Professional Documents
Culture Documents
AudioXpress 2014 02
AudioXpress 2014 02
AudioXpress 2014 02
www.audioxpress.com
audio xpress
www.audioxpress.com
Standards Review
Audinate Dante
Making Digital Audio Networking Easy
You Can DIY!
Vacuum Tube Oscillator
By Atto Rinaldo
Audio Electronics
A High-Voltage Delay
for Tube Amplifiers
By Jan Didden
Sound Control
Acoustical Diffusion and Scattering
Audio Praxis
The Brave New World
of Loudness Control
FEBRUARY 2014 By Jon Schorah
The Authority on Hi-Fi DIY
Your #1 Source
for
NEW & NOS Vacuum Tubes,
DIY Parts & Components
and
Audiophile Accessories.
AZUMA
1-866-681-9602
Debit, Visa, Mastercard, Amex, PayPal, Money Order & Bank Draft
Contents
Features
12 Solid State Logic’s Sigma lost in the old paradigm of peak
Innovative Mix Engine normalization.
By Miguel Marques
This review details the features 44 The Ultimate Living Room
of Solid State Logic’s stand- Home Theater
alone summing mixer, a digitally By Ethan Winer
controlled device that can provide Good acoustics affect the sound
automation to any console. quality more than the choice of
receiver or loudspeaker.
14 High-Definition Digital
By Gary Galo 48 Vacuum Tube
“Move Over PCM and Make Room Low-Frequency Oscillator
for HiRes DSD and DXD Audio,” By Atto Rinaldo
a presentation at the 135th Design and build a vacuum
Audio Engineering Society (AES) tube oscillator utilizing current
Convention. components.
Columns
Departments
STANDARDS REVIEW
18 Audinate Dante (Part 1) QUESTIONS & ANSWERS
Making Digital Audio Networking 40 The Original 6 From the Editor’s Desk
Easy Sound Designer
7 Client Index
By João Martins An Interview with Dan Dugan—
Audio Engineer, Inventor, and
8 What’s News
SPEAKERS Nature Sounds Recordist
32 Super Subwoofers By Shannon Becker
64 Member Profile
A Battle of the Titans
By Mike Klasco and Steve Tatarunis HOLLOW-STATE ELECTRONICS 66 Industry Calendar
58 The Tone Character of
SOUND CONTROL Tube Guitar Amplifiers
36 Acoustical Diffusion and By Richard Honeycutt
Scattering Websites
By Richard Honeycutt
audioxpress.com
voicecoilmagazine.com
cc-webshop.com
@audioxp_editor audioxpresscommunity
© Segment LLC 2014 Regular Contributors: Bill Christie, Dennis Colin, Joseph D’Appolito, Vance Dickason, Jan
Printed in the US Didden, Gary Galo, Chuck Hansen, Richard Honeycutt, Charlie
Hughes, Mike Klasco, G. R. Koonce, Ward Maas, Miguel Marques,
Nelson Pass, Bill Reeve, Steve Tatarunis, David J. Weinberg
United States
Carlo van Nistelrooy
860-289-0800
c.vannistelrooy@elektor.com
Elektor Labs
Wisse Hettinga SUPPORTING COMPANIES
+31 46 4389428
w.hettinga@elektor.com
ACO Pacific, Inc. 38 Home Theater Shack 64
COLUMNISTS
Brazil
João Martins
+31 46 4389444
j.martins@elektor.com
Vance Dickason has been working as a professional in the loudspeaker industry since 1974. He is
the author of Loudspeaker Design Cookbook—which is now in its seventh edition and published in
Portugal English, French, German, Dutch, Italian, Spanish, and Portuguese—and The Loudspeaker Recipes.
João Martins
+31 46 438944 Vance is the editor of Voice Coil: The Periodical for the Loudspeaker Industry, a monthly publication.
j.martins@elektor.com Although he has been involved with publishing throughout his career, he still works as an engineering
consultant for a number of loudspeaker manufacturers.
India
Sunil D. Malekar Dr. Richard Honeycutt fell in love with acoustics when his father brought home a copy of Leo Beranek’s
+91 9833168815
ts@elektor.in landmark text on the subject while Richard was in the ninth grade. Richard is a member of the North
Carolina chapter of the Acoustical Society of America. Richard has his own business involving musical
instruments and sound systems. He has been an active acoustics consultant since he received his PhD in
Russia
Nataliya Melnikova electroacoustics from the Union Institute in 2004. Richard’s work includes architectural acoustics, sound
+7 965 395 33 36
system design, and community noise analysis.
