Xaoc Devices ODESSA User manual
ODESSA
variable
spectrum har-
monic cluster
oscillator
Model of 1975
operator’s manual rev. 1975/1.0
2
SALUT
Thank you for purchasing this Xaoc Devic-
es product. Odessa is an additive oscillator,
which means the output signal is synthesized
by adding a multitude of sinusoidal compo-
nents (up to 2560 harmonic partials). By
manipulating their parameters it is possible
to obtain a broad range of unearthly sounds
as well as classic saw, square, and pure sine.
Odessa offers a set of controls for shaping har-
monic spectra based on number of partials
and their distribution in frequency and ampli-
tude; all of which is illustrated by a spectrum
analyzer comprising 12 multicolor LEDs.
The series of harmonics can be squeezed or
spread apart, tilted, or pruned by a comb-like
frequency response, resulting in a huge vari-
ety of spectra. Animating the comb response
phasing. All partials are frequency-related to
a common fundamental and controlled by a
single volt/octave input. Additionally, the sig-
nal can be frequency modulated by exponen-
tial and linear (through-zero) means. Also, up
a fat and dense cluster, or a powerful chord.
INSTALLATION & SETUP
The module requires 24hp worth of free space
in the eurorack cabinet. Always turn the pow-
er off before plugging the module into the bus
board using the supplied ribbon cable. Pay
close attention to power cable pinout and
orientation. The red stripe indicates the neg-
ative rail and should match the dot or –12V
mark on both the bus board and the unit.
Odessa is internally secured against reversed
power connection, however, rotating the 16-
pin header may cause serious damage to
other components of your system because
it will short circuit the +12V and +5V power
lines. Always pay particularly close attention
to the proper orientation of your ribbon cable
on both sides! Also, observe that there are sev-
eral pin headers on the board. connecting
the power cable to an incorrect head-
er will destroy your odessa! The unit
should be fastened by mounting the supplied
screws before powering up. To better under-
stand the device, we strongly advise the user
to read through the entire manual before use.
MODULE OVERVIEW
-
rect access to all parameters in a one-knob-
per-function arrangement. Observe that it
also follows the traditional synthesis layout,
wherein pitch and voicing are controlled on
the left side, the main timbral features are
centrally located, and additional effects are
controlled at the right. Signal outputs are
located in the bottom array of jacks. The arc
of multicolor LEDs offers a rough overview of
the spectrum of the signal, from very low to
the highest audible frequencies.
Pitch frequency is controlled via the pitch
cv v/oct input 1which accepts voltages
in –5V…+10V range. The coarse 2and
fine 3pair of potentiometers set the pitch
throughout the entire audible range (16Hz to
20kHz) without the need for external voltage.
Additionally, pitch can be modulated from the
exp fm input 4that accepts ±5V, with depth
controlled by the dedicated attenuator above
5. As with other Xaoc Devices products, the
slider LED illuminates to show the absolute
degree of modulation by lighting up for both
negative and positive voltages. Bear in mind
that while the pitch knobs alone cover the
entire frequency range, at extreme knob po-
sitions modulation from the exp fm input will
module
explained
3
not change the pitch any further. This limita-
tion does not include the pitch cv v/oct in-
put which is handled by a separate precision
A/D converter.
The big, central partials knob 6limits the
harmonic partials, from 1 to 512 per voice.
This limit can also be controlled by CV using
the jack below 7. Note that the response
appears to be stepped, especially at the be-
ginning of the range, because it causes con-
secutive partials to be turned on and off. At
the minimum position, only the fundamental
frequency is audible. Also note that due to au-
tomatic volume compensation, lower frequen-
cies become quieter as higher components are
added to the spectrum.
The red spectral tilt knob 8controls how
quickly the amplitudes of consecutive partials
decrease with frequency. At the middle posi-
tion, partials decrease slowly, similar to the
spectrum of a saw wave. At the minimum po-
sition, the decrease is so rapid that mostly the
fundamental is audible. At the maximum posi-
additional comb response is engaged), which
results in a buzzy, narrow pulse waveform.
