BW Broadcast DSPXtra-FM Quick start guide

DSPXtra
incorporating the
Broadcast Audio Processor
Operational Manual
Version 2.30
www.bwbroadcast.com
RMS Leveller

Table Of Contents
2
Table Of Contents
Warranty 3
Safety 4
Forward 6
Introduction to the DSPXtra-FM 7
DSPXtra-FM connections 8
DSPXtra-FM meters 9
DSPXtra-FM status LEDs 10
Quickstart 11
An introduction to audio processing 12
Ariane multiband leveller 13
Source material quality 16
Pre-emphasis 16
The DSPXtra-FM and its processing structure 17
The processing path 17
Processing block diagram 19
User levels 20
The DSPXtra-FM menu system overview 21
The menu structure 22
Processing parameters 23
Setting up the processing on the DSPXtra-FM 30
Getting the sound that you want 38
Managing presets 42
Factory presets 43
Remote control of the DSPXtra-FM 45
Remote trigger port 52
Security code locks 53
Clock based control 54
Specifications 55
USB drivers Appendix A
Preset Sheet Appendix B

Warranty
3
Warranty
BW Broadcast warrants the mechanical and electronic components of this product to be free of defects in mate-
rial and workmanship for a period of one (1) year from the original date of purchase, in accordance with the war-
ranty regulations described below. If the product shows any defects within the specified warranty period that are
not due to normal wear and tear and/or improper handling by the user, BW Broadcast shall, at its sole discretion,
either repair or replace the product.
If the warranty claim proves to be justified, the product will be returned to the user freight prepaid.
Warranty claims other than those indicated above are expressly excluded.
Return authorisation number
To obtain warranty service, the buyer (or his authorized dealer) must call BW Broadcast during normal business
hours BEFORE returning the product. All inquiries must be accompanied by a description of the problem. BW
Broadcast will then issue a return authorization number.
Subsequently, the product must be returned in its original shipping carton, together with the return authorization
number to the address indicated by BW Broadcast. Shipments without freight prepaid will not be accepted.
Warranty regulations
Warranty services will be furnished only if the product is accompanied by a copy of the original retail dealer’s
invoice. Any product deemed eligible for repair or replacement by BW Broadcast under the terms of this war-
ranty will be repaired or replaced within 30 days of receipt of the product at BW Broadcast.
If the product needs to be modified or adapted in order to comply with applicable technical or safety standards
on a national or local level, in any country which is not the country for which the product was originally devel-
oped and manufactured, this modification/adaptation shall not be considered a defect in materials or workman-
ship. The warranty does not cover any such modification/adaptation, irrespective of whether it was carried out
properly or not. Under the terms of this warranty, BW Broadcast shall not be held responsible for any cost result-
ing from such a modification/adaptation.
Free inspections and maintenance/repair work are expressly excluded from this warranty, in particular, if caused
by improper handling of the product by the user. This also applies to defects caused by normal wear and tear, in
particular, of faders, potentiometers, keys/buttons and similar parts.
Damages/defects caused by the following conditions are not covered by this warranty:
Misuse, neglect or failure to operate the unit in compliance with the instructions given in BW Broadcast user or
service manuals.
Connection or operation of the unit in any way that does not comply with the technical or safety regulations
applicable in the country where the product is used.
Damages/defects caused by force majeure or any other condition that is beyond the control of BW Broadcast.
Any repair or opening of the unit carried out by unauthorized personnel (user included) will void the warranty.
If an inspection of the product by BW Broadcast shows that the defect in question is not covered by the war-
ranty, the inspection costs are payable by the customer.
Products which do not meet the terms of this warranty will be repaired exclusively at the buyer’s expense. BW
Broadcast will inform the buyer of any such circumstance. If the buyer fails to submit a written repair order within
6 weeks after notification, BW Broadcast will return the unit C.O.D. with a separate invoice for freight and pack-
ing. Such costs will also be invoiced separately when the buyer has sent in a written repair order.
Warranty transferability
This warranty is extended exclusively to the original buyer (customer of retail dealer) and is not transferable to
anyone who may subsequently purchase this product. No other person (retail dealer, etc.) shall be entitled to
give any warranty promise on behalf of BW Broadcast.
Claims for damages
Failure of BW Broadcast to provide proper warranty service shall not entitle the buyer to claim (consequential)
damages. In no event shall the liability of BW Broadcast exceed the invoiced value of the product.
Other warranty rights and national law
This warranty does not exclude or limit the buyer’s statutory rights provided by national law, in particular, any
such rights against the seller that arise from a legally effective purchase contract. The warranty regulations men-
tioned herein are applicable unless they constitute an infringement of national warranty law.

BW Broadcast Technical Manual Page 4
Safety Instructions
DETAILED SAFETY INSTRUCTIONS:
All the safety and operation instructions should be read before the appliance is operated.
Retain Instructions:
The safety and operating instructions should be retained for future reference.
Heed Warnings:
All warnings on the appliance and in the operating instructions should be adhered to.
Follow instructions:
All operation and user instructions should be followed.
Water and Moisture:
The appliance should not be used near water (e.g. near a bathtub, washbowl, kitchen sink, laundry tub, in a
wet basement, or near a swimming pool etc.).
The appliance should not be exposed to dripping or splashing and objects filled with liquids should not be
placed on the appliance.
Ventilation:
The appliance should be situated so that its location or position does not interfere with its proper ventilation.
For example, the appliance should not be situated on a bed, sofa rug, or similar surface that may block the
ventilation openings, or placed in a built-in installation, such as a bookcase or cabinet that may impede the
flow of air through the ventilation openings.
Heat:
The appliance should be situated away from heat sources such as radiators, heat registers, stoves, or other
appliance (including amplifiers) that produce heat.
Power Source:
The appliance should be connected to a power supply only of the type described in the operating instructions
or as marked on the appliance.
Grounding or Polarization:
Precautions should be taken so that the grounding or polarization means of an appliance is not defeated.
Power-Cord Protection:
Power supply cords should be routed so that they are not likely to be walked on or pinched by items placed
upon or against them, paying particular attention to cords and plugs, convenience receptacles and the point
where they exit from the appliance.
Cleaning:
The appliance should be cleaned only as recommended by the manufacturer.
Non-use Periods:
The power cord of the appliance should be unplugged from the outlet when left unused for a long period of
time.
This symbol, wherever it appears,
alerts you to the presence of
uninsulated dangerous voltage inside
the enclosure—voltage that may be
sufficient to constitute a risk of shock.
This symbol, wherever it appears, alerts
you to important operating and mainte-
nance instructions in the accompanying
literature. Read the manual.
CAUTION: To reduce the risk of electrical shock, do not
remove the cover. No user serviceable parts inside. refer
servicing to qualified personnel.
WARNING: To reduce the risk of fire or electrical shock, do
not expose this appliance to rain or moisture.
1.3 SaFEty InStrUCtIOnS

