Martin Audio MLA MINI, Руководство для продвинутых пользователей

Страница: 86 / 256

MLA Mini Display 2.2 V1.0
86
MLA MINI
DISPLAY 2.2
inaccuracies of up to 8dB which fairly obviously would make the system unusable.
Martin Audio have perfected the computer model for our Multicellular systems to such a degree
that the predicted performance can reach accuracy of as close as 1dB of measured results which
has made the whole system not only possible but extremely accurate.
Numerical optimisation uses the computer model to check the performance for a given
mechanical orientation or DSP configuration, and compares the result on the venue slice to
what you have asked the system to achieve. It will then try an alternative option to see if it is
getting closer or further from the desired result. If it is further from what you want it to achieve
it will ignore that configuration and pursue the more successful angles or EQ curves, trying
configuration after configuration before it has attempted close to every possible combination and
has got as close as possible to the desired result.
If you consider the first optimisation which calculates the splay angles, for an MLA system there
are eight possible angles available between each cabinet; 0.5, 1, 2, 3, 4, 5, 6 and 7.5° If you
have a reasonable size array of say 16 enclosures that means a total of 8 x 8 x 8 x….8, (fifteen
8’s!) or 35,184,372,088,832 possible combinations of angles between all cabinets. Clearly
analysing each of these combinations involves a substantial amount of processing but pales into
insignificance compared to the hundreds of thousands of calculations required when running
numerical optimisation for the DSP; every band of EQ frequency, bandwidth and gain, phase
and FIR filters for every one of the cells of every cabinet. This is why a powerful PC is required
and the EQ optimisation can take up to fifteen minutes or so for a large array in a complex
arena. Even the most powerful PC available today couldn’t try every possible combination in a
reasonable time scale so complex algorithms are used so that when a certain sequence of filter
parameters are consistently demonstrating results further from the goals that have been set, the
system will abandon those avenues and try others. With this system highly accurate parameters
can be produced that are as close as possible to the desired goals in a reasonable time frame.
This is a very basic explanation, in fact it is Martin Audio’s development of the revolutionary
but phenomenally complex digital algorithm that the application uses to refine the optimisation
which give the best possible results in a reasonable time frame. Not only that but key parts of
the optimisation process are multithreaded which means that the system is able to exploit the
latest 64-bit, quad core processors to speed up the entire process.
Whilst this system is processor hungry clearly it is a very realistic and practical method to obtain
the best possible results from an array where so many variables are at your disposal. Trying to