This parameter can be also controlled by ex-
ternal CV (±5V) via the dedicated jack 9.
The three knobs near the right edge of the
panel control a comb-like frequency response
that is imposed on the spectrum. Note there is
certain settings), instead, a frequency-domain
shaping function is applied to the amplitude
of each harmonic partial. The density knob
10 controls how dense the notches of the comb
are: from zero (no notches at all) through
moderate (just a few notches) to a maximum
-
tered out (assuming minimum warp setting).
This parameter can be controlled by external
CV via the dedicated jack 11 that accepts
±5V, and is scaled by the slider potentiometer
above 12 . The warp knob 13 controls the
uniformity of the comb response: from linear,
where notches are equidistant in frequency
-
ear response, where notches are very dense
at the bottom of the spectrum and become
more distant for higher overtones (similar to
a phaser effect). This parameter can also be
controlled by CV via the dedicated jack 14
that accepts ±5V, and is scaled by the slider
potentiometer above 15 . The peaking knob
16 controls the shape of the comb response:
from narrow notches (at the minimum posi-
tion) through moderate up to wide notches
with narrow peaks in the response (at the
maximum position). This parameter can be
controlled by a ±5V CV via the jack below 17 .
tension 18 is a very important and sensitive
parameter that determines whether the sinu-
soidal partials generated by Odessa adhere to
a harmonic pattern wherein frequencies are
strictly integer multiples of the fundamental
frequency. There is a little dead zone in the
central position of this knob that helps to set
it to zero. With tension set to sharp (above
the middle), partials are more spread apart so
that their frequencies increase more quickly
than with a linear law (e.g. the harmonic se-
ries) and the spectrum becomes sparse. With
tension
are more condensed so that their frequencies
increase slowly and create a dense non-har-
monic cluster that may resemble noise. This
parameter can be controlled by a ±5V CV via
the jack below 19 .
4
Odessa features two main outputs of the
synthesized signal: odd partials 20 and
even partials 21 . It is possible to split the
harmonic spectrum so that even and odd
numbered partials are separately present at
those outputs yet always mixed with the fun-
damental partial. An additional fundamen-
tal output 22 offers a simple signal of the
fundamental frequency: either a sinusoid or a
square wave that can be employed for syncing
selects one of these two options.
The two knobs at the top ( 23 and 24 ) to-
gether with their associated CV inputs
( 25 and 26
address the distribution of spectral compo-
nents between the odd partials and even
partials outputs. When the bank length
parameter is set to 0, both outputs offer the
same full signal. If set to +1, each output of-
fers the same fundamental (1st partial) plus
its even (2nd, 4th, 6th, etc) and odd (3rd, 5th, 7th,
etc) overtones, respectively. With different
lengths, odd and even sequences of partials
are split between the two jacks. Addition-
ally, all partials except the fundamental
may be frequency-scaled by an integer fac-
tor: from simple fractions (1:2, 1:3 down to
1:8) to multiples (2, 3… up to 8), selected by
harmonic factor. The direction of bank
length (either turned left or right) selects
which of the two outputs will receive the
Odessa is capable of delivering 1, 3, or 5
stacked voices of its synthesized signal, se-
lectable by the voices button 27 . When a
single voice is selected (button lit green),
the spread knob 28 and its corresponding
CV jack below 29 have no effect. With 3 or 5
voices (button lit orange or red), the spread
front panel
overview
2
19
2518
3
28
31
6
1
4
30
29
5
5
the interface
13
10
14
15
12
16
9
2623 2432 8
27
20 1122721
17
6
parameter controls the degree of symmetric
detuning of the additional voices around the
central voice.