Elektor.Russia@gmail.com
Mike Klasco is the president of Menlo Scientific, a consulting firm for the loudspeaker industry, located
Turkey
in Richmond, CA. He is the organizer of the Loudspeaker University seminars for speaker engineers. Mike
Zeynep Köksal
+90 532 277 48 26 specializes in materials and fabrication techniques to enhance speaker performance.
zkoksal@beti.com.tr
Steve Tatarunis has been active in the loudspeaker industry since the late 1970s. His areas of
China interest include product development and test engineering. He is currently a support engineer at
Cees Baay
+86 21 6445 2811
Listen, in Boston, MA, where he provides front-line technical support to the SoundCheck test system’s
CeesBaay@gmail.com global user base.
The APx525 analyzer can be combined with the new Loudspeaker Production Test suite Audio Precision
for work on integrated devices such as powered speakers, TVs, and Bluetooth headsets. www.ap.com
For work involving driver evaluation or final production test, a simpler APx515 can be
used.
Audio Precision
www.ap.com
I/O meters mostly occupy the Sigma front panel. A knob level can control monitoring and talkback functions, and two user-defined buttons can
be assigned to functions in the control panel. The Sigma also includes an external audio input using mini TRS input and has a standard 0.25” TRS
headphone output.
High-Definition Digital
Direct-Stream Digital (DSD) recording/playback technology takes a
different approach to existing pulse-code modulation (PCM)-based
systems and delivers the high-resolution stereo or multichannel
By
Gary Galo audio found on the Super Audio CD (SACD). DSD encodes audio
(United States) data using 1-bit samples taken at 2,822,400 samples per second.
This is 64 times faster than the rate used on standard audio CDs,
which enables the digital representation on SACD audio to more
closely follow the analog source signal that is being encoded.
DSD technology appears to be gaining in popularity among
audio industry professionals.
In October 2013, the Audio Engineering Society flexibility and ease of use and far greater signal pro-
(AES) 135th Convention kicked off AES week in New cessing capability, than raw DSD editing.
York with a 2-h monthly meeting focused on “Move Digit a l e X t r e m e D e f initio n ( DX D) is a n
Over PCM and Make Room for Hi Res DSD and DXD ultra-high-resolution recording format developed
Audio.” Three of the world’s leading proponents of by Merging Technologies, which offers the Pyramix
the new ultra-high-resolution recording formats— multitrack recording system. DXD operates at sam-
Dominique Brulhart of Merging Technologies, Morten pling frequencies of either 352.8 kHz (8 × 44.1) or
Lindberg of 2L, and John Newton of Soundmirror— 384 kHz (8 × 48), up to 32-bit floating. It also sup-
lead the discussion. ports all the standard sampling frequencies from
Sony developed the DSD recording format for the 44.1 to 192 kHz. DXD is a PCM format (the title of
SACD. A 1-bit system, DSD originally operated at a the presentation was a bit misleading in that regard)
2.8-MHz sampling frequency, providing a 100-kHz and is not new.
bandwidth. Conventional 2.8-MHz DSD recording is Merging Technologies exhibited an earlier version
often referred to as DSD64 because the sampling rate of the Pyramix system at the 123rd AES Convention
is 64 times that of the 44.1-kHz CD standard. (The in 2007. One of DXD’s virtues is its conversion tech-
“2.8 MHz” designation has been rounded off.) But in nology to and from DSD is incredibly sophisticated
recent years, the base-sampling frequency has been (352.8 kHz is the preferred sampling frequency for
doubled to 5.6 MHz and even quadrupled to 11.2 MHz. optimum conversion to DSD). Because DXD is a PCM
These new DSD formats are often referred to as format, the Pyramix recording system offers the ease
DSD128 and DSD256 (i.e., Double DSD and Quad DSD). and flexibility recording engineers expect from PCM.
Although the SACD has never become much more than However, because the conversion to DSD is so trans-
a fringe audiophile format, the DSD recording system parent, DXD makes an excellent recording and mas-
has a small but loyal following in the audio industry. tering system for a finished product in DSD format,
One of the main problems with DSD is the difficulty and whether an SACD or a raw DSD file.
expense of editing. PCM editing offers much greater Brulhart gave an overview of the formats and
described the new ultra-high sampling frequencies
used for DSD. Indeed, the “merging” in Merging
Technologies focuses on bridging the gap between
the two digital formats, and the panelists noted that
the conversion algorithms will further improve in
The configuration screen
the future.
of Merging Technologies’
Horus network interface
supports recordings from Demos and Testimonies
44.1 to 192 kHz, DXD and Lindberg is the proprietor of 2L, a small Norwe-
up to DSD256 modes. gian company specializing in high-resolution digital
audio. 2L has also issued recordings in high-resolu- only those spots—often only 10 ms in length—to DXD
tion PCM formats on Pure Audio Blu-ray discs and on to perform the edit, and then convert back to DSD. The Horus Networked
SACDs. About 50% of its sales are high-resolution Newton noted that Merging Technologies’ recording Audio Interface from
downloads. equipment follows the Ravenna protocol described Merging Technologies
Lindberg noted that attendees who were hoping in my 2008 article about the AES 2007 Convention quickly became one of the
for a boxing match between DSD and DXD would (see Resources). The technology uses CAT-5 cabling most popular AD/DA and
microphone pre-amplifiers
be disappointed. Both formats are excellent, in his from the performance space to the control room.
