That Syncing Feeling (Part 4)
Chris Smith continues his series on synchronisation
Chris Smith continues his synchronisation series, with Synchronisation in music - A definitive guide
The past two or three years have seen a proliferation (no, a complete deluge!) of MIDI sequencing software. Take a look at Table 1 and you will see what I mean. The days when Pro 24 was practically the only professional sequencer choice for the studio are long gone. Where does all this software come from? Surely all MIDI musicians have a sequencer by now, or have they?
Hardly a month goes by without a new program, or at least a new version of an old program being released onto a market that must, by now, be approaching saturation point. New customers can only be first time buyers or converts from other brands, and with so many to choose from the competition must be pretty fierce.
Despite this, music software remains ridiculously expensive. It seems the least important feature is the price, or maybe people simply suspect the quality of cheaper programs. This argument does not hold true for keyboards. More cheap portable and home keyboards are sold than any other type of instrument. Why? because they are cheap. Loads of musicians would like an M- 1 instead, but they simply can not afford one.
I would not deny that some music software is extremely good, and probably took years of research, but research is a fixed cost and only has to be paid for once. With the marginal costs so low (disk copies and mass produced computer peripherals) and so many new products chasing a steady market, surely price competition has its part to play. Instead producers like to spend their money (ie your money) on expensive marketing and advertising campaigns, when maybe a price reduction would yield the same result.
Okay, now this is a computer music magazine, so let's look at the stuff that brings them together - the software. What should you be looking for in a MIDI sequencing program?
There are at least five ways to synchronise your sequenced music. We have internal sync, external sync to MIDI clocks, direct to SMPTE, MTC, and tap input. Different sequencers give you different options with different combinations of hardware peripherals.
You will inevitably be using this method most, so make sure your sequencer choice is easy to start, stop, fast forward/backward, and autocue in both play and record modes. You tend to use these functions a lot when building up sequences and awkward controls soon get irritating. Most sequencers give you a choice of controls for these functions - the mouse, the computer or the MIDI keyboard.
Programmable metronome and count in facilities are useful, and you should be able to program as many tempo changes as you like throughout your song. Better programs allow accelerando and decelerando over specified bars (gradual increases or decreases in tempo) and even 'fit time' by which the sequencer calculates the correct tempo to make a specified number of bars and beats last a specific length of time (in minutes and seconds). This usually requires a tempo resolution of 0.01 bpm and is useful for background music.
All modern sequencers output MIDI clocks and Song Position Pointers but it is useful to be able to disable output clocks in some MIDI systems.
This is by far the most common method by which MIDI equipment is synced together. Most MIDI hardware (sequencers/drum machines) and UK software on the Atari, Amiga and IBM compatibles use it exclusively. In a 'sequencer only' based system this method works well and is quite sufficient. To lock to real time (ie tape) however, you need more than MIDI clocks.
Tape sync methods were discussed at great length in the first article of this series. Most sequencers now have some kind of hardware peripheral equipment using FSK or SMPTE codes to lock to tape, or you can use a separate, discrete synchroniser that will read SMPTE and send MIDI clocks to any sequencer.
Computer add-ons have the advantage over discrete synchronisers in that tempo maps are set up and saved as part of the sequence file (although MTC equipped synchroniser/sequencer systems share this facility). On the other hand computer peripherals are nearly all 'specific' to one particular computer and sequencer package, so if you go for that option make sure it has all you will ever need.
A good tape sync system will be based on SMPTE (the standard time code, with all four rates catered for), will allow you to set up start time to bit accuracy and will allow a complex map of tempos to be set up and edited throughout the song. Also look for the ability to input MIDI events (usually notes-ons) at specific SMPTE times (hrs, mins, secs, frames) as well as musical times (bars and ticks).
All of the above methods of synchronisation will exact a rigid tempo, derived from precise electronic clocks. Much of live music output is free of this restraint and has a subtle human element of randomness in its tempo. For computer systems to imitate this requires either the sequencer program to introduce its own calculated tempo randomness, or for an external human clock source to dictate tempo variations which can be recorded. Many sequencers incorporate both options, but the latter method also allows the user to synchronise his MIDI sequences to other music recorded previously on tape, or played live by a band.
The sequencer is set to tap input and synchronised to SMPTE code laid down alongside existing music on multitrack tape. The user then taps out the beats (usually on a MIDI keyboard or drum pad) in time with the recorded music and the sequenced music follows the beat, making and recording all the subtle tempo changes needed to stay in time. The MIDI sequence can then be manipulated as required and is saved with its corresponding tempo map, so that it can be recalled and synchronised at any time.
In practise tap time input is a little awkward but the results can be very worthwhile. A good sequencer program will help out by letting you adjust things like tap capture windows, maximum tempo variation, initial tempo etc. One more hardware enhancement lets you take the idea a stage further by listening to audio cues from the tape (eg snare or bass drum) and deriving the tempo directly.