The lin fm input 30 together with the as-
sociated slider 31 offers a deep through-ze-
ro frequency modulation of the signal. The
modulator input is AC coupled (cut below
20Hz) and accepts full-bandwidth signals up
to 10Vpp. Bear in mind that while the funda-
mental frequency is modulated to the degree
you set, the overtones are modulated much
wider because the depth scales with their
relative frequency. With a wideband carrier,
the spectrum of an FM’d signal explodes into
MHz range and most of it will be removed by
the anti-aliasing protection. Classic clangor-
ous FM sounds are obtained with just a few
harmonic partials. note: Place the slider at
minimum when no modulation is applied in
values read by the A/D converter which
could impact pitch stability.
SPECTRUM ANALYZER
The arc of 12 multicolor LEDs 32 shows the
power density spectrum which is the
name of the power contribution of differ-
ent frequency components of the signal in a
number of disjoint bands. Here, the 12 bands
cover the entire audible frequency range in
exponentially spaced intervals (0.8 octave
per band): below 35Hz, 35 to 63Hz, 63Hz to
113Hz, 113Hz to 204Hz, 204Hz to 367Hz,
367Hz to 661Hz, 661Hz to 1.19kHz, 1.19kHz
to 2.14kHz, 2.14kHz to 3.85kHz, 3.85kHz to
6.94kHz, 6.94kHz to 12.5kHz, above 12.5kHz.
Certainly, with only 12 bands it offers only a
crude overview of what is going on. note:
the color temperature is mapped from dB
scale. Although the LEDs turn off below a
certain level, this does not mean there are
no spectral components in a given band, but
rather that they are too quiet to show.
THE MEANING OF SPECTRUM
PARAMETERS
The various parameters offered by Odessa
have been selected by observing how spec-
tra of different sounds vary, and how these
differences could be generalized to a set of
global features without losing the ability to
synthesize a broad range of sounds. The de-
the most common waveform in synthesizers:
the sawtooth wave. It became so popular be-
cause it is quite easy to generate in analog
circuits, and also because it’s a good starting
point for many synthetic timbres.
The partials parameter controls the num-
ber of harmonic components in a signal,
from 1 to 512. Turning it down limits the
spectrum to the initial N partials, until the
note: you can’t turn off the fundamental.
However, since it is separately available on
a dedicated output, you can subtract it from
your signal using a simple patch.
WORKING CLASS ELECTRONICS
MADE IN THE EUROPEAN UNION
XAOCDEVICES.COM
POWER
CONNECTOR
STRIPE
FUNDAMENTAL OUT
SINE
SQUARE
V·OCT
LIN FM
NULL
TR 1 TR 2
REV. 06.2019
1975 VARIABLE SPECTRUM
HARMONIC CLUSTER OSCILLATOR
CAUTION! BEFORE CONNECTING
THE EXPANDER MODULES PLEASE
REFER TO THE MANUAL! IMPROPER
CABLE ORIENTATION WILL
CAUSE SERIOUS DAMAGE!
CAUTION! DO NOT ATTACH THE
POWER CONNECTOR TO ANY OF THE
EXPANDER HEADERS!
LEIBNIZ
SUBSYSTEM
TERMINAL
STRIPE
HEL
EXPANDER
fundamental output
configuration jumper
So as to avoid aliasing, Odessa does not pro-
duce partials whose frequencies would exceed
the maximum frequency of 21kHz. Therefore,
the usable range of the big knob depends on
pitch frequency. For high-pitched sounds, the
frequency of most of the overtones would be
too high, hence increasing this parameter
above a certain number will not create any
audible effect.
There is a fundamental problem with addi-
tive synthesis as it relates to dynamic range:
a sum of many sinusoids may be much loud-
er than just a single sinusoid (or a few). For
can exceed 50dB. For practical reasons,
Odessa applies a perceptually optimized vol-
ume compensation to the output signal. You
may notice that the low-frequency partials
become quieter when the energy of higher
partials increases. The result is similar to
what you hear when comparing the loudness
of different waves from a traditional VCO.
