Designer Notebook: This Time It’s Personal
Rane MM 42 processor outfitted for IEM needs
This Designer Notebook as submitted by Rane Corporation. Live Sound makes every effort to eliminate any use of marketing inspired hyperbole.
|
|
 |
Not since a venerable roadie turned a loudspeaker to face the stage has a technology completely revolutionized what and how a performer hears more than in-ear personal monitoring systems (IEM). |
As IEM technology improves and prices continue to drop, performers of all calibers, from weekend worship bands to world-class touring performers, are making the switch to what is often called “ears” for short.
|
|
|
The introduction of dual and triple driver transducer designs, improved construction techniques, and the use of passive filters and crossover networks to optimize the frequency response for musical applications are some of the more recent IEM improvements. A host of affordable universal fit designs are also now available, making personal monitors accessible to those who can’t afford custom molded designs. |
Even with these improvements, there’s one element that cannot be completely
factored into any design human hearing. Every performer hears differently
due to their physiology, preferred listening volume, years of abuse standing
next to the drummer (insert favorite joke here), etc.
To compensate for these personal differences, monitor engineers use outboard
compressors, equalizers, and limiters to achieve satisfactory ear mixes.
Multiply the required equipment by the number of processed mixes and it’s
easy to see how things can get expensive and fill precious rack space
in a hurry.
THE WISH LIST
This is where the new Rane MM 42 monitor processor enters the picture,
conceived to combine flexible, quality equalization and dynamics processing
into a single rack-space device with several features, such as an on-board
headphone amp and cue bus link between devices, to meet the requirements
of both artists and engineers.
The MM 42 list of features resulted from input acquired from touring sound
companies and A-list monitor engineers. Without question, the product
went through more re-definition and re-design stages than any other in
recent company history as we developed a better understanding of this
application’s unique requirements. The wish list included:
• IEM-specific signal processing: shelf/cut filters, multi-band compression, parametric equalization, and multi-band peak limiting.
• The ability to process a single stereo mix or two independent mono mixes (or each side of a stereo mix individually).
• Subwoofer output for feeding bass “shakers” or powered subwoofers.
• Beefy on-board headphone amp for cueing each mix pre- or post-processing.
• Cue bus link between devices for accessing any monitor mix without re-patching headphones.
• Direct outs (pass through) for each input for distributing a common mix to multiple units.
Figure 1 shows a block diagram for the MM 42. An input matrix
is used to mix and route incoming signals to each of three output-processing
chains: out 1, out 2, and sub (woofer). Each main output processing chain
offers (in order of signal flow): high and low shelf/cut filters, three-band
compressor, five-band parametric EQ, variable output level and a three-band
peak limiter.
Processing for the sub output is simpler, consisting only of low- and
high-cut filters for creating a band-limited mix without requiring an
external crossover. This low-frequency mix is then fed directly to a powered
bass shaker or subwoofer arrangement.
|
|
|
MAKING ASSIGNMENTS |
In a dual mono arrangement, input A might feed out 1 as a mono mix for
the guitarist, while input B feeds out 2 for the drummer. Input B would
also be assigned to the sub output, which connects in turn to a powered
tactile shaker mounted to the drum throne.
The extra inputs are useful when there are performers on stage who share
a similar basic blend of instruments backup singers, for example but
each require their own (slightly different of course) mix of solo instruments
or voice. Rather than duplicate the same generic band mix across multiple
aux sends, create a single stereo band mix and feed it to two of the inputs
of the first MM 42.
Then, use the direct outputs of this MM 42 to daisy chain the generic
mix to the inputs of the second performer’s MM 42, use the direct outputs
of this MM 42 to feed the inputs of the third performer’s MM 42, and so
on. The direct outs are active-buffered, so there is no signal loss even
when daisy chaining across multiple devices.
With the band mix distributed to each performer, the additional MM 42
inputs are used to augment each performer’s mix with the solo mix for
each. The relative balance between band mix and solo mix for each performer
can be adjusted using the front panel input level controls.