The merits of MTC were clearly put forward in Part 3 of this series. The standard is well adopted in the US and is steadily being introduced in Europe. Most software sequencers have, or are in the process of incorporating MTC (often called direct time lock), but hardware equipment still lags behind. MIDI systems, it seems, will always require a combination of MTC and MIDI clocks, but once past the tape sync interface there is nothing wrong with that.
Suffice to say, now that the main advantage of MTC equipped sequencers is that while they offer all of the above synchronisation facilities, they also offer full compatibility between hardware and software from different manufacturers. For example, you should be able to use tape synchroniser A to direct time lock sequencer X and tap input tempo against music on tape. Now save the sequence with its tempo map using the standard MIDI file format, load it up into sequencer Y and direct time lock it to the same tape with synchroniser B.
Table 1 is a list of sequencers available in the UK, marking their facilities regarding synchronisation. I would not suggest for a minute that you should or should not go for a system on the grounds of sync alone. Sequencers have a wealth of desirable and undesirable features to consider, but this is at least a good place to start.
Even if you do not have much MIDI equipment right now, if you are still playing music in two or three years time, you will have. A typical MIDI set up may have several sources of synchronisation; a computer or hardware sequencer, a tape sync unit, a drum machine or two, a synth with a built in sequencer or rhythm unit, and so on. Only one of these sources can be the master clock. Everything else has to be either slaved to the master, or used purely as an expander module set to respond to notes only and not to MIDI clocks.
My word of advice is be careful when setting up your MIDI cables. The main points to be aware of are:
1) Most sequencers (especially drum machines) always echo thru MIDI clocks, although some units allow you to turn off out going clocks. It is all to easy to unwittingly wire up a system where clocks are looped back into the sending device and get echoed round and round the system causing chaos. This is especially true using tape sync methods (see diagram A).
2) It is very difficult to design a system using a normal MIDI chain system (ie out to in to thru to in to thru etc) where the master clock source can be changed without having to repatch the MIDI cables (see diagram B).
The problem is always that MIDI thru connections go to MIDI ins, so there is nowhere left to plug in the MIDI outs (if you get my drift), If you want to utilise your clock sources efficiently you really need some kind of MIDI routing or merging hardware - which leads on to point 3.
3) MIDI clocks just cannot be merged. When two clocks meet, even if they originate from the same source, they will then be two clocks. This can cause double tempo rates or even worse MIDI chaos.
There are hardware MIDI processing boxes that will deal with clocking problems properly (ie by only letting one through at a time) but check that these will provide the routing and merging facilities you want, and will want in the future, before you buy.
Many people I talk to adopt the 'it works as it is, so I'm not touching it' attitude. All they need is to bear the above points in mind and invest in one or two extra bits of MIDI hardware and they could really unleash the power of their system. Do not be afraid to experiment. Moving MIDI cables around can never harm your equipment, although it can get very confusing.
A good MIDI system set up should be flexible enough to allow all the components in it to talk to each other via MIDI notes, MIDI clocks and system exclusive data.
Much of TV and promotional video production consists of animation or computer generated graphics. These are generally used to add colour and visual excitement to otherwise conventional presentations and announcements at the start and end of programs, hence the name video titling. More recently such graphics have been mixed with video effects and live action to form a central part of the actual content of TV programs.
Channel 4 engineers have been pioneers in this field and are using relatively affordable techniques to great effect in shows like Star Test and The Chart Show. Both of these utilise the Amiga computer to generate and manipulate pictures and text. In contrast the BBC tend to use graphics mainly for titling or presentation of facts and figures in a more exciting way, such as Going for Gold or news presentation. They use more expensive systems such as Quantel Paintbox on Vax computers, or Masterpiece systems.
So why is there this discrepancy between the two company approaches? Well, at the Beeb nearly all home-grown programs are engineered in-house and high equipment costs are easily justified (and expected). Production costs rarely reflect viewing figures. At Channel 4 programs are made to a budget and ratings are generally much lower, so production is often contracted out to specialist video companies who have the equipment and expertise to complete the work at an economical and fixed price.
Take a look at Star Test (Channel 4 audience around 700,000), an interview show where personalities are interviewed by computer. Graphic image artists Simon Dunstan and Chris Mills use the Amiga 200 to great effect, together with Delux Paint software, Newtek's Digiview and a PD page flipper locked to off-line video using Bausch gen-locks.
Although a video suite is used for the final production, in theory they could run Amiga graphics over the entire output of the show all at once.
We now have specialist video production companies setting up all over the place, taking on work not just for TV, but for advertising, corporate videos, education and training films. Such companies supply the right combination of equipment, engineering and artistic talent, and realistic pricing that the market now demands.