The tilt parameter determines how quickly
the spectrum decays (how quickly the ampli-
7
The spectrum of a saw wave contains all overtones in a naturally decaying harmonic series: the am-
plitude of each harmonic partial is inversely proportional to its number: An=A1/n (A1is the amplitude
-
ic partials missing because their spectra are shaped by a Sinc function which introduces a series of
notches: An=A1
spectrum
parameters
0 0.5 1 1.5 2
1
0.5
0
-0.5
Time
Amplitude
0 0.5 1 1.5 2
1
0.5
0
-0.5
-1
Time
5 10 15 20 25 30
1
0.8
0.6
0.4
0.2
Partial number
0
05 10 15 20 25 30
1
0.8
0.6
0.4
0.2
Partial number
Amplitude Amplitude
sawtooth and pulse waveforms and their respective harmonic spectra (truncated for clarity).
Amplitude
8
tudesofpartialsdecrease withfrequency).This
-
in the partial amplitude formula: An=A1/n.
This parameter can change from 3 (very
quick decay, dull sound), through 1 (like in
the sawtooth wave) down to nearly 0 (almost
impact on the resulting energy of the signal.
Together, the density, warp, and peak-
ing parameters control the comb-like fre-
quency response imposed on the spectrum.
The notches of the comb are produced by a
warped Sinc function in the frequency do-
-
fers to the relative frequencies of partials, so
it scales with pitch. Depending on the den-
sity parameter, there may be zero to 256
notches, hence at minimum, the spectrum
is smooth, and at maximum, each second
partials remain), provided there is no warp.
Thus, with all other parameters set to de-
fault, turning the density knob morphs the
but quite differently than a simple crossfade!
0 0.5 1 1.5 2
Time
1
0.5
0
-0.5
-1
Amplitude
0 0.5 1 1.5 2
Time
1
0.5
0
-0.5
-1
Amplitude
5 10 15 20 25 30
Partial number
0
1
0.8
0.6
0.4
0.2
Amplitude
5 10 15 20 25 30
Partial number
0
1
0.8
0.6
0.4
0.2
Amplitude
sawtooth wave after limiting the spectrum
to the initial 5 partials
sawtooth wave after limiting the spectrum
to the initial 10 partials
Warping the comb response results in a
non-uniform distribution of the notches in
frequency. As warp increases from zero,
the notches become more concentrated in
low frequencies. note: at high density and
warp values, notches may be so dense in low
frequencies that they interfere with the har-
monic pattern which may lead to unexpected
holes in the spectrum. At low peaking val-
ues, notches of the comb response are very
narrow. At high peaking values, the notches
become wider and peaks become more nar-
The tension parameter has a crucial impact
on the harmonicity of the signal. At the neu-
tral (middle) position, frequencies of all si-
nusoidal partials are integer multiples of the
fundamental: Fn=n×F1, which is a necessary
condition for obtaining a periodic waveform.
This results in equidistant partials through-
out the spectrum. With tension above 0, the
upper partials are spread apart (the frequen-
cies increase quicker than the partial num-
bers) which yields an inharmonic, metallic
timbre with an often more sparse spectrum
tension below 0, the distances
between upper partials become smaller and
smaller (the frequencies increase slower than
the partial numbers) which yields a dense,
rough, inharmonic cluster that resembles
noise. Large negative values of tension may
even result in the spectrum folding over itself
to a degree where certain partials have lower
frequencies than the fundamental.
caution: Pitch and intonation behaves par-
adoxically with these inharmonic sounds and
-
ditional sounds. Furthermore, these sounds
can cause ear fatigue and are best used spar-
ingly in musical contexts.
