All rhythmic sequencing hardware devices and software applications have internal clocks that keep time and drive their sequencing engines.

The accuracy of these internal clocks range from surgically precise to terminally unstable.

People often say they don't like the 'sound' of a particular drum machine or sequencer/workstation/sound module combination. Less obvious, however, is their subjective judgement of the machine in question has a lot less to do with the sounds it makes and much more to do with how accurately it re-plays these noises against a theoretically perfect tempo grid.

To make things more complicated, within every sequencer or drum machine there is an offset between its own internal pattern tempo-grid and the external outgoing or incoming Midi Clock/Din Sync signal it either generates or slaves to.

In a perfect world they should be sample accurate.

This is almost never the case.

A rare few get close, most are average and the vast bulk are so far out you could be forgiven for thinking you were listening to another machine altogether.

This is one reason why a group of individually good-sounding sequencers and/or drum machines synchronised together by Midi Clock and Din Sync directly, usually sound average at best and diabolical at worst.

If you are interested in knowing just how bad things really are, an easy test that requires no complicated equipment is to run two of your favorite machines in sync playing identical 16th rim shot patterns or even metronome clicks and record both units - one panned hard left and the other panned hard right - and record them as a single stereo .wav file for a few bars.

Open up the recorded file in any stereo software audio editor application (Soundforge or similar) and zoom in to look at the fronts of the 16ths you have captured - you will be amazed at how far apart two supposedly 'synced' drum machines actually are!

Now measure the number of actual samples between the very front of every 16th on each machine respectively. They should be identical. The more variation, the greater the rhythmic 'slop' and the worse the machine in question 'sounds'.

 The Roland TR-808 running under its own steam is a tight box on the whole.

Sample the Rim Shot into a good hardware sampler and edit the front so it is nice and tight. Now program hard quantized 16th notes in any software sequencer application on the best XP or OS-X system you can find with any USB-Midi interface and get it to trigger your Rim Shot sample. The result will be unusable. Period. Measure it with our little test. You will be astounded at just how bad professional software sequencer applications are at keeping time in the outside world.

All this has nothing to do with sound quality.

Use an Akai Z Series 24/96kHz sampler if you have one for the test. The results will be the same.

Comparing our original TR-808 with sequenced samples in the above test you will always be drawn to the former.

What matters most to our ears and musical brain is not the sounds themselves but the clocking engine that plays them back in time.

Now try the same test but this time dump the samples into an AKAI MPC-60 Mark II and sequence up the same 16th pattern inside the MPC sequencer itself. The result - outstanding. Funk in effect! Why? Because it plays back the samples in time with great precision under any conditions. Its sequencer timing clock doesn't sag, slop, rush, drift, push, pull or pause when it is busy doing other things.

Of course you can sample-align audio these days in any PC based Digital Audio Workstation after you have recorded it no matter how sloppy the sync is to begin with. You can even get a close free-running tempo match on any drum machine or sequencer, record it in isolation and fit a re-aligned loop to the track you're working on easily enough without too many time-stretch artifacts creeping in but that is not really the point of this discussion.

With musicians, real magic happens when band members play with, to or against each other.

To best illustrate how this relates to synchronisation, listen to a personal favourite recording you own by good musicians either captured live in one take or recorded track-by-track played by the individual band members against what had previously been recorded.

Now imagine that same piece of music was recorded while those same band members were drunk at the recording session.

You would recognise the song of course but would it have the same feel or emotional impact as the original?

This is our analogy for sloppy sync.

Next, imagine the notes of the song were written down and each band member then recorded just their part in isolation - that is - without listening to any of the other player's parts. Now take those separate parts and construct the song in a computer.

Again the song would be recognisable but would it be worth listening to?

This is our analogy for composition by mouse and grid.