BW Broadcast Technical Manual Page 5
Safety Instructions
Object and Liquid Entry:
Care should be taken so that objects do not fall and liquids are not spilled into the enclosure through open-
ings.
Damage Requiring Service:
The appliance should be serviced by qualified service personnel when:
- The power supply cord or the plug has been damaged; or
- Objects have fallen, or liquid has been spilled into the appliance; or
- The appliance has been exposed to rain; or
- The appliance does not appear to operate normally or exhibits a marked change in performance; or
- The appliance has been dropped, or the enclosure damaged.
Servicing:
The user should not attempt to service the appliance beyond that is described in the Operating Instructions.
All other servicing should be referred to qualified service personnel.
CE CONFORMANCE: This device complies with the requirements of the EEC Council
Directives: 93/68/EEC (CE Marking); 73/23/EEC (Safety – low voltage directive);
2004/108/EC (electromagnetic compatibility). Conformity is declared to those standards:
EN50081-1, EN50082-1.
WARNING: This equipment generates, uses, and can radiate radio frequency energy.
If not installed and used in accordance with the instructions in this manual it may cause
interference to radio communications. It has been tested and found to comply with the
limits for a Class A computing device (pursuant to subpart J of Part 15 FCC Rules),
designed to provide reasonable protection against such interference when operated in
a commercial environment. Operation of this equipment in a residential area is likely to
cause interference, at which case, the user, at his own expense, will be required to take
whatever measures may be required to correct the interference.
CANADA WARNING: This digital apparatus does not exceed the Class A limits for radio noise emis-
sions set out in the Radio Interference Regulations of the Canadian Department of Communications.
Le present appareil numerique n'emet pas de bruits radioelectriques depassant les limits applicables
aux brouillage radioelectrique edicte par le ministere des Communications de Canada.

Forward
6
FOrWard
Thank you for your purchase of the DSPXtra-FM digital audio broadcast processor.
Over the last decade the staff at BW have observed broadcast processors from afar with fascination and intrigue
and shared a keen interest in digital audio processing. In January 2002 we decided on a whim that it was time to
have a go at designing our own digital audio processor. We knew that if we were going to design an audio proc-
essor we had to do it the BW way and make the processor the most cost effective fully featured all-in-one broad-
cast processor on the planet. The one aim, to make an audio processor that offered all of the features found in
the other more costly processors but at a fraction of the price. That processor was the DSPX!
Why DSPX? Like most products the concept is conceived nameless. The name DSPX came about for several
reasons. Firstly, the BW team couldn't decide on a name we liked so the concept became DSP processor X. As
the project developed it soon became clear that the processor was to be a processor for all seasons and so in
true algebraic fashion X represents many things. DSPX was born.
18 months later and it was time for a more advanced model. We bring you, the DSPXtra-FM. The Xtra is the
result of continued research and development and the inclusion of the award winning Ariane RMS leveller con-
cept from TransLanTech sound of New York city.
DSPXtra-FM is built-on DSPX-FM but with Extra added features. More DSP horsepower, improved distortion
control, six bands of peak limiting and that all important award winning Ariane RMS leveller.
Once again, thank you for your purchase; we hope you enjoy the DSPXtra-FM!
BW Broadcast Team

Introduction
7
IntrOdUCtIOn tOthE dSPXtra-FM
The BW Broadcast DSPXtra-FM is a new generation of digital audio signal processor that can be used to proc-
ess audio ready for FM and digital broadcasting such as DAB, HD Radio or internet streaming.
Using the latest multi-band DSP technology the DSPXtra-FM offers a versatile and powerful tool in creating a
loud punchy on-air presence.
The DSPXtra-FM has been designed and built from scratch using a new approach to the design of a digital audio
processor that incorporates the most up to date components. Cutting edge technologies allow the DSPXtra-FM
to produce similar results to other processors in the market but in a simpler more cost effective way. The advanc-
es we have made have allowed us to pass the savings on to our customers.
What's Under the Lid?
The DSPXtra-FM is driven by a fast 8 bit micro-controller which controls an array of specialised analogue and
digital circuits: These include 24-bit A/D and D/A converters, analogue level control circuitry, 22 x 24 bit DSP's,
an ethernet port, a trigger port, a USB port, an RS232 port, over 200 LEDs (metering), an LCD screen, 2 sample
rate converters, a headphone jack and memory devices to hold the software and firmware.
The Processing Architecture!
After input selection the 24-bit digital audio signal is passed through conditioning circuitry before being pre-
sented to the Ariane RMS levelling block. The four band RMS leveller corrects for input level variations and also
improves consistency. The block also can provide stereo enhancement with its 'matrixed' mode of operation.
The output of the AGC feeds the EQ and bass enhancement sections before being split into six bands by linear
phase time aligned filters. The six bands are processed by dynamic audio limiters on each band. The unique
dual processing paths allow simultaneous processing for FM and digital radio. Look-ahead limiting and distor-
tion cancelling clipping ensures your signal is kept to a strict maximum while maintaining crystal clear sound. A
supersonic sample rate DSP stereo encoder provides MPX generation with fantastic stereo separation.
The easy to use front panel control system and LED metering display, afford the user with ease of use and
setup.
Comprehensive control of every processing parameter is available to the user both from the front panel control
system and by remote (computer) control.
At a fraction of the size, weight and price of its rivals, the DSPXtra-FM is a clear winner.
Dynamic, fresh & innovative...
... The DSPXtra-FM

