STEINBERG

SATELLITE

The Satellite accessory supplied with Cubase/Cubeat is still causing problems to users. The biggest difficulty comes from a misunderstanding of what Satellite is designed to do. Reading from the manual, in the introductory section... "while Satellite can work with any sequencer, its full potential is revealed with Steinberg's Cubase/Cubeat software. You can only edit sounds that use the Steinberg Synthworks format."

I must have said this a thousand times now, and I can assure the non-believers that the above text really is in the manual. What Satellite can do is look at the Bank files created by Synthworks programs, and then allow you to access them. The access is fairly comprehensive — you can send the whole Bank of sounds, send one sound, 'glue' the Sys Ex dump of a sound into Cubase/Cubeat's tracks, or use the Macro Editor sliders to edit the sound. These Synthworks files are the only files that the Sliders will work on. You cannot perform a bulk dump into a totally different section of the Satellite (the Bulk Dump Utility) and then use the Macro Editor sliders to change the sound. Changing the Instrument Names makes no difference, all you are doing is telling the Satellite that the next type of Synthworks file to be loaded will be type 'X'. This has no bearing on the Bulk Dump sections at all.

So why can't you use the Macro Editor on Bulk Dump files? Well, the reason is not really that complicated. It has to do with the fact that you have to tell the Satellite which type of Synthworks file you want to load. This is done on the Instrument Setting page. When a file is loaded from Synthworks, the program can analyse the file, and knows that the data for Voice XXX lies between byte YYY and byte ZZZ on the disk file. It can now wrap that up in the appropriate Sys Ex headers and send it out, or send out modifications to it (with the Macro Editor) because it knows which bits need to be changed. However, within Bulk Dumps, we cannot know where individual voices lie (and in any case, many Bulk Dumps are not just synth voices), so what do you extract to send out? Where does any voice start or finish? What would you change to alter the voice's release time?

CONTROLLER MAPPING

Cubase and Cubeat have a Controller Mapping feature on the MIDI Definitions page. This allows any controller to be mapped to any other during both record and thru. It is important to realise that aftertouch is not a MIDI controller. It is in a class of its own, in the same way that notes are not pitch bends. When I refer to aftertouch, I mean Channel Pressure, and not Polyphonic Pressure.

In a simple case, controller mapping can be used to change Breath Control information into MIDI volume data. This can be performed easily in real time, because we are are changing like for like. Consider a situation when it would be useful to change modulation data into aftertouch, or vice versa. Why? Well what would happen if you had a sequenced piece where you had made extensive use of aftertouch to introduce vibrato into your lilting refrain, and now you decide that a sound from another keyboard or expander would be better. But where has the vibrato gone? The new unit doesn't respond to aftertouch, only modulation data. What do you do? Go to Logical Edit and convert it.

The problem with this situation is that controller events are three bytes long with the 'variable' in the third position, and channel aftertouch is a two byte message with the variable in the second position. Using the Logical Edit's Condition parameters to accept only aftertouch data, you can now set the Result field to Controller, set the Value 1 column to 1, so that we will be generating modulation data. Now drag Variable A from the aftertouch data to the Value 2 column of the Results field. This will move the active byte from the second position in the aftertouch data to the third in the new modulation data when you click on Insert or Transform. You may find it useful to look at page 22-12 of the Cubase manual. A word of warning: we count our Value 1 and Value 2 settings after the status byte, so a three byte message is [status: Value 1: Value 2], and a two byte message is [status: Value 1].

This idea can be extended as far as you want or need. There are only two limiting conditions — you need a general understanding of the Logical Edit page, and at least a basic idea of the structure of MIDI data. For example, you could take a modulation sweep and use its Value 2 to create notes of varying pitch (Value 1) for crazy impossible glissandos. In this case it will be necessary to use Logical Edit again to give the notes a real length.

CHIPS WITH EVERYTHING

There are new versions of the SMP24 and the Timelock software. The latest version of the SMP24 is V1.65, and Timelock is now up to V1.41. The changes are minor, but may be significant to your operation. Firstly, both devices now write the type of SMPTE/EBU that an SSL desk can understand, and the drop frame flag is set when writing code at 30fps (drop). Previously, both devices could read drop frame SMPTE, but didn't set the flag when writing Timecode.

