The Quantise and Groove strategies on Notator have always been one of its strengths, and these are now enhanced still further in version 3.0. Whilst Quantise adapts timing to a preset error-correcting template, and Groove applies the same facility to 'swing' feel or user-defined templates, Adaptive Groove automatically selects the optimum error-correction strategy from a range of simultaneously active templates. This allows for far more sophisticated feel in a part than any traditional quantise would be able to handle. (Adaptive Groove '16C+24' for example, is a combined straight 16ths and 16th triplet strategy, with a 'swung' 16ths syncopation applied to the 16ths only, according to Preset Groove 16C, one of the pre-determined 'swing' quantises). Up to eight templates, selected from the range of Quantise values, Preset Grooves and User Grooves, can be combined into an Adaptive Groove Set; eight sets are available per song. There are seven Adaptive Groove Sets already programmed, which in practice seem to cover most eventualities, thereby probably rescuing most users from the need to ever attempt to understand how to set up their own!

A second level of correction can be applied via the descriptively named 'Bend to...' parameter, with the option to bend from the original positions, or from the new positions after the first template has been applied. Adaptive Grooves have further Segment and Minimum parameters, which determine the period over which a template may be active, and the minimum number of notes that must be available before a note will be shifted to a template position. Finally, Groove templates can be given an Advantage value between 1 and 10 — templates with a higher Advantage figure exercise more influence on timing than those with a lower figure.

This really does all work, tightening up timing whilst allowing the maximum amount of 'feel' to still come through — surely the most sophisticated and effective quantise strategy yet devised. I recommend Version 3.0 for this alone.

NOTATION

The notation side of the program has received a number of small enhancements, with the addition of time-saving features such as Diatonic Insert, and the correction of some omissions, such as the ability to handle grace notes. Diatonic Insert, as the name suggests, automatically places inserted notes (from the partbox) only at the correct pitches for the scale of the current key signature.

Grace notes are at least now possible, although their handling is still a little more convoluted than most of us would prefer — Notator's polyphonic voice separation, which facilitates display of independent parts within a single stave, is used to achieve differentiation between the main part and any notes designated as grace notes. The miniature notation facility is then used on those notes only (via Shift-Alternate-M). Horizontal Shift (in the Note Attributes window) can then be used to further enhance the grace note effect if the default micro-shift doesn't look quite right. It solves the problem, but not very elegantly — surely an option to have the system automatically designate any notes below a user-defined length as grace notes would be simpler? Manually deleting the few spurious notes that would inevitably be displayed as a result of this would probably be quicker and easier than using the polyphonic separation facility.

Another new feature, MIDI Meaning, will go down well with those who prefer to edit via the notation display. This allows you to define which notation symbols will affect the actual MIDI data. Accent symbols can add velocity, or a staccato instruction can actually shorten a note.

Complex tuplets, quintuplets, septuplets, etc. (up to 15!), can now be entered via the new Note Attributes window. You call this up by double clicking on the note entered as the first of the tuplet set. You then have only to enter the appropriate number to define the tuplet ('7' for a septuplet, and so on) and the time that the notes will occupy (half note etc.). It is not a feature that you will use every day — indeed many people may never use it at all — but if you are preparing a score, it is desperately frustrating to come upon something that the system simply cannot deal with, and it is good to see Notator now able to handle this efficiently.

Mixed Triplets, such as quarter notes and eighth notes in the same triplet, are now theoretically possible via the Note Attributes parameters, using the Syncopation facility. I must admit, I haven't so far been able to get this to work, so either it doesn't or I am doing it wrong!

PRINTING

At last we have a Page Preview mode, so that the layout of pages can be viewed before committing to print. Unlike many word processors, the preview page itself is 'active'; you can alter the margins just by dragging them, and you can change the number of bars in a line, either locally, or globally. You can step through the pages of a large piece, and print just one of them, if necessary. Page Preview is destined to save a lot of time, frustration, and wasted paper!

Titles, lyrics and instructions can now be further enhanced by a wider choice of fonts. Two additional GDOS fonts are supplied with the system — standard Helvetica and Times Roman (more familiar to regular GDOS users as the Swiss and Dutch proportional fonts). Any Atari GDOS font with the '*.FNT' extension should be compatible — fortunately the nuts and bolts of GDOS operation, all that stuff with the drivers and the tedious ASSIGN.SYS file, are all taken care of by the program. Shouldn't all GDOS usage have been like this? GDOS uses bit-mapped fonts and therefore large sizes can appear rather 'stepped' on-screen, although they still print well (perhaps the vector fonts of the rumoured new GDOS will similarly be able to be used). Downloadable fonts will probably not seem a particularly valuable feature to many users, but to anyone using the program primarily for score printing origination, they will be essential in allowing the appearance of a Notator page to compete with dedicated score printing software.

RMG AND MACROS

Notator's Realtime MIDI Generator (RMG) system of on-screen faders now has a fader grouping facility, allowing any fader to become a master to any of the others. One small point to watch out for is that relative offsets are lost if any of the group reaches data value 127 or 0 before the master does — if the master is the lowest value in the group, once they have all reached 127, they will then all be stuck at the same value! This is very easy to do accidentally, and I think it would be preferable if the system were able to recover the relative positions after they have been levelled in this way.

Sys Ex handling within RMG has been further enhanced by the incorporation of graphic switches for use where this is a more appropriate control source. Furthermore, it is now possible to delete and re-record the fader data of a single channel without having to exit the RMG window.