solutions used for high-
view, providing a level of transparency that enables This minimizes the amount of microphone cabling
resolution audio recordings.
the recording engineer to forget about the recording he transports for a location recording, which results In the picture, Merging
system and concentrate on microphone placement. in a substantial cost savings. He normally buys the Technologies’ President,
Lindberg prefers Brüel & Kjær 130-V omnidirectional CAT-5 cable locally and leaves it on site—it’s less Claude Cellier is supervising
condenser microphones and makes nearly all of 2L’s expensive than shipping it back and forth and it can a recording session where
recordings in a five-channel surround format, gen- always be reused for future sessions at the same the Horus is used as an I/O
erally using only one microphone per channel. Spot location. Newton said the differences between DSD connection node to RAVENNA
microphones are not used in most of their record- IP Audio Networks.
ings, and equalization is only needed on about 10%
of them. Lindberg records and masters in the DXD
format and does not go back and forth between DXD
and DSD. When the master is finished, he makes a
one-time conversion to DSD.
Newton has been the proprietor of Soundmirror
since 1971. Soundmirror is a Boston, MA-based com-
pany that specializes in high-resolution digital record-
ing. It was responsible for the remastering of the
RCA Victor Living Stereo material issued by BMG on
SACD. Newton prefers recording directly in DSD for-
mat. But, he admitted that however spectacular the
results, using DSD for recording, mixing, and edit-
ing is very expensive and time consuming. In cases
where cross-fade edits are needed, he will convert
“I asked Merging berg played a military band recording that was very
[Technologies] to pay impressive in terms of detail and depth perspective.
attention to the headphone The stereo layer from this SACD was created from
amplifier when Horus was
his five-channel pickup. Newton played a Haydn Sym-
being developed. These
phony performed by the Oregon Symphony Orchestra,
often sound very poor on
other devices. I must say recorded in DSD at 11.2 MHz using a pair of Brüel &
they definitely listened Kjær 2006 microphones for the stereo pickup. The
to me because it is really playback loudspeakers were a pair of NHT Pro Model
outstanding. It also saves A-20 Active Monitors, a powered two-way loudspeaker
me from taking a separate system with a 6.5” treated paper cone woofer and a
amp and all the cables butyl rubber surround, and a 1” metal dome tweeter
and power supply,” says with a textile surround. Amplifier power was rated at
Jean-Daniel Noir from
at 2.8, 5.6, and 11.2 MHz is audible even in simple 250-W continuous RMS, and the system also included
the specialist recording
recording/playback demonstrations, and not just in its own outboard control amplifier.
company, JDN—based in
Gland-Vich, Switzerland. editing and processing. The meeting venue was a fairly intimate space.
An assortment of high-resolution digital record- The loudspeakers had no difficulty supplying clean,
ings was played to demonstrate the working methods detailed sound, though the bass, as expected from
and recording philosophies of the participants. Lind- such a small woofer, was deficient. Unfortunately, the
At the October 2013 New York Chapter of the Association for Recorded Sound
Collections (ARSC) meeting, producer/engineer Jon M. Samuels and project
consultant Joseph Patrych gave a presentation on Sony Classical’s 41-CD collection
Vladimir Horowitz: Live at Carnegie Hall. (Photo courtesy of Sony Classical)
Resources
2L, www.2l.no.
Merrill Replica
ES-R1
Turntable manufactured in the USA by George Merrill
Audinate Dante
(Part 1)
In a two-part article series, we will explore the Dante can be found throughout the pro audio live
reasons why Dante technology became the “de facto” sound market. But most importantly, Dante is also
standard in audio networking and reveal the com- reaching the commercial installation, broadcast,
pany’s perspective of “surfing the wave” of audio and recording studio segments.
over IP (AoIP). We will also discuss the company’s The Dante networking solution became widely
ability to cleverly manage existing industry efforts accepted among many pro audio manufacturers and
and commercial requirements to build a unique is currently deployed in thousands of installations
example of marketing success. As Audinate’s CEO worldwide because of its self-configuring network
Lee Ellison explains, because “the great thing about architecture. Dante uses standard IP over 100-Mbps
standards is that there are so many of them.” and 1-Gb Ethernet, enabling easy setup and auto-
Built on existing networking protocols and stan- matic discovery of devices on the network with
dards, Audinate’s Dante technology is a self-de- one-click signal routing and user-editable names.