Connections
8
LOAD PRESET
REMOTE
AES/EBU
EDIT
UP
LOAD
SAVE
DSPXtra
OUTPUT
PROCESS
INPUT
UP
LOAD
U1=BASS FACE INS
DEL
DONE
ENTER PASSCODE
INS
DEL
DONE
ENTER NEW PASSCODE
3779
U2=ROUND FACE
U3=CLASSICAL 3
(A)
SAVE PRESET UP
GO
U1=BASS FACE
U2=ROUND FACE
U3=CLASSICAL 3
(A)
INS
DEL
DONE
NAME PRESET U2
MY PRESET 3
UP
LOAD
SAVE
DSPXtra
OUTPUT
PROCESS
INPUT
LI MITERS OUTPUT
ARIANE LEVELERINPUT
DSPXtra
100
90 80 70
60
50
40 30 20 10
PILOT
00
-3 -6 -9
-12 -15 -18 -21 -24 -27 -30 -33 -36
-39
-42
-1 -2 -3 -4 -5 -6 -7 -8 -9
-10 -11 -12 -15 -18
-6 -4 -2 -0
+2 +4 +6 +8
+10 +12 +14 +16 +18
GATE
CLIP GATE
+18 +16 +14 +12 +10 +8 +6
+4
+2 -0 -2 -4 -6
HOLD
-2 -4 -6 -8
-10 -12 -14 -16 -18 -20 -22 -24
HOLD -2 -4 -6 -8 -10 -12 -14 -16 -18 -20 -22 -24
123456
dB dB dB %
RL RL
MPX
+ / L - / R
IDR IDR
1 2 3 4
+ / L - / R + / L - / R + / L - / R
+20 +20
DSPXtra-FM
ANALOGUE
INPUTS
SCA
INPUT
MPX
OUTPUT
PILOT
OUTPUT
SYNC
INPUT
AES / EBU
INPUT
RS232
&
TRIGGER PORT
LAN
PORT
GND
LIFT
MAINS POWER INPUTANALOGUE
OUTPUTS
HEADPHONE JACK METERING USB
PORT
STATUS
LEDS
SOFT
KEYS
CONTROL
KNOB
AES / EBU
OUTPUT
DSPXtra-FM FRONT AND REAR PANELS

Meters
9
dSPXtra MEtErS
The DSPXtra has several hundred LEDS that provide ‘always on’ instant IO and processing metering.
I/O metering
The input meters show the level of the input audio. The meters are ‘hooked in’ to the DSP code after the input
level selection and mode options. The clip LED’s represent the onset of the A/D convertors clip-point and these
LED’s should not light under any circumstances. Adjust the input gain control of the DSPXtra if they do.
The output meters represent the level in dB below full scale output. This output level is the peak output level of
the processing and has nothing to do with the actual output level set by the analogue and digital output level
options.
The output meters show a smaller dynamic range compared to the input ones. This reflects the smaller dynamic
range of the audio once processed by the DSPXtra. If we were to have used the same scale as the input meter-
ing we would not see a lot of activity on the LED’s, with all of the LEDS on most of the time.
The multiplex output metering represents the composite outputs peak level. This is a representation of the out-
put in relation to the peak composite level of the processing and not the actual level set by the multiplex output
level control.
The IO meters follows an approximation of the PPM level of the audio waveform while a floating ‘peak hold’ dot
tracks the absolute value of the waveform
G/R metering - Ariane leveller
The Ariane metering displays gain being applied with more leds indicating more gain, the metering having eight
columns, two columns for each band. The scale is -6 to +20dB in 2dB steps.
When the leveller is operating in matrix mode the left hand column of each band (1 through 4) represents the
Right plus Left component (aka the SUM signal) while the right hand column of each band represents the Left
minus Right component of the audio (aka the DIFFERENCE signal).
When the leveller is operating in stereo mode the left hand column of each band represents the left channel and
the right hand column represents the right channel.
When in bypass the metering returns to the 0dB gain position to indicate unity gain through the leveller
(bypassed).
There are two other leds above each gain column. One is the gate for the band and the other is the IDR for the
band. More information on the gate and IDR is contained in the processing setup section of this manual.
G/R metering - Six band limiter
The six band limiter displays gain reduction from 0dB to -24dB in 2dB steps.
There is only one meter per stereo channel and the value shown is the largest gain reduction of the left and right
channels. Under normal operation (with a stereo audio feed) this is fine but you may observe strange metering if
the channels are not very balanced in level or you are using the DSPXtra to process two separate mono feeds.
The red LEDS above each gain reduction meter indicate HOLD when lit. More information on HOLD function is
contained in the processing setup section of this manual.
LOAD PRESET
REMOTE
AES/EBU
EDIT
UP
LOAD
SAVE
DSPXtra
OUTPUT
PROCESS
INPUT
UP
LOAD
U1=BASS FACE INS
DEL
DONE
ENTER PASSCODE
INS
DEL
DONE
ENTER NEW PASSCODE
3779
U2=ROUND FACE
U3=CLASSICAL 3
(A)
SAVE PRESET UP
GO
U1=BASS FACE
U2=ROUND FACE
U3=CLASSICAL 3
(A)
INS
DEL
DONE
NAME PRESET U2
MY PRESET 3
UP
LOAD
SAVE
DSPXtra
OUTPUT
PROCESS
INPUT
L I MITERS OUTPUT
ARIANE LEVELERINPUT
DSPXtra
100
90
80
70
60
50
40
30
20
10
PILOT
00
-3
-6
-9
-12
-15
-18
-21
-24
-27
-30
-33
-36
-39
-42
-1
-2
-3
-4
-5
-6
-7
-8
-9
-10
-11
-12
-15
-18
-6
-4
-2
-0
+2
+4
+6
+8
+10
+12
+14
+16
+18
GATE
CLIP GATE
+18
+16
+14
+12
+10
+8
+6
+4
+2
-0
-2
-4
-6
HOLD
-2
-4
-6
-8
-10
-12
-14
-16
-18
-20
-22
-24
HOLD
-2
-4
-6
-8
-10
-12
-14
-16
-18
-20
-22
-24
123456
dB dB dB %
RL RL
MPX
+ / L - / R
ID R I DR
1 2 3 4
+ / L - / R + / L - / R + / L - / R
+20 +2 0
DSPXtra-FM