There are, of course, other creative uses for the extra inputs: mix in
FOH, monitor world, or on-stage talkback feeds so musicians and engineers
can communicate with each other; add a blend of ambient microphones to
each performer’s mix; or feed a specific mix of instruments just kick
drum and bass, for example to the sub output.
FILTERING OPTIONS
Once the inputs are mixed and routed, signal passes through a basic filter
section consisting of 12 dB-per-octave Butterworth filters with selectable
shelf or cut characteristics. Shelf filters function as simple bass and
treble tone controls to add some extra thump or air to the overall mix
and are especially useful when working with monitor consoles lacking dedicated
EQ for each group output. Alternately, the filters can be set to cut mode
to band-limit the main output mix, as might be done when using a combination
of IEMs and a bass shaker or subwoofer.
|
|
|
A multi-band compressor was chosen to provide precise control of the dynamics of each frequency range within a mix. Strapping a broadband (full range) compressor across the entire
mix has the inherent problem that the loudest signal in any frequency
range a strong, low frequency kick drum, for example dominates
the compressor’s response, resulting in an unnatural pumping and
breathing effect. |
The solution is to divide the audio signal into distinct bands and compress
each frequency range separately. For example, it is possible to tighten
up just the low frequency components of a mix say 200 Hz and below
without affecting the mid and high frequency regions.
The MM 42’s compressor section uses a phase-compensated, 24 dB/ octave
Linkwitz-Riley crossover to split the audio signal into three user-defined
frequency bands. Each band is then compressed to taste with its own associated
threshold, ratio, attack and release controls before being summed back
together and passed on to the parametric EQ section. The compressor offers
a fixed soft-knee characteristic centered about the threshold region.
The five-band parametric equalizer section makes it possible to achieve
a variety of effects, ranging from simple sweetening of mixes to precise
boosts or cuts to compensate for anomalies in a performer’s hearing or
a particular earpiece’s response. In talking to the various transducer
manufacturers, it quickly became obvious that there is no one magic curve
for a particular brand of earpiece or user.
The only thing all manufacturers agree on is that flat response is NOT
the goal when equalizing earpieces. Most earpieces are designed to achieve
a particular response curve, and outboard equalization is used to further
tailor this curve to the performer’s subjective taste or hearing needs.
During development of the MM 42 a Bruel & Kjaer (B & K) 4157 Ear Simulator
was used in conjunction with an Audio Precision test system to accurately
measure the frequency response of several commercially available universal
fit earpieces. The B & K device simulates the volume and shape of the
main ear canal and presents the earpiece with acoustic impedance approximating
that of the human ear.
Figure 2 and Figure 3 show the before and after response
graphs respectively of a popular model of earpieces equalized flat to
within 1 dB using the MM 42.
WHAT DID YOU SAY?
Noise-induced hearing loss is an unfortunate reality for many performers
who have spent years being blasted by guitar amps, drums, and traditional
wedge-based systems. Often, this hearing loss is asymmetrical, with each
ear exhibiting a slightly (or sometimes radically) different frequency
response. To account for this disparity it is necessary to balance the
response between ears by equalizing each side differently.
To address this requirement, the processing parameters for each of the
main output chains can be manually linked or unlinked as desired. For
example, it’s possible to unlink the parametric EQ section and apply different
EQ curves to each side of the mix, yet keep the compressor section linked
to prevent image shift during gain reduction.
|
|
|
Given that IEM systems couple the transducer directly within a performer’s ear, some form of “brick wall” limiting is required to prevent excessively loud sounds regardless of their duration from reaching the performer at full volume. Most beltpacks have a built-in limiter that works well to protect against unexpected transients or to prevent an overzealous performer from dialing their mix to 11. |
However, these limiters do nothing to prevent the RF (radio frequency)
overmodulation effects and distortion caused by overdriving the front
end of a wireless transmitter. Proper gain structure and limiter use upstream
of the transmitter input are the only sure ways to mitigate such effects.