The number of functions duplicated by, or activated solely by, keyboard entry has increased dramatically in the last couple of versions, as the supplied new pull-out Keychart shows. A Macro function has now been added, enabling combinations of keystrokes and, supposedly, mouse movements to be stored and recalled via a single shifted key. This potentially useful facility is sadly undermined by the unreliability of the recording of mouse movements, and the fact that it not always possible to specify an absolute value into certain parameters. For example, I would find it very useful to be able to set up a macro to execute the functions that I always use to process a MIDI guitar track: Delete Short Notes, enter a negative delay in the track parameters box, then Normalize the data via the 'N' key (to fix the data in its new position prior to quantising), and finally select an appropriate Quantise or Groove.

As you can see, to be able to do that lot from a single key would be rather an asset! Unfortunately it would appear to be beyond the present macro implementation. Even if the mouse pointer makes it to the correct menu, it will not always select the right item when it gets there, and if the system does not take the same amount of time to scroll to a value as when the macro was recorded, it will abandon the process at whatever figure it has reached. The Short Notes value will scroll to the correct figure the first time it is used but, due to the program's helpful ability to retain the last used value for many parameters, it will subsequently add itself to the value each time it is used. This is a shame, for a powerful macro implementation is a great asset to any system — as it is, the use of macros really needs to be confined to simple things, like automatic track name entering, when they work reliably.

PATTERN OVERVIEW

There are actually too many small but significant enhancements to the working of the program to cover them all in the space available. However, among the features which I found particularly useful are Pattern Overview, which allows an instant view of which patterns contain data and whether they are used in the present arrangement or not, and the inclusion of a display of small 'invisible pattern' numbers at the top of the arrange listing. This allows you to see which patterns are active in different levels, even though their entries may, in a complex arrangement, have disappeared off the top of the display. Track Play Parameter values can now also be conveniently incremented in logical steps, with a single click giving transpose in octaves, or loop in whole bars and so on. The operation of the very useful Force Legato function has also been tidied up, making its interaction with Overlap Correction more logical.

Yet all is not perfect with Version 3.0 by any means. There are too many small irritations that have consistently annoyed at least this user from the outset that have survived the opportunity for eradication presented by this second re-write. There is still no 'Individual Un-Delete' in the edit page — if you overshoot by one note on a scrolling delete, you have to Undo the whole edit, and start again. I would much rather have individual un-delete than individual de-quantise, which, paradoxically, is present. (To achieve a mixed quantised/un-quantised effect, I find it makes far more sense to leave the whole track un-quantised and then use spot quantise where appropriate.)

The only Exit box from the Edit window is still placed low down on the right-hand side, where it will be obscured by all but the smallest of notation displays. I know you can use the Return or Enter keys on the keyboard, or the GEM 'close' box in the corner of the window, but neither is quite as convenient as a better placed Exit box would be.

Recording Arrange Mutes still produces unpredictable effects — occasionally, adding another mute will turn off any that are already present. No particular damage is done, for all mute data seems to remain correctly recorded, and will play back perfectly next time round, but it is disconcerting nonetheless, and the kind of thing that one would prefer to see disappear with a re-write. Similarly, the tendency for a cycled pattern to momentarily display the following pattern as it wraps around is still present. This can be the source of some aggravation when editing values in the track parameters box during cycled playback, as you can find that you have inadvertently edited a value in the wrong pattern!

My final gripe concerns what appears to be a small degree of Operating System incompatibility. Older TOS 1.2 Ataris seem to display peculiarities with Notator V3.0 that they have not shown on any previous issue of the program. The Track Exchange/Merge box is incomplete, lacking its normal Cancel facility, and Merging is accompanied by a small screen corruption in the track box — the facility still works correctly however, and the corruption disappears next time that area of the screen is re-drawn. (Early Atari TOS 1.2 machines are the ones which will execute the Save Desktop operation without a dialogue box; later TOS 1.2, and 1.4 require a confirmation before saving.) If there is any question of TOS incompatibility, then it should be rectified as soon as possible, for there are rather a lot of TOS 1.2 machines out there, and many of them will be owned by people who bought the C-Lab programs when they first came on to the market. Without them there would be no version 3.0. I am also slightly disturbed by the upgrade pricing policy — a V3.0 Notator upgrade kit costs the not insignificant sum of £49, yet it does not include the larger 'Notator SL' style binder necessary to house the new 400 page manual. These are an extra £10. Surely existing users deserve better than this?

FINALLY...

I have been a C-Lab user from the outset, and have seen the program evolve and expand to its current, arguably pre-eminent position in the ST sequencer world. Version 3.0 continues that growth, with both solid enhancements and entirely new features, fitted neatly into the existing framework. It would have been nice to have had one or two more of the longstanding little irregularities cleared up at the same time however, and it might be argued that Creator users get rather less out of this upgrade than Notator owners, the gap being perhaps insufficiently reflected in the upgrade price differential.

Version 3.0 of the famed C-Lab sequencers will undoubtedly continue their remarkable, and deserved, success story. The power and flexibility on offer, I believe, stands comparison with any other Atari-based system on the market. As all too many present systems demonstrate, power alone is never enough — it is the ability to efficiently direct and control that power that really counts, and it is precisely here that the unique quality of the C-Lab user interface places Notator on a level all of its own.

FURTHER INFORMATION

Notator V3.0 £499 Inc VAT.
Creator V3.0 £299 inc VAT.

Sound Technology plc, (Contact Details).