Status LEDs
10
StatUS LEdS
The DSPXtra-FM front panel contains three status LEDS.
REMOTE: Indicates that the DSPXtra-FM is currently talking to a remote computer. This will flash
during an update of the firmware or remote control with the remote control application.
AES/EBU: Indicates the presence of a valid AES/EBU signal connected to the digital audio input
of the DSPXtra-FM.
EDIT: Indicates that you are currently editing a parameter.
FUSE
LOAD PRESET
LOW HIGH
MIDHMIDL LOW HIGH MIDHMIDL
WB
LRLRMPX
CLIP CLIPCLIP
GATE
INPUT AGC OUTPUTLIMITERS
REMOTE
AES/EBU
EDIT
0
-3
-6
-9
-12
-18
-24
-30
-36
-42
-48
-56
-2
-4
-6
-8
-10
-12
-14
-16
-18
-20
-22
-24
-2
-4
-6
-8
-10
-12
-14
-16
-18
-20
-22
-24
0
-3
-6
-9
-12
-18
-24
-30
-36
-42
-48
-56
110
100
90
80
70
60
50
40
30
20
10
PILOT
dB
dB
dB
dB %
RS232
DSP X
UP
LOAD
SAVE
DSPX
OUTPUT
PROCESS
HOLD
INPUT
UP
LOAD
U1=BASS FACE INS
DEL
DONE
ENTER PASSCODE
INS
DEL
DONE
ENTER NEW PASSCODE
3779
U2=ROUND FACE
U3=CLASSICAL 3
(A)
SAVE PRESET UP
GO
U1=BASS FACE
U2=ROUND FACE
U3=CLASSICAL 3
(A)
INS
DEL
DONE
NAME PRESET U2
MY PRESET 3
UP
LOAD
SAVE
DSPX
OUTPUT
PROCESS
INPUT

Quickstart
11
QUICk Start
1. Install the processor into the rack.
2. Connect AC power to the unit, and turn on the power.
3. Connect the analogue and / or digital audio inputs.
4. Select the analogue or digital input as the source of the processing with the 'INPUT SELECTION' parameter
which can be found in the 'INPUT' menu. Apply audio and observe the input meters. For analogue inputs, adjust
the 'INPUT’ menus 'INPUT LEVEL CONTROL' so that the input meters do not clip. We recommend setting the
mixing board or audio source to full level output (even clipping) prior to adjusting this control. This ensures that
the processor's A/D converter will not clip under any conditions.
FM USE:
5. Select the pre-emphasis setting for your region (input menu). 75 µs for USA and 50 µs for Europe.
6. Connect the audio outputs as required and set the output level and de-emphasis settings for the analogue and
digital outputs to match any external links, stereo encoders, or transmitters that require left and right audio inputs
or an AES/EBU input. Make sure the output mode is set to FM for the output in question.
7. If you are using the MPX Output (preferred), navigate to the ‘STEREO’ menu and adjust the ‘MPX OUTPUT
LEVEL’ to match the transmitter (or link device) that follows the processor. Adjust for 100% modulation with
audio.
8. Select a factory preset (see Managing presets).
dr USE (daB/hd radIO/drM/FMEXtra/StrEaMInG):
5. Set the pre-emphasis setting to OFF from the input menu.
6. Connect the audio outputs as required and set the de-emphasis for the analogue and digital outputs to OFF,
MODE to DR and set the output levels for the analogue and digital outputs to match the equipment that the proc-
essor is connected to.
7. Select a factory preset (see Managing presets).
8. Navigate to the look-ahead menu and adjust the SHELF EQ control to suit. This sets the brightness of the DR
outputs sound.
dUaL USE: FM + daB/hd radIO/drM/FMEXtra/StrEaMInG:
5. Select the FM pre-emphasis setting for your region (input menu). 75 µs for USA and 50 µs for Europe.
6. Connect the audio outputs as required and set the de-emphasis for the analogue and digital outputs to OFF,
MODE to DR and set the output levels for the analogue and/or digital outputs to match the DR equipment that
the processor is connected to (EG: codec/computer).
7. If you are using the MPX Output, (preferred) navigate to the ‘STEREO’ menu and adjust the MPX ‘LEVEL’
Output to match the transmitter (or link device) that follows the processor. If one of the analogue or digital out-
puts needs to feed another piece of equipment that cant take a composite MPX input then make sure you select
FM mode for that output with the appropriate de-emphasis setting to match the corresponding piece of a equip-
ment in the FM chain.
8. Select a factory preset (see Managing presets).
9. Navigate to the look-ahead menu and adjust the SHELF EQ control to suit. This sets the brightness of the DR
outputs sound.
NOTE: The front panel headphone jack connects to the analogue outputs so the sound may be excessively
bright if pre-emphasis is engaged and de-emphasis on the analogue outputs is set to off.