There are basically 2 ways to get animated (ie drawn) or computer generated graphics onto video tape with live action. Firstly, you can draw your pictures on paper, photograph and record them onto video tape, then mix the live action video playback with the animation video through a vision mixer while recording onto a new tape.
This method (see diagram C) is a little cumbersome but remains the only way to get 'drawn' images (ie cartoons) onto video.
A more modern approach is to get the artist to use a computer graphics package to generate pictures, or to use a video camera and computer to 'digitise' a picture. Both methods allow the computer to 'store' pictures in software for later retrieval and to manipulate pictures with respect to their size, screen position, colour content, shape and so on.
To get software images onto video tape requires a gen-lock device for synchronisation. In simple terms this device looks at the incoming video from a VCR playing the live action, derives the frame sync pulse and feeds the corresponding sync signal into the computer to synchronise its monitor output. This ensures that the graphics video output is in sync with the live action video input and the gen-lock can mix them together for recording onto another VCR (see diagram D).
Now each picture can be loaded onto the computers monitor, manipulated as required, mixed with the appropriate peace of live action and recorded bit by bit. Synchronisation between the playing and recording VCRs is only necessary here if a series of tight edits are required, dropping in pictures and action at specific times.
If graphics are being added to an otherwise complete production the computer may be powerful enough to recall or generate a number of sequential images to record animated action over live action all in one take.
Simple graphics programs are available that allow you to input text and logos in various fonts, sizes and colours and screen positions. These can scroll up or across the computer screen to be recorded as credits over video. More complex graphics packages allow you to 'capture' video frames for manipulation, or build up cartoon-like moving images frame by frame (like video games graphics).
Anything appearing on the computer screen can be recorded on video using a gen-lock, even mouse movements and GEM icons. The easily recognisable Amiga GEM system is put to good use for The Chart Show's information presentation for example.
Well it seems that each computer finds its own market. While IBMs are in the office, Acorns are in schools, Ataris are in the studio and Spectrums are in the home (in the attic), the Amiga has found itself a nice little niche in the video field. A combination of very good graphics, true multitasking, easy to use GEM, and plenty of gen-lock systems and software around mean that the Amiga has to be first choice for the amateur or semi-pro enthusiast.
Unfortunately for us musicians looking into video production we may well have to either use two computer systems, or move our sequencing across to the Amiga. The Atari cannot synchronise its video output externally and as yet no genlock systems are available. A case of nice MIDI sockets, shame about the video.
Gen-lock systems start at around £100 (eg Micro-gen) complete with basic titling software. Even at this price the facility is impressive. You do not need special software to run your favourite games or graphic pictures with the TV in the background, or record them onto your VCR. You must watch out for compatibility problems though, between gen-lock, VCR, software and monitor. Always try to get a good demonstration and explanation before you buy.
More expensive gen-lock systems give better picture quality, more stable and universal sync. (PAL/NTSC), fade controls, and more flexible titling/drawing software. As prices and facilities increase towards frame grabbing and 3D image processing, more computer hardware is needed such as memory expansion, hard disk drive, graphics/video cards, PC bridge cards and so on. Surely though, from the casual user to the professional, the Amiga system offers the widest choice.
As I mentioned before, most professional software (at professional prices) is written for IBM compatible micros with special hardware add-ons for VCR synchronisation as well as screen handling and graphics.
Amigas do have some competition in the semi-pro market coming up in the near future though. Apple Macs are now available in the UK for around £1000, opening the door to American video and audio software here.
Atari and Yamaha have their PCs (IBM compatible) just around the corner, and Acorn have a number of systems available or under development for the extremely powerful Archimedes computer.
Okay there are some good software/hardware packages around for music and video productions, and the new Japanese VHS recorders do give excellent picture quality for the price. If you are looking for a complete system for audio/visual production though, you have three basic problems.
1) Video and music (MIDI based) software and hardware is directed towards completely different markets and on the whole is written for different computers. There is no reasonably priced system that deals with video and MIDI sequencing/sampling requirements simultaneously or synchronously on the same host.
2) VHS recorders are essentially designed for recording TV programs and little else. Machines with even rudimentary editing facilities are prohibitively expensive and over complicated.
3) While the SMPTE standard is put to good use in both the video and audio production fields individually, as yet few facilities apply the available technology to unite the two. This means that effective audio/visual synchronisation and control equipment will remain an expensive luxury for some time to come.
Perhaps I am being over enthusiastic. Perhaps the public get what the public wants. But if and when a manufacture or group of manufacturers can offer a comprehensive, expandable and integrated equipment list for complete production in the video medium then the market will reach its full and proper potential.
Well, I did not have room for the glossary this month - sorry about that but it will be here next time, promise. I will also be looking at the really expensive stuff - direct to disk sampling packages to see how they sync up. There will also be a list of company numbers and contacts that you may find useful, mainly those I found helpful in compiling what you have been reading.
Feature by Chris Smith