On the SMP24, the 'running status' selection for each output is now set entirely on Cubase (or more correctly, by MROS) and cannot be changed again on the SMP24. This increases the speed that the SMP24 processes MIDI data under MROS control. Most importantly, the operating system that previously only resided in EPROM in the SMP24 can now be updated by MROS itself. The driver for the SMP24 in the MROS folder will actually send the higher level functions 'down the wire'. This means that from now on updates will be as simple as loading the latest disk version of any MROS software.

Timelock V1.41 improves the dropout detection, but it is very important to use the latest MROS driver if you are using Cubase. This comes automatically on V2.0 of Cubase, but if you want to check what you've got, the TIMELOCK.DRV file in the MROS folder should be 1495 bytes long.

C-LAB

PAPER LENGTH

Amazing-but-true facts, part 47: although the DIN standard for paper length is a German invention, their DIN A4 length in continuous stationery (also known as 'fanfold') for computers is in fact exactly 12", whereas their (and our) cut A4 is the correct 11 2/3" long, or more precisely 297mm long by 210mm wide.

The default printer driver ('adaptation') that comes ready loaded with Notator is for 24-pin printers and for 12" paper, even though it is called 'A4': this is because we didn't realise that 'true A4' continuous stationery existed.

Provided you have set your printer to its A4 setting, this adaptation is fine. If your printer has no such setting (hello NEC P6 and Star LC24-10 owners!), you will need to use the new adaptation provided with V3.0 which tells the printer to expect 'true A4' paper length. Without it, the sheet perforations will become more offset with each successive sheet, because the printer thinks it has 12".

When you buy paper, insist on 'true A4', and measure it if necessary (see above). There are more varieties of paper sizes than the mind can comfortably take in, a situation that is sometimes complicated by erroneous descriptions on the boxes. A good weight to go for is '80 grams'.

SYNCHRONISATION

One of Human Touch's great strengths, when combined with Unitor, is its ability to create a new 'Sync Reference' when fed with an audio signal from a tape — so you are able to create a new sequence to go with existing music, even if there is no existing sync track. What you end up with is Creator or Notator locked to the tape, having used, say, the kickdrum as the continuous tempo source.

The success of the operation is all down to the quality and distinctiveness of the audio signal's beats, and your skill and patience in programming a Tempo Interpreter setting that allows as little variation in tempo as posible, something which is is easy to achieve with practice. You should aim to minimise the Window and Maximum Tempo Change values. With standard commercial material, you should probably end up with a gently varying tempo on each beat of less than 1/2bpm either side of the true tempo.

This slight 'warble' may be too much for certain applications that require a rock-steady tempo from start to finish. Here, it may be necessary to put another cunning plan into operation.

Set Creator/Notator's tempo, in Internal Sync mode, to be as close as possible to that of the music on tape (it helps if you switch on Play Click, and direct the MIDI click to a kickdrum or similar). Feed the SMPTE timecode into Unitor and open the Sync Reference window. Start the tape, and click the Pickup Frame switch as a soon as you hear the count-in or music. This creates an initial 'Sync Reference'. Now spend some time repeatedly starting the song on tape, and adjusting the SMPTE Offset value until the first MIDI click in the sequence (and the two or three beats after it) are exactly in time with the tape. To achieve this, Unitor allows you to alter the SMPTE Offset value on the fly while the tape is running, and directly hear the results.

Once you have done that, wind the tape all the way to the end of the song and repeatedly shuttle around the last handful of beats, this time adjusting the tempo on the fly. Now you are trying to get the very last MIDI click beat in sync with what's on tape. Make full use of the Sync Reference tempo's four decimal places to get a precise result. Caution: the (MIDI) tempo you start off with (see beginning of previous paragraph) must be as close as possible to the tempo of the music on tape to minimize the risk of the sequence being a whole beat or more out by the end of the song. As a check, you can watch the Main Bar Counter when you are making your final adjustments, to ensure it is showing the correct bar and beat.

That's it! The combination of a precise SMPTE Offset at the start with fine tuning of the tempo at the end allows you to achieve a rock-steady linear tempo.