Latest Version; Numerical Optimisation - Martin Audio MLA MINI Advanced User's Manual

MLA MINI
DISPLAY 2.2
Windows 10; no other operating systems are supported. We would recommend a 64 bit i7
processor running at 2.6GHz minimum with a minimum of 8GB of RAM. File sizes are not
particularly huge so a high-capacity Hard Drive is not necessary but an SSD drive will be faster
and more reliable. The mathematical calculations involved in creating optimisations use an
application called MatLab Components and when running this uses a great deal of computational
power so it is best not to scrimp on CPU power and speed when specifying a PC to run the
application.
Users have reported that Display2.2 works perfectly well on an Apple Mac using Bootcamp or
under a virtual platform such as VM Ware Fusion or Parallels, (these options still require a copy
of Windows Vista, 7, 8 or 10 to be installed) however this is not supported by Martin Audio.

LATEST VERSION

From version 2.2 onwards there is greater interaction with Vu-Net v2.0 so it is important to
use that version or later to get the best performance from one of our multicellular systems. It
is perfectly ok to use older versions of Display2 with Vu-Net v2.0, or optimisations created in
Display 2.2 with Vu-Net v1.2.5, there is complete cross compatibility but the best results will
definitely be obtained using the latest versions of both applications.
One particular step forward is a lowering of the frequencies at which the optimisation is effective.
This is always difficult as part of the optimisation process involves attempting to contain the audio
distribution within areas you have specified which is very difficult to do with lower frequencies
as they become more omnidirectional the lower the frequency. The Martin Audio R&D team have
however continued to research methods to extend the low frequency control and we learnt a great
deal developing the MLA Mini system which has always had LF included in the optimisation.
Version 2.2 now optimises lower frequencies using FIR filtering but without increasing latency
which is a remarkable achievement. Note that your multicellular system MUST be running the
latest firmware for the optimisations to be effective at low frequencies. MLA and MLA Compact
must be running 2.2.57 and 2.12.44 respectively. These were released in 2015 so all system
should be running them as a matter of course.
Note that if opening .mat files created in Display version prior to version 2.2, the low frequency
optimisation will not be used. If you wish to re-optimise older projects, the best option is to
import the venue shell and then add the array manually to create a new project with all the latest
benefits.

NUMERICAL OPTIMISATION

At this point it is worth taking a moment to explain in detail one of the concepts behind how
Display operates; Numerical Optimisation is an easily misunderstood concept, a few people
make the incorrect assumption that it is some sort of "averaging" system meaning that the result
is a compromise which couldn't be further from the truth.
What made multicellular systems possible was the development of a phenomenally accurate
computer model of the system which works hand in hand with the precision acoustic design
of the cabinets. There are a colossal number of factors that affect an array performance; the
performance of individual drivers, the horn design, the angle between enclosures, the number
of enclosures et cetera. Martin Audio have taken the analysis of array performance to the most
advanced level ever attempted, including boundary effect modelling which includes the effect
that all other cabinets in an array have on a single driver; essentially each individual cabinet has
a baffle extending above and below it formed by the other enclosures in the array. For cabinets
in the middle of the array this is fairly symmetrical but for cabinets towards the top and bottom
it becomes more and more asymmetrical which has a highly significant effect on the individual
module performance and therefore the overall array performance.
The diagrams below show the response of a single array element which is the second box down
in a six boy array. The first picture ignores the effect of its neighbouring enclosures. The second
picture shows the dramatic effect that the enclosure above and the four below have on the
response. Ignoring this effect will clearly compromise the ability to accurately predict how the
array will perform, in fact measurements show that ignoring the boundary effect can result in
MLA Mini Display 2.2 V1.0
85