Future Proof Surround Sound Mixing using Ambisonics
brochure
1. Omada is a new way to measure and control sound
developed at the Norwegian University of Science
of Technology (NTNU). We are actively seeking
industrial partners to take the patent pending
OMADA technology to the commercial market.
2. Could you please
repeat that?
What does directivity index mean to
you? What does balance between DI
and white noise imply? An easier
question is what does quality of sound
mean to you, let’s say in a business
meeting by video conference?
An everyday situation is that far too
many miss important parts of the dis-
cussion. This might be because of poor
sound quality or because a disturbing
source of noise is nearby.
Sorry, what
did you say…?
To deal with this everyday situation
you need equipment that can handle
all sources of noise and isolate voices
from each other. Now you will be able
to clearly distinguish and hear your two
colleagues clearly in a video conference.
- Despite the disturbing background
noise in their office 2000 miles away.
I can´t hear you!
What you will find if you search for it,
is hardware microphones that is really
good for a unidirectional recording of
your voice. You will then of course need
a microphone for each and every partici-
pant in a phone call, and the system is
rigid.
We dear say that you will not find equip-
ment that is able to efficiently deal with
and process the natural dynamically
soundscape that surrounds us in our
lives. There is just far too many sources
of sound for the equipment to handle
and return good sound quality to the
listener.
Applications
Advantages
• All signal processing functions integrated in one single optimization
framework
• Maximum directivity and optimization of wanted sound/voices
• Robust and flexible array processing
• Lower processing power and memory requirements
• Computationally efficient in implementation and steering
• Easy to integrate into existing communication devices
The algorithm is optimized for applications using
a microphone array in a number of fields:
• Audio/speech communication
• Video conference
• Room acoustics
• Music recording
• Soundfield analysis
• Underwater acoustics
Now there is a software/algorithm with muscles and
intelligence enough to make use of the microphones to
handle real life situations. -Meet OMADA!
Loud and clear!
3. An optimal microphone array directiv-
ity algorithm - OMADA for short - is a
beamforming framework implement-
ing spherical arrays that introduces
new optimization methods for signal
processing. This new method gives a
detailed analysis of the sound field in
the room, and allows you to point your
microphone array towards the important
areas of the soundscape, while sup-
pressing external noise sources.
The OMADA technology can be used
in combination with advanced micro-
phones to give a precise and desired
pattern of directivity.
OMADA can in combination with
spherical microphone arrays intelligently
capture all the desired near-end voices,
and at the same time automatically
suppress all the significant interferences
coming from the three dimensional (3D)
directions other than that of the desired
voices.
As mentioned earlier, could for example
two parties in a video-conference avoid
being disturbed by a noisy roof fan with
the OMADA technology.
The technology has been developed
at the Department of Electronics and
Telecommunications at the Norwegian
University of Science and Technology
in Trondheim, Norway.
The technology
The Directivity Index (DI) tells by how
many dB a microphone enhances one
direction compared to all other direc-
tions (on average), so a higher number
is better. Directivity is important
because it helps indicate how much
sound will be directed towards a spe-
cific area compared to all the sound
energy being generated by a source.
For example the directivity pattern can
be manipulated to subdue noise as well
as echo, reverberation or interference,
all in whatever direction it might come
from.
4. A visual illustration of a directivity example for a micro-
phone above a meeting table, based on the use of a spheri-
cal array of order 2 (this picture) and 5 (right most picture).
With OMADA one can make do with fewer microphones
while maintaining excellent directivity. This also means
lower hardware cost.
The main ability of OMADA is to give less unwanted noise/sound to the listener. One of the other abilities of the OMADA algorithm is to enhance the directivity
by, for example, filtering noise from sound in 3D space. The pictures to the left and right show the directivity in all directions in the room shown in the middle.
In this example, the look direction (maximum sensitivity) is defined with the small red circle, and the directivity is coded in colors with red as high sensitivity,
and dark blue as low sensitivity.
From its location, you can specify the microphone equip-
ment “to look for” or focus on sound from exact locations
in space, by using OMADA algorithm. Then it is up to you
to choice either to enhance or suppress the source.
Here one can see how able OMADA is to process the
sound signals in order to enhance the preferred sound.
