metering
Real-time volume changes in Cuesheet
Recording in sync with timecode
Marker window listing times for editing
Number of markers increased to 10
Marker setting "on the fly" (space bar)
Playback of initial section of selected area (Left Shift)
Playback of final section of selected area (Right Shift)
DAT back-up using digital I/O
Save preferred configuration

All existing ADAS owners updated free of charge

Plasmec have sorted out most of the hardware problems; ADAS records to and plays back from hard drive and the audio quality is very good, if a little noisy. The current problems are with the software, and program development takes time - a lot of time.

Some of the main problems will have been sorted by the time you read this with the release of v1.4, but there are others which are of a high priority, especially the updating of cue-sheet edits which makes the cue sheet very awkward to use, and the addition of key equivalents wherever possible. I've no doubt that these will be attended to and other worthwhile features added.

Personally, I'm not convinced that a direct-to-disk system running on an Atari ST can be regarded as a fully-professional system. That said, the instances of professionals using less-than-professional gear and liking it are too many to count. Beyond (or beneath) the professionals there are plenty of other ST users who really owe it to their music to check out ADAS. The bottom line is that ADAS may well represent access to technology which would otherwise remain out of your financial reach. You might say that technology has just broken the sound barrier once again.

Price £849 with psu

More From Plasmec Systems Ltd, (Contact Details).

Also featuring gear in this article

Previous Article in this issue

Korg 01/W FD Workstation

Next article in this issue

Multitimbrality Made Simple

Music Technology - Copyright: Music Maker Publications (UK), Future Publishing.


Music Technology - Nov 1991

Review by Vic Lennard

Previous article in this issue:

> Korg 01/W FD Workstation

Next article in this issue:

> Multitimbrality Made Simple

Help Support The Things You Love

mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.

If you value this resource, you can support this project - it really helps!

Donations for October 2021
Issues donated this month: 8

New issues that have been donated or scanned for us this month.

Funds donated this month: £50.00

All donations and support are gratefully appreciated - thank you.

Please Contribute to mu:zines by supplying magazines, scanning or donating funds. Thanks!

Monetary donations go towards site running costs, and the occasional coffee for me if there's anything left over!

Small Print

Terms of usePrivacy