Bear in mind that inharmonic spectra yield
aperiodic waves due to individual sinusoids
being no longer synchronized in phase. When
you turn the tension parameter off the cen-
tral (zero) position, the relative phases of
all signal components begin to drift away
from each other. Thus the original waveform
becomes more distorted and will remain as
such, even after returning tension to zero. It
is possible to re-synchronize the phase and
restore the waveform, however, this produc-
es an audible click in the signal due to discon-
9
AmplitudeAmplitude
5 10 15 20 25 30
Partial number
0
1
0.8
0.6
0.4
0.2
5 10 15 20 25 30
Partial number
0
1
0.8
0.6
0.4
0.2
5 10 15 20 25 30
Partial number
0
1
0.8
0.6
0.4
0.2
Amplitude
comparison of tilted spectra in extreme and middle
positions of the tilt knob
10
tinuity, therefore it is not done automatical-
ly. To do so, press the voices button. Besides
selecting the number of unison voices, this re-
syncs the individual voices and also restarts
all sinusoids within each voice.
signal, the sinusoidal partials generated
by Odessa may be organized into harmon-
ic banks and split between the odd par-
tials and even partials outputs. bank
length
partials above the fundamental that belong
to even and odd banks, respectively. har-
monic factor
partials within one of the banks. note: the
fundamental (1st partial) is excluded from
the banks and is always present on both out-
puts with its original frequency unaffected.
For example, with length set to 3, the odd
partials output contains the fundamental
mixed with 5th, 6th, 7th, 11th, 12th, 13th ... etc,
while the even partials output contains
the fundamental mixed with 2nd, 3rd, 4th, 8th,
9th, 10th… etc. Additionally, if factor is set to
2, all partials (except the fundamental) on
the even partials output will have their
0100 200 300 400 500
Relative frequency
1
0.8
0.6
0.4
0.2
Amplitude
0100 200 300 400 500
Relative frequency
1
0.8
0.6
0.4
0.2
Amplitude
0100 200 300 400 500
Relative frequency
1
0.8
0.6
0.4
0.2
Amplitude
0100 200 300 400 500
Relative frequency
1
0.8
0.6
0.4
0.2
Amplitude
warped comb response with minimum peaking (upper)
and maximum peaking values
tilt + comb frequency response for low (upper)
and high (lower) density values
11
05000 10000 15000 20000
Frequency [Hz]
1
0.8
0.6
0.4
0.2
Amplitude
05000 10000 15000 20000
Frequency [Hz]
1
0.8
0.6
0.4
0.2
Amplitude
EXPANDABILITY
The shrouded expansion header is dedicated
to the universal Xaoc Hel expander mod-
ule which offers paraphonic generation of
3-voice and 5-voice chords in Odessa. Please
refer to the manual of Hel for details.
Odessa also offers integration with the Leib-
niz Binary Subsystem via another expansion
header at the back of the unit. Connecting
an expander like Xaoc Lipsk, allows one to
enable and disable the individual groups
bank length
parameter. To engage this control, press and
hold the voices button for 1 second until it
starts to blink. At this point, a high state of
each binary line (for example, a button acti-
vated on Lipsk) turns one group on. The low-
est bit (B0nd, 3rd,
and 4th), the next higher bit (B1) affects the
second group (5th, 6th, and 7th), and so on. The
highest bit (B7) affects all remaining partials
(from 23rd up).
note: the result of switching individual par-
tials may be subtle. For a more radical effect
Amplitude
5 10 15 20 25 30
Partial number
0
1
0.8
0.6
0.4
0.2
Amplitude
5 10 15 20 25 30
Partial number
0
1
0.8
0.6
0.4
0.2
signal spectra at the odd partials and even partials
outputs for length=3 and factor=2
inharmonic spectra resulting from positive (upper)
and negative (lower) values of tension
MAIN
FEATURES
Powerful additive
synthesis engine
with up to 2560
sinusoidal
partials
0.5Hz to 21kHz
frequency range
with resolution
of 0.006Hz
Harmonic
or inharmonic
spectra shaped
by tilt and
warped comb
response
Volt/octave
pitch control
Through-zero
linear FM input
Up to 5 unison
and detuneable
voices
Three signal
outputs
TECHNICAL
DETAILS
Eurorack synth
compatible
24hp, 63mm depth
Current draw:
+110mA/ -80mA
Reverse power
protection
EASTERN BLOC TECHNOLOGIES MADE IN THE EUROPEAN UNION
WARRANTY TERMS
XAOC DEVICES WARRANTS THIS PRODUCT TO BE FREE OF DEFECTS IN MATERIALS OR WORKMANSHIP,
AND TO CONFORM WITH THE SPECIFICATIONS AT THE TIME OF SHIPMENT FOR A PERIOD OF ONE YEAR
FROM THE DATE OF PURCHASE. DURING THAT PERIOD ANY MALFUNCTIONING OR DAMAGED UNITS
WILL BE REPAIRED, SERVICED, AND CALIBRATED ON A RETURN-TO-FACTORY BASIS. THIS WARRANTY
DOES NOT COVER ANY PROBLEMS RESULTING FROM DAMAGES DURING SHIPPING, INCORRECT INSTAL-
LATION OR POWER SUPPLY, IMPROPER WORKING ENVIRONMENT, ABUSIVE TREATMENT OR ANY OTHER
OBVIOUS USER-INFLICTED FAULT.