Introduction To Audio Processing
12
IntrOdUCtIOn tOaUdIO PrOCESSInG
Most audio processors use a combination of compression, limiting and clipping to 'funnel' the dynamic range
down, reducing the peak to average ratio in each stage. A cascaded arrangement of compressor, limiter and
clipper produces the best results. The first stage of processing usually operates in a slow manner, the process-
ing getting progressively faster and more aggressive as the audio passes through the chain. The instantaneous
peak clipper or look-ahead limiter is the final stage of the chain and sets the final peak level.
The images below illustrate a section of audio as it passes through a typical audio processor.
The first image to the right is an unprocessed section of audio.
The images that follow represent compression of the input wave-
form, followed by limiting and then finally peak clipping.
Compression
Compression reduces the dynamic range of the audio waveform
slowly in a manner similar to a trained operator riding the gain.
Compression is usually performed on the RMS level of the audio
waveform and the ratio of compression is usually adjustable.
Compression is usually gated to prevent gain riding and suck-up
of noise during silence or quiet periods.
Limiting
Limiting is a faster form of compression that employs faster time
constants and higher ratios to produce a denser sound while
controlling peaks based upon the peak level of the audio wave-
form. Excessive limiting can create a busier packed wall of sound
effect.
Clipping
Clipping the audio waveform will not produce any audible side
effects if performed in moderation. Excessive clipping will pro-
duce a form of distortion that produces a tearing or ripping
sound. Clipping can also be used as an effective method of high
frequency peak control when used in conjunction with distortion
controlling filtering.
Look-ahead limiting
Often used instead of a clipper in systems that feed bit rate reducing audio codecs, look-ahead limiting exam-
ines the audio waveform and prepares a gain control signal in advance of the delayed audio waveform arriving.
This prevents overshoots while minimising distortion. A look-ahead limiter behaves in the same way as a soft
clipper. Competent look-ahead limiters will usually be of the multi-band variety.

Ariane
13
thE tEChnICaL and thE PhILOSOPhICaL BaCkGrOUnd OF thE arIanE
dIGItaL aUdIO LEvELLEr
(Or, Why Automatic Level Control Doesn’t Have to be Processing’s Weak Link!)
What is Levelling?
Levelling is the automatic process of reducing the long-term dynamic variations in level of an audio source. Also
called Automatic Gain Control or, AGC.
Why Use Levelling?
Levelling is required for multiple purposes, but the most important is to force a given program’s level variations
to best conform to the requirements of human hearing of the consumer in their expected listening environment.
Average levelling control is done on the basis of the audio signal’s average level over time, while peak levelling
control is based upon the highest instantaneous value. While peak amplitude value can be determined by simple
measurement of the audio’s absolute value at any given moment in time, average control can only be derived by
analysis over a specified period of time.
Levelling for effect
Levelling can be for artistic purposes, or simply as a control utility. These two separate purposes are distinction
are not usually made as most processors’ effect is taken as a inseparable. The Ariane was created specifically
to allow control without any effect. Automatic level control can be implemented in ways that are very audible.
For our purposes, however, levelling is being considered purely for its merit as a utility for the maintenance or
improvement of audible perception. The Ariane has been designed to have little or no sound effect of its own
The Arianes level-managed output will allow the following dynamics processing to work with an extremely con-
sistent signal, so that the expected effect will itself be more consistent and predictable.
Levels and Human Hearing
It is accepted among many learned students of human perception that human hearing, complex as it is, is more
sensitive to the long-term average of sound level than it is to instantaneous peaks. The human ear/brain per-
ception system integrates sound levels over time. That is, a sound of a given level that is sustained for a length
of time will sound louder than a shorter time burst of the same level. Depending upon duration and intensity, a
particular sound can actually be lower in level than another, yet be perceived to sound louder. The sensitivity of
human hearing to time as well as amplitude is critical when considering how to control listening levels over time.
RMS Average
One very useful way to determine average level is called Root Mean Squared, or RMS for short. This is a math-
ematical process where the audio’s absolute value is squared, then an average of this signal is taken over a
given length of time, then the square root is taken of that average. The result is a control signal that corresponds
to the power level of the audio over time, rather than its peak level.
RMS measurement, with its requisite integration, corresponds nicely with human hearing. By using an RMS sys-
tem of measurement and control, a level controller can work remarkably well in conjunction with human hearing.
Thus, by choosing to analyse our signal with RMS detection, we can build a simple system whose control is
much less distracting to the human ear than a simple peak-based control system.
The Nature of Program Audio
The creation of programs and presentations of audio generally requires the mixing and level adjustment of multi-
ple sources. It is the nature of ‘raw’ unprocessed audio sources (such as live microphones) to be wide ranging in
their levels, consistency and other qualities. These sources are the ones that most obviously will require a level-
ler’s automatic assistance. Yet it is also the nature of audio that some sources, such as produced music, will be
very consistent in level and quality, even while their initial absolute level may be unknown. These sources will
not require as much control, if any, as the unprocessed material. What is most typical of all, is that these multiple
types of sources may need to be conformed to each other before their final combined presentation to the end
user. Taking advantage of coincidental properties of human hearing and program audio, it doesn’t take a genius
to note that the audio for programming should be presented in a way that humans find most easy to accept.
While extreme loudness variations are disturbing, corrective control systems themselves can add effects that are
equally if not more disturbing. In other words, in audio levelling, the cure can be worse than the illness. By seek-
ing out and utilizing only those aspects of audio control that are utilitarian AND pleasing to the ear, while keeping
at arm’s length those methods of audio control that are generally found to be distracting, and then matching this
knowledge base with an understanding of the nature of typical program audio, an intelligent control system can
be designed that best meets the needs of successful, listenable programming over the long term.
Windowing Release
While we humans respond to average level in determining loudness, we also crave detail. That is, minute varia-