The MM 42’s three-band peak limiter section is located immediately before
the main outputs and sets the absolute maximum level fed to a downstream
transmitter or wired beltpack. A three-way crossover splits the signal
into user-defined bands, as is done in the compressor section.
A subtle but important design aspect is that the limiter and compressor
do not share a common crossover. The user can choose completely different
frequency settings for the limiter than the compressor, thereby optimizing
each dynamics section. It’s also possible to disable the multi-band feature
and use each section in a simple, full range mode.
STILL SUBJECTIVE
By design, the MM 42 does not offer any form of reverb, stereo image enhancement,
or audio delay. Reverb is a highly subjective area, and many engineers
interviewed eschew the use of reverb on an entire mix, preferring instead
to use a quality outboard reverb to sweeten the “money channels” and add
this effect to the rest of the mix.
The jury’s also still out on stereo enhancement and delay; some users
say spatializing the mix or delaying the ear mixes to align with the bleed
from the house PA helps separate vocals from instruments and improve intelligibility
but just as many others say these features are unnecessary. We decided
to leave them off for the time being and wait until a more general consensus
is reached. (Plus we were running out of room on the front panel for more
buttons...)
Instead, we focused on offering a number of extras to make the MM 42 as
easy as possible to set up and use in the heat of a show. A partial list
of features includes: momentary and latching modes for the cue bus, having
the gain reduction meter automatically switch to follow the currently
selected compressor or limiter band, and fast switching between mixes
when working in dual mono mode.
An on-board headphone amplifier was considered an absolute necessity.
Because the MM 42 connects in-line between the monitor console and a wired
or wireless monitor system, the engineer needs some means of easily cueing
(soloing) the processed mix, without resorting to y-cable splits or other
such complicated arrangements.
Wireless transmitter systems have an on-board headphone amplifier, but
it’s only feasible to listen to one mix at a time. You must constantly
re-patch headphones between transmitters to listen to various mixes.
The MM 42 includes both 1/4-inch phone and 1/8-inch mini jacks on the
front panel, allowing connection of all types of headphones and earpieces.
Both jacks can even be used simultaneously if desired. For the budget
conscious artist (or in an emergency situation), the MM 42’s headphone
amp can even take the place of a hardwired beltpack system for the less
mobile performers on stage.
In addition to activating the headphone feed, pressing the front panel
cue bus button routes the selected mix to the rear panel cue bus and cue
out connectors. Figure 4 shows a typical touring setup consisting
of a monitor console, multiple MM 42s, and a mix of wired and wireless
monitor systems.
In this example, the MM 42s are connected together via the cue bus link.
Once this link between devices is established, the engineer simply plugs
her headphones or earpieces into any MM 42 (it doesn’t matter which one)
and presses the front panel Cue Bus button of the appropriate MM 42 to
solo that performer’s mix.
|
|
|
It’s not necessary to constantly re-patch headphones when moving between different mixes. Within each device, the signal sent to the cue bus is selectable pre- or post-processing, allowing the monitor engineer to A/B the end result without affecting the artist’s mix (the main outputs are always post-processing). |
Figure 4 also shows an interesting use for the dedicated cue outs.
In order to hear exactly what the performer hears, a spare transmitter/beltpack
combination is connected to the cue outs, allowing the engineer to monitor
the signal both post-processing and post-transmitter.
Also, should a performer’s transmitter or beltpack fail during a show
it’s a logical matter to cue the affected performer’s MM 42 mix and swap
beltpacks to use the spare system. Alternately, the cue outs connect to
unused channels on the monitor console, allowing the engineer to solo
the processed Cue Bus signal, individual channel inputs, or the pre-processed
group outputs.
BETTER LATE THAN NEVER
Digital signal processing (DSP) offers features, flexibility and cost
benefits that can’t be fully realized with traditional analog designs.
All processing in the MM 42 uses fixed-point Motorola DSP components and
converters operating at 48 kHz sample rate with full 24-bit resolution.
Internally, DSP calculations use 48 bits for maximum accuracy.
An obvious benefit of using DSP came late in the MM 42’s production schedule.