LEGACY SUPPORT
IF SOMETHING WENT WRONG WITH A XAOC PRODUCT AFTER THE WARRANTY PERIOD IS OVER, NO
NEED TO WORRY, AS WE’RE STILL HAPPY TO HELP! THIS APPLIES TO ANY DEVICE, WHEREVER AND
WHENEVER ORIGINALLY ACQUIRED. HOWEVER, IN SPECIFIC CASES, WE RESERVE THE RIGHT TO CHARGE
FOR LABOR, PARTS AND TRANSIT EXPENSES WHERE APPLICABLE.
RETURN POLICY
THE DEVICE INTENDED FOR REPAIR OR REPLACEMENT UNDER WARRANTY NEEDS TO BE SHIPPED IN
THE ORIGINAL PACKAGING ONLY, SO PLEASE KEEP IT JUST IN CASE. ALSO, A FILLED RMA FORM MUST
BE INCLUDED. XAOC DEVICES CAN NOT TAKE ANY RESPONSIBILITY FOR DAMAGES CAUSED DURING
TRANSPORT. PRIOR TO SENDING US ANYTHING, PLEASE CONTACT US AT SUPPORT@XAOCDE-
VICES.COM. NOTE THAT ANY UNSOLICITED PARCEL WILL BE REJECTED AND RETURNED!
GENERAL INQUIRIES
FOR USER FEEDBACK SUGGESTIONS, DISTRIBUTION TERMS, FEEL FREE TO CONTACT XAOC DEVICES AT
THE CURRENT PRODUCT LINE, USER MANUALS, FIRMWARE UPDATES, TUTORIALS, AND MERCHANDISE.
WRITTEN bY M. bARTKOWIAK. PROOfREADINg AND EDITINg bY b. NOLL. DESIgNED bY M. ŁOjEK.
ALL RIGHTS RESERVED. CONTENT COPYRIGHT © 2019 XAOC DEVICES. COPYING, DISTRIBUTION OR
ANY COMMERCIAL USE IN ANY WAY IS STRICTLY PROHIBITED AND REQUIRES THE WRITTEN PER-
MISSION BY XAOC DEVICES. SPECIFICATIONS ARE SUBJECT TO CHANGE WITHOUT PRIOR NOTICE.
Other manuals for ODESSA
1
Table of contents
Other Xaoc Devices Synthesizer manuals
Xaoc Devices
Xaoc Devices ROSTOCK User manual
Xaoc Devices
Xaoc Devices SEWASTOPOL User manual
Xaoc Devices
Xaoc Devices KARL MARX STADT User manual
Xaoc Devices
Xaoc Devices BELGRAD User manual
Xaoc Devices
Xaoc Devices KARL MARX STADT User manual
Xaoc Devices
Xaoc Devices KATOWICE User manual
Xaoc Devices
Xaoc Devices WARNA 1983 VOLTAGE DISTRIBUTOR II User manual
Xaoc Devices
Xaoc Devices JENA User manual
Xaoc Devices
Xaoc Devices HEL User manual
Xaoc Devices
Xaoc Devices SAMARA User manual