Ariane
14
tions create the ‘flavour’ of the sound, and help distinguish, for example, a flute from an oboe. Details ‘live’ in the
fast variations that are mostly confined to the peaks, so preserving the fast variations while making level correc-
tions only relatively slowly maintains this audio flavour. In the overall scheme of audio processing, the less we
rely upon fast peak control and the more we rely upon slow average control, the more life-like our audio will be.
If confining our wide-range level changes to a slow control will reduce the destruction of audio detail, then stop-
ping the control altogether totally eliminates any possibility of detail eradication.
How And Why Windowing Release works
In a simple level control, when audio gets too loud it is turned down, and when it is too soft it is turned up. But
what if we had a third possibility, i.e., if it is at the proper level, do nothing?
This is what windowing release is all about.
First presented in the CBS Laboratories “Audimax” level controller of the 1960s, the concept of windowing
release (which CBS called “Platforming” and is also described as ‘hysteresis release’) is an elegant way to
have your cake and eat it, too. The basic idea is that if the acceptable (and desirable) short-term variations of
the audio remain within a fixed range, the audio level is considered to be correct. The control system should
make NO adjustment until the audio moves away from the correct level. By determining a decibel range within
which any variation in the audio’s level is considered to be acceptable, we can set our equipment to then make
changes only when it leaves this range. This ‘do nothing’ range could be as small as a fraction of a decibel, or
as wide as we want to make it, though practical experiments have shown the useful range is from about one to
fifteen decibels.
Should the audio’s energy go outside either the top or bottom thresholds of our imaginary ‘window’, the control
system will briefly engage to correct the audio level, moving the window with it. Once our window of acceptable
variation is set at the newly appropriate level, the system once again disengages control and does nothing.
The result is that audio levels can be aggressively controlled, yet not sound it. All the detail, the instantaneous
variations, the ‘flavour’ remains, even though the audio is held at a consistency of level that the human ear typi-
cally finds to be very comfortable. The sound is extremely natural, mainly because our system spends most of
its time doing nothing. Multiple Band (‘multi-band’) AGC level controller that uses a single, wide-band system
will not alter the spectral balance of the controlled audio in any way. This is a good and admirable goal. But a
wide-band AGC system has a serious flaw: a highly-audible control side effect called spectral intermodulation.
When one instrument that is confined to a limited portion of the spectrum (such as a bass drum) is higher in
energy level than the rest of the spectrum combined, the level intensity and rhythmic variation of that instrument
will dominate the control and ‘modulate’ the AGC’s automatic level variation over the entire spectrum. When
the rhythmic characteristics of the lower-level instruments (such as strings) are sustained and held longer, the
modulation effect is clearly obvious. It is most annoying when the modulating instrument is at one extreme of
the spectrum where the ear is least sensitive while the modulated instruments are in the middle of the most ear-
sensitive area of midrange frequencies. Various solutions have been attempted this “hole-punching,” such as
slowing down the attack and release of the wide-band system, or desensitizing the low frequency sensitivity of
the control loop. Using a windowing release also helps. But these ‘solutions’ only slightly reduce the fault, they
do not fix it.
The solution is to control different portions of the spectrum with some degree of independence. In most cases
it can be shown that, for a given amount of dynamic gain change, by splitting the audio into several frequency
bands any control ‘artifacts’ are much less audible to the human ear than using a single band.
There is however, a potential problem with multi-band processing that can make it a deal-breaker: “spectral
skewing.” Spectral skewing is what happens when the operating characteristics of the multiple bands cause
them to drift way from each identical gains when differing material comes along at differing levels. The effect
can be deadly: a song can have a different spectral balance depending only upon how hard the processor is
driven. Song fades turn muffled or shrill, depending on processor settings. The solution to the spectral skewing
problem is to make each band’s detection and control systems react with the same characteristics, i.e., reaction
times (attack and release), compression ratios, thresholds, etc., and, importantly, set the crossover frequencies
such that there is a similar amount of energy in each band. This is the approach undertaken with the Ariane, and
spectral skewing very effectively becomes a non-issue. Any resultant spectral shift will in most cases be unno-
ticed and in some cases, considered an improvement.
Feedforward control
Historically, automatic gain controllers have used a detection scheme located at the output of the proces-
sor which was fed back to control the gain at the input. This is a self correcting system, called “feedback” or
“closed-loop.” Until accurate and stable linear-to-logarithm (and vice versa) converters became common, feed-
back processing was the only economical option. The problem with feedback processing is that the control
range is limited and non-linear from one end of its control range to the other. This makes feedback best suited
only for relatively small amounts of variable gain, i.e, less than 20 dB. This is OK for peak limiting, where the