When increasing the order from 2 to 5 (this picture), the
performance is improved as the look direction reduces
in width, and the sidelobes get highly attenuated.
0 dB
-2
-4
-6
-8
-10
-12
-14
-16
-18
-20
-22
-24
-26
-28
-30 Db
OMADA at work
5. 0
5
10
15
20
Suppression[dB]
Sidelobe suppression with broad main lobe
150 1000 8000
Frequency [Hz]
OMADA
non−optimized
Predicted, order 3
0
5
10
15
20
Suppression[dB]
Sidelobe suppression with single main lobe
150 1000 8000
Frequency [Hz]
OMADA
non−optimized
Predicted, order 2
Sidelobe suppression with single main lobe
When a directional microphone has one enhanced direction (the main lobe), then all the
other directions will not be suppressed equally much. The sidelobe suppression specifies
how many dB, at least, the other directions are suppressed. This means that a higher num-
ber of dB is better. Over a wide frequency range, the OMADA algorithm gives around 5 dB
higher suppression.
Sidelobe suppression with broad main lobe
A spherical microphone array has large flexibility, and it is possible to shape a “broad”
main lobe, picking of several talking persons at once. The sidelobe suppression for such
a directivity pattern also shows a clear advantage for the OMADA algorithm. In general,
the more complicated directivity pattern is wanted, the better the advantage might be with
OMADA.
OMADA at work
6. Directivity index for 2.nd and 3.rd order
A spherical microphone array has problems to maintain a high DI at low frequencies, and
the OMADA algorithm makes it straightforward to give a gentle roll-off towards the lower
frequencies, and well-behaved values at high frequencies. The non-optimized microphone
might give good values for a limited frequency range, but breaks down at low frequencies
(a DI below 0 dB is actually worse than the simplest possible, single microphone).
An order-3 microphone can give a higher DI than an order-2 microphone.
For comparison: A Cardioid has a DI of 4.8 dB, a Supercardioid 5.7 dB,
and a Hypercardioid 6.0 dB.
0
5
10
15
20
DirectivityIndex[dB]
Directivity Index, Spherical array
150 1000 8000
Frequency [Hz]
OMADA
non−optimized
Theoretical, order 2
0
5
10
15
20
DirectivityIndex[dB]
Directivity Index, Spherical array
150 1000 8000
Frequency [Hz]
OMADA
non−optimized
Theoretical, order 3
OMADA at work
7. Best paper award IEEE WASPAA 2009
The authors received the Best Paper Award out of 181 paper submissions at the biennial
IEEE WASPAA conference 2009 (www.waspaa2009.com) for their paper “Robust spherical
microphone array beamforing with multi-beam-multi-null-steering, and sidelobe control”.
The inventors
Prof. Peter Svensson, NTNU
Professor Svensson has more than 20 years of experience in acoustics research from Swe-
den, Japan, Canada and Norway. He is now a professor at the Department of Electronics
and Telecommunications at the Norwegian University of Science and Technology in Trond-
heim.
PhD Haohai Sun, NTNU
PhD Sun has written his Phd on the invention and has won both academic and innovation
awards for his work. As a senior DSP/CAE engineer Sun has gained experience with algo-
rithm implementation and optimization. He finished in 2011 his PhD at the Department of
Electronics and Telecommunications at the Norwegian University of Science and Technol-
ogy in Trondheim.
The Norwegian University of Science and Technology (NTNU) in Trondheim represents
academic eminence in technology and the natural sciences as well as in other academic
disciplines. Our 7 faculties and 53 departments educate over 20.000 students, and every
year over 3300 master and doctoral degrees are awarded. Over 200 man-years in academic
or scientific positions work or about 2000 R&D projects at all time, at more than 100 labo-
ratories.
NTNU Technology Transfer is the contract partner and has exclusive rights to the commer-
cialization of all the IPR at the university. We are a 100% owned and subsidiary of NTNU,
and have over 50 active projects at any point in time. To learn more about our available
technologies, please visit our website: www.tto.ntnu.no
Contact information
For more information about NTNU Technology Transfer
or interest in this technology, please contact:
Knut Wilhelm Knutsen
NTNU Technology Transfer AS
knut.knutsen@ntnu.no
Cellphone +47 92460646