The initial design placed the output level control after the limiter section,
with a maximum level of 0 dB (attenuate only).
However, many users asked for the output level control to be located before
the limiter and provide additional gain so they can turn the processed
mix level up without affecting the compressor settings, with the limiter
still determining the maximum overall output level.
|
|
|
Such a change would have required a significant hardware change
in the analog days; instead, it took just a few minutes to restructure
the DSP code and a simple firmware update via MIDI to modify existing
units. |
Latency, or propagation delay as it’s often called, is an unavoidable
attribute of any digital system. Latency is the amount of time it takes
to convert an analog signal to its digital representation, process it
in some fashion (compress, EQ, limit), then convert it back to an analog
signal. It’s measured in milliseconds (ms), and is dependent on the chosen
hardware and the amount and efficiency of internal processing performed
on the signal.
As more digital devices enter the signal path between microphone and ear,
engineers must be conscious of the latencies introduced by digital consoles
and outboard processors and strive to keep the overall system latency
as short as possible typically 5 ms or less. Any longer and the processed
signal becomes distinguishable from the body’s natural voice feedback
due to bone conduction, creating an unnecessary distraction for the performer.
With that in mind, the MM 42 was designed with a fixed propagation delay
of 1.5 ms, which is well within acceptable limits when combined with the
2 to 3 ms latencies common to digital consoles.
FITTING IT ALL IN
How do you fit six tons of stuff into a one-ton can? We’re not sure either,
but the MM 42 posed a similar challenge. The front panel includes a large
display, six encoders, twelve momentary contact switches, two headphone
jacks, a level-control potentiometer and forty-one light emitting diodes.
The rear panel is a veritable jack farm, with four Neutrik XLR/TRS combo
jacks, two XLR jacks, seven TRS jacks, MIDI jacks, an IEC power connector
and RJ 12 jacks for the Cue Bus. Take all of this hardware, mix in a universal
internal power supply, bring to a boil while stirring constantly... and,
hopefully, it all fits. Well, it almost did.
The clearance on our standard metal chassis was short by 0.15 of an inch.
A new sheet metal chassis was developed to allow enough clearance for
the stacked TRS jacks and front panel assembly.
Choosing a connector for the cue bus link between units was a bit of a
challenge. A means of connecting multiple MM 42s together was needed,
but we also wanted to differentiate the cue bus from the cue outs so users
wouldn’t inadvertently connect the current-summing Cue Bus to console
inputs. A standard RJ 12 telephone-style jack was chosen, thereby satisfying
the “Can I buy a replacement cable at Radio Shack?” requirement mandated
by all experienced touring professionals.
Since few sound engineers carry their own telephone-style cables to gigs,
the MM 42 ships with short cables for the cue bus; in a pinch, any normal
telephone cord with at least four conductors will work. And no, it’s not
possible to dial up the MM 42 and mix the show from home...
Another challenge was finding enough space for the large number of LEDs
used in the input, output and gain-reduction meters. The maximum possible
circuit board size was 1.5 inches by 4.25 inches and, just to add to the
fun, the shafts for five of the encoders were in the way.
Obviously the LED driver chips and a surface-mounted connector would have
to go on the backside of the board. As seemed to be par for the course,
double-sided surface mount boards were a new challenge.
But the good people in our manufacturing department were willing to give
it a go and indeed were able to pull it off. The double-sided board stacks
perfectly over the board for the pushbuttons and the encoders.
At this point, the front panel was coming together nicely except for the
headphone jacks. With only an inch of space remaining on the right side
of the front panel, it seemed impossible to fit a 1/4-inch phone jack,
1/8-inch mini phone jack and a level potentiometer.
After a bit of head scratching, an effective solution was proposed: rotate
the headphone board 90 degrees so it sits vertically in the available
space, then mount the 1/8-inch jack on the back side of the board, beneath
the 1/4-inch jack.
Wouldn’t you know, it all fit.
Sheldon Radford is a product technology engineer and Michael Rollins is a senior digital design engineer for Rane Corporation.
January 2004 Live Sound International
