Recording - December 2015 - 59

Sync boxes
A sync box is a device that translates between different
kinds of sync signals, allowing different devices to lock in
with one another. Arturia, Doepfer, Kenton, Korg, Roland,
and others make them, each with different input and output
possibilities. On our website (http://is.gd/RECGetInSync),
you'll find detailed tables with all kinds of useful data on a
bunch of different devices, including sync boxes.
Some drum machines and sequencers can act as sync boxes,
even though they aren't advertised as such. For example,
Arturia's BeatStep can perform MIDI-to-CV/Gate conversion (via
USB) and the BeatStep Pro version can act as a sync box, standalone or with a DAW-look for a review soon. Vintage devices
like the Roland MSQ-100 sequencer and TR-707 drum machine,
or the Sequential TOM or Kawai R-100 rhythm boxes, can perform many sync, triggering, or even clock-dividing duties. See
the tables provided online for more details.
Hey, something's off: latency and jitter
When two or more devices sync properly, everything just
sounds right and "tight." We don't think much more about it.
Only when sync goes wrong or something doesn't feel right
do we scratch our heads and become unwilling detectives.
There are many culprits causing poor sync, but the two
usual suspects are latency and jitter. Each device or system
can contribute its own bit of latency and jitter into the studio
chain. So the detective work includes determining where the
timing discrepancies originate. Further, we must discriminate
between one-way (outgoing) and round-trip latency and jitter.
Latency is the time delay from the musical event's expected
time to its actual occurrence. For example, a drum kick should
typically fall exactly on the measure's downbeat or "on the
grid." When we work entirely "in the box" (DAW) with virtual
instruments, the latencies are typically compensated under the
hood for us. Jitter is the measure of random variation in latency.
The more latency, the less timing accuracy (everything's hitting
late rather than on time); the more jitter, the less precision (everything's all over the place rather than "in the pocket").
Device manufacturers and producers sometimes get the
terms "accuracy" and "precision" mixed up. Let's set the
record straight...
~ "sample-accurate" means there's no latency (or fully compensated latency).
~ "sample-precise" means there's no jitter (no random timing
variance from the average latency).
Timing accuracy and precision are typically expressed in
either milliseconds (ms) or audio sample count. If a device
has 10 ms of latency, then at 48 kHz that's 480 samples,
and at 96 kHz that's 960 samples. How we precisely define
jitter is a matter of much discussion among engineers and
mathematicians, but there are some free software apps that
let you measure jitter yourself using your DAW. Innerclock
Systems offers their Grid Lock software plug-in, which characterizes jitter at 8th and 16th note resolution. If you want to
measure at a higher clock (full 24 PPQN) resolution, then you
may try Expressive Electronics' LAS. http://expressive
electronicsformusicians.com/midi_jitter.html
The degree of acceptable (perceptible) jitter is a topic of hot
debate, especially on gear forums. Some people are more
attuned, noticing timing irregularities better than others.
Particularly with percussion tracks (or anything with sharp transients), jitter can have a subconscious, detrimental affect on a
song's appeal. The general consensus is that jitter under 1 ms
may be acceptable, but anything over 2 ms becomes perceptible.
The important point is that if a groove feels right, it probably is. But if a track invokes a 'meh' response, it might be due
to jittery / sloppy event timing. If quality is important to us
and we want to make hits, then we must not settle for merely "okay" timing in our tracks.

Figure 1

Figure 2
On the grid
In the current gear literature and on online forums, you'll find mention
of sync timing being "on the grid" or not. That's just a shorthand way of
saying that musical events align on bars, beats, 16th notes, etc. The
thing to be aware of is that within the Digital Audio Workstation, there
are actually two grids: the audio grid and the MIDI grid. When sync is
correct, the events on both the audio and MIDI grids line up perfectly.
Let's take the simple case of synching a MIDI drum machine as slave
to an Ableton Live session. Live is the clock master, sending MIDI RealTime messages (Start, Stop, Continue, Clock, and Song Position Pointer)
to the hardware. The path of these MIDI messages can vary greatly
depending on the interface connections. For example, one might route
MIDI clock to a dedicated interface (via FireWire, USB, or Thunderbolt)
which then outputs to classic 5-pin DIN MIDI ports. Alternatively the MIDI
is directly routed to the drum machine via a USB cable.
Latency: the devil we know
Of the two culprits, latency is the easier one to deal with. Most
DAWs offer myriad options for compensating audio and MIDI
latencies, with both global settings and track-specific timing. You've
probably dealt with these settings at some point, but when you add

Figure 3





http://www.is.gd/RECGetInSync http://www.expressiveelectronicsformusicians.com/midi_jitter.html http://www.expressiveelectronicsformusicians.com/midi_jitter.html

Table of Contents for the Digital Edition of Recording - December 2015