Ariane
15
range of control is usually less than 15 dB, but for the purpose of a modern AGC this range is too ‘limited.’ A
“Feedforward” (or “open loop”) detector analyses the signal before the variable gain element and creates a con-
trol signal that can accurately mirror level variations in the incoming audio. With a feedforward system the effec-
tive range is many decades of decibels, and its linearity of control is virtually perfect from one extreme of gain
to the other. For an initial AGC for audio of an unknown level, a feedforward system makes great sense. In the
Ariane, the control range is well over 40 dB, and the control characteristics are the same at both extremes.
Introducing The Ariane
The Ariane is a:
• High-quality digital,
• Two-channel,
• Multi-band,
• RMS-derived,
• Windowing release,
• Feedforward control…
automatic levelling processor for general purpose audio level management.
Each of these features of the Ariane were chosen to solve a particular problem that has manifested itself in
previously available levelling solutions. Processors with one or more of the above features have long been avail-
able, but the Ariane is the first leveller to bring this unique and significant combination of features together in a
unified design.
The above waveform chart demonstrates the peak level of a piece of typical music audio, in blue, superimposed
with its RMS energy value, the black line. Note that the RMS level does not closely follow the peak level but,
rather, represents the short-term average density of the audio. Now let’s superimpose the Ariane control system
upon that same RMS waveform.
The horizontal red, dashed line represents the Gate threshold. Whenever the RMS energy goes below that level
(numbers 1 through 7, above), the release is gated, and the Ariane’s gain holds steady, with no change. When
this is true, the broad horizontal band is red (the vertical dimension of the multi colour band represents the
Ariane’s “window”).
When the audio’s RMS energy drops below the window, but still remains above the Gate threshold, the window
releases (numbers 8 through 13), and the window band is green, indicating an increase in gain.
Whenever the RMS energy stays inside within the range of the window, the band is Yellow, and the audio level
is unchanged, just as it was when it was gated.
In both charts, time goes from left to right, and higher amplitude is upward; in the lower chart, the broad multicol-
our band’s position vertically indicates more gain the lower it goes.
It’s easy to see that the Ariane can be adjusted to spend most of its time holding steady, with no level change.
This is the little bit of ‘magic’ that makes the Ariane sounds so good!

Source Quality And Pre-emphasis
16
SOUrCE MatErIaL QUaLIty
The DSPXtra-FM has the ability to substantially improve the quality of your ON-AIR broadcast. However the
DSPXtra-FM can only work with what you provide it. The best performance will be obtained when the DSPXtra-
FM is fed with very clean source material. After dynamic multi-band re-equalisation is performed poor quality
source material will sound poorer when processed with the DSPXtra-FM.
We strongly advise against the use of MP3's and other compressed audio formats for audio storage. If you must
use compressed audio we advise rates of 256 Kbps and higher. Linear formats are always to be preferred.
Compressed audio formats employ frequency masking data reduction techniques to reduce the bit-rate. Through
re-equalisation the DSPXtra-FM can violate the frequency masking characteristics of the bit reduction process,
creating distortion that was inaudible prior to the DSPXtra-FM processing.
PrE-EMPhaSIS
If you are using the DSPXtra-FM to process for FM broadcast you will need to enable the pre-emphasis filter in
the DSPXtra-FM. Even though your STL or transmitter may contain pre-emphasis we recommend disabling it,
letting the DSPXtra-FM handle the pre-emphasis for the transmission system. The DSPXtra-FM uses sophis-
ticated processing methods to limit the high frequency energy of the pre-emphasis curve while maintaining a
'bright' sound. Using de-emphasis and then pre-emphasising again will only degrade performance and possibly
casue overshoots, resulting in lower average deviation.
The exception to the rule is when the DSPXtra-FM is feeding discrete left and right outs to a compressed audio
STL. Bit rate reduction codec's do not cope with pre-emphasis very gracefully and artifacts will be generated.
The best option in this case is to de-emphasise the output of the DSPXtra-FM prior to the STL system. At the
transmitter site the pre-emphasis can be enabled in the transmitter to restore the processed signal back to nor-
mal prior to transmission.
The best solution is always to locate the processor at the transmission site. This way overshoots are minimised
and quality is maintained.

Processing Structure
17
thE dSPXtra-FM and ItS PrOCESSInG StrUCtUrE
The DSPXtra-FM broadcast audio processor can be used for processing audio prior to broadcast on FM and
digital radio services. Digital radio encompasses DAB, HD Radio (IBOC) and other radio based broadcasting as
well as internet radio, also known as streaming. The DSPXtra-FM can also be used effectively for audio post pro-
duction and mastering, ideal for giving CD’s that HOT sound. It is also possible to use the DSPXtra-FM for other
audio level control/equalization applications such as night clubs and bands. However, this manual will only be
referring to the use of the DSPXtra-FM for FM and digital radio processing.
Before we discuss the processing structure in full we would like to tell you a little about the final peak limiting
stages of the DSPXtra-FM. The DSPXtra-FM employs dual output paths for peak control. Your processing appli-
cation may need you to configure the DSPXtra-FM in a certain way. Selecting the wrong output path and or not
configuring the other settings that affect it may seriously downgrade your audio quality.
The first peak control path is known as 'FM' as it is typically used when processing signals for FM broadcast.
It employs distortion controlled clippers to limit the peaks of the signal. Distortion controlled clipping is the best
method for preserving as much high frequency energy as possible, important when the high frequency loss char-
acteristics of the FM broadcast de-emphasis curve is taking into account. Distortion controlled clipping produces
harmonic distortion which if used moderately can produce a sizzling bright sound but can result in a ripping or
tearing sound if used excessively (overdriven).
The second peak control path is known as 'DR' (digital radio) and is the desired method of peak processing
when the output feeds a codec that employs ‘bit rate reduction compression techniques’. The 'DR' path employs
look-ahead limiting as opposed to clipping. Look-ahead limiting produces less artifacts than conventional clip-
ping, so will reproduce the original audio more accurately with less bits of digital information because it is not
wasting bits encoding non-audible clipping artifacts. Look-ahead limiting produces less harmonic distortion but
produces IM distortion if over driven resulting in a packed, busy sound.
The DSPXtra-FM can be configured so that each peak control path can be routed to either of the digital or ana-
logue outputs. The stereo encoder is always fed with the 'FM' path. One popular configuration for FM radio sta-
tions is to use the DSPXtra-FM to process their FM broadcast and to have the 'DR' path feed their digital radio
service or web stream, each service optimally processed for that medium. We suggest that digital radio services
always use the 'DR' path but you are free to experiment with both options.
thE PrOCESSInG Path
Input selection and conditioning
The DSPXtra-FM offers the user input selection, gain control and a selection from a range of stereo/mono
options. The audio is then routed through defeatable high pass, phase rotating and pre-emphasis filters.
A silence detector provides automatic primary to secondary input failure switching.
Ariane RMS leveller.
For a transparent input levelling function, the DSPXtra-FM employs an RMS detected multi-band leveller
Ariane. With an award winning windowing gating concept, the Ariane improves consistency and presents a
uniform audio signal to the following stages. This important stage is described in more detail throughout this
manual. Ariane levller can also provide stereo enhancement through its matrix mode of operation.
Bass enhancement
The DSPXtra-FM offers three forms of bass enhancement.
1. A 12dB/Octave shelving filter with up to 12dB of gain.
2. Bass tune control.
3. A peaking filter that can be set to provide up to 6dB of gain on 1 of four frequencies with a choice of 4 Q's.
This can be thought of as a simple bass parametric.
Xover
The DSPXtra-FM employs linear-phase time aligned digital FIR filtering to split the audio spectrum into 6
bands while maintaining sonic transparency.
Multi-band Limiters
Each band has its own dynamic peak limiter. Multiple time constant based detectors with built in adjustable
hold and delay functions significantly reduce distortion.
Mixer
The six bands are 'virtually' mixed together at this stage. In truth, the six bands have become three. The

Processing Structure
18
three bands are fed off into the two peak processing paths.
Distortion controlled clippers (PEAK CONTROL PATH 1)
The DSPXtra-FM clipping algorithms peak limit (clip) and linear phase filter the audio in three bands for maxi-
mum distortion control before being fed to the final clipper stages.
Look-ahead limiter (PEAK CONTROL PATH 2)
The DSPXtra-FM look-ahead limiter is one of the most sophisticated available in a broadcast processor.
Processing is performed in three bands for maximum transparency and clarity. A cut shelving filter is provided
to compensate for the effect of pre-emphasis when the DSPXtra-FM is used to process FM signals at the
same time.
Output selection, processing and routing
The DSPXtra-FM allows the user to select where each processing path is routed to and provides output level
controls. A de-emphasis option is provided on both the analogue and digital outputs. The digital output sam-
ple rate can also be configured to a variety of settings.
The analogue output also has one extra path that can be routed to its outputs. This is a lower latency (delay)
path that bypasses the final clipper stages and reduces delay by more than 4mS. This can be used to help
with talent (DJ's) who can not get to grips with the delay of the main processing path. This output if selected
should only be used as a studio monitor signal and should not be used ‘on the air’ as the peak clipping stage
in the monitor processing path is not over-sampled or anti-aliased.
The analog and MPX outputs will be affected by the ITU limiter, if it is engaged. The purpose of this limiter is
to comply with the ITU BS.412 standard. The standard calls for limiting of the power of the composite signal
being broadcast. When activated, the ITU limiter will drastically reduce the loudness of your signal.
Stereo encoder
The DSPXtra-FM's DSP stereo encoder takes its inputs from the FM path of peak processing. The audio fed
to the stereo encoder does not pass through any de-emphasis circuits. The stereo encoder is highly over-
sampled and offers superb stereo performance. A composite clipping function is provided for those who wish
to use it, as well as a pilot protection filter.

Block Diagram
19
ANALOG
DIGITAL
A/D
SRC
HPF
INPUT
LPF
BPF BPF HPF
P.R.
BASS
SHELF
BASS PEAK
PRE
LOW
LIM
LMID
LIM
HMID
LIM
MB LIM
MID
LIM
LPF
BASS CLIP
BPF
MID CLIP
HPF
HF CLIP
N-1
DELAY
ATTACK RELEASE
LOOKAHEAD
SOFT
HARD
N-1
DELAY
ATTACK RELEASE
LOOKAHEAD (DR MODE)
N-1
DELAY
ATTACK RELEASE
N-1
DELAY
ATTACK RELEASE
LPF
15
19
OUTPUT MPX
PILOT OUT
D/A
MPX
COMP
CLIP
DIGITAL
SRC
DE
ANALOG
D/A
DE
PILOT
SYNC
SCA IN
MPX OUT
DSPXtra-FM PROCESSING BLOCK DIAGRAM
MON
DR
DR
FM
FM
LOW
AGC
MID1
AGC
HIGH
AGC
ARIANE LEVELER
LPF
BPF BPF HPF
MID2
AGC
SHIGH
LIM
HIGH
LIM
BPF BPF
MATRIX
/
STEREO
selection
MATRIX
/
STEREO
selection
LPF
FAT
SLIM
LPF
MED
ITU
LIMIT

User Levels
20
USEr LEvELS
The DSPXtra-FM has the ability to restrict the use of some of the more advanced processing parameters by
selecting a beginner mode. This mode control is located at the top level of the processing menu tree, the options
being beginner and advanced.
Selecting advanced will allow access to every menu and processing parameter while beginner ‘hides’ certain
processing controls and menus. The menu structure diagram on the next page illustrates what is hidden when in
beginner mode.
It is worth noting that when in beginner mode the ‘HIDDEN’ processing parameters are still active, but defaulted
to a safe value that works well with the processing parameters that are left on and selectable. Another reason for
doing this is to stop the scenario of setting an advanced control to a certain value and then switching to begin-
ner mode and having no way of fully controlling the processing because the ‘HIDDEN’ parameter is left in a state
that makes the audio sound bad. One downside of this is that if you make a preset in advanced mode and then
switch to beginner mode the sound will change as the hidden parameters default to safe values. To avoid issues
like this we suggest starting with beginner mode and then switching to advanced mode to refine the preset if
necessary. We see no reason why you would want to do this in reverse but we thought it best to mention it so
that when switching to beginner mode you don't wonder why you have just lost your preset that you just spent
an hour working on. One good reason for using the save preset facility at various stages when making a new
preset!