scribed “plug-and-play networking solution, which In contrast to previously existing audio network
delivers synchronized media with ultra-low latency, technologies, Dante distributes digital audio plus inte-
simplifying the installation and configuration of digi- grated control data with imperceptible latency, sam-
tal media networks.” Dante is currently the market’s ple-accurate playback synchronization, and high-chan-
leading digital audio networking solution, adopted nel counts. The technology reliably uses standard
Sydney, Australia-based
Audinate became a global by approximately 150 OEMs to date (some still to network infrastructures, even with high-quality digital
company in just 10 years be publicly disclosed). Those OEMs have already signals in high-sampling frequencies.
thanks to the success of its developed a unique ecosystem of hundreds of com- As licensed technology, Dante also offers a
Dante technology. patible products. Currently, applications that use combined hardware and software toolset, enabling
A Little History
In the analog domain, the main concern dealt
with noise interferences and audio signal degrada-
tion from long cables. With the adoption of com-
puter networks, the focus evolved from the simple
transmission of digital audio signals in point-to-
point, to finding the best way to efficiently distribute
and route audio without compromises in quality—if
only using, in practice, the Ethernet physical layers.
From standards such as AES/EBU (currently
AES3), a multichannel version was created and
Multichannel Audio Digital Interface (MADI) AES10
Dante Controller is a free
software application for a broad range of technology sectors including soft- became the first solution to enable large routing sys-
Windows (7 and 8) and Mac ware, computer, and telecommunications. His expe- tems and the distribution of up to 56 channels (later
OS X. It can be used to set rience enabled him to quickly become an industry expanded to 64 in the 2003 revision). The hardware
up and manage audio routes spokesman on all media network-related subjects. was expensive because it was based on coaxial
in a Dante audio network As Ellison is quick to instantiate when we refer cable, but the possibilities still generated interest,
and configure Dante devices, to Dante’s success as a “de facto” standard, “we especially for use with broadcast applications.
providing real-time network are becoming an enabling technology, rather than a Other digital point-to-point audio transmission
monitoring functionalities.
competing technology. We are enabling things that interfaces, such as TDIF and ADAT, survived for
could have never been done before… quite some time together with AES3 and MADI, but
“Maybe the reason is that, from the beginning, they required dedicated cables. Digital signals were
we architected Dante looking at using the “transmitted” or distributed point-to-point in most
IP technology as a foundation, unlike other of the applications. But it become apparent that cre-
companies who approached this industry in ating a true network topology, in which the signals
the past. Primarily, those were audio compa- can be routed freely to any number of points with
nies developing another network technology, bidirectional flows of independent data sharing the
while we are not an audio company, we are same infrastructure, made sense. The flexibility this
a networking company. We have also the would enable encouraged many R&D departments to
largest engineering team out there with a start examining solutions that would use standard
background on all various forms of network- Ethernet network switchers, routers, and cabling.
ing, so that’s why we where able to do that. Audio Engineering Society (AES) convention
“We knew there would be some additional exhibitors have demonstrated such systems since
challenges and more stringent requirements 2000. Among them was Sony’s demonstration of the
in transmitting uncompressed audio. From Pro-Audio Lab Oxford (2003) of SuperMac (10/100 Mb)
the beginning our initial focus was really and HyperMac (1 Gb), which were later recognized
the Live Sound. Commercial installation is as the AES50 standard (2005). The technology was
a larger market than Live, but we felt we later sold by Sony Oxford to Klark-Teknik. Today it
could address the more stringent require- serves as the foundation for the successful range
ments of live sound and we knew we had an of Midas and Behringer (The Music Group) digital
Dante uses automatic
architecture from the perspective of being a scalable consoles and live sound solutions.
device discovery and Zero
platform, because we were using IP.” Interestingly enough, the technology was intro-
Configuration Networking—
Internet Engineering Task But this is not the entire story. Essentially, Audi- duced as a “multichannel audio interconnection”
Force (IETF) Zeroconf nate responded efficiently to a diversity of require- solution running on CAT-5/6 concurrently with con-
protocols. ments and different goals from the manufacturers trol data and (potentially) enabling independent
Resources
Audio Engineering Society (AES), “AES standard for Audio Appli- S. Schmitt and J. Cronemeyer, “Audio Over Ethernet: There are Many
cations of Networks—High-Performance Streaming Audio-over-IP Solutions—But Which One Is Best for You?” www.dspecialists.com/
(AoIP) Interoperability,” AES67-2013. sites/default/files/publication/110126a_networking_paperembedded
world2011_paperen_final_st_jc.pdf.
———, “Analog to Digital Audio in the 21st Century,” White Paper,
www.audinate.com. T. Shuttleworth, “Emerging Technology Trends Report,” Audio Engi-
neering Society (AES) Technical Committee on Network Audio Systems,
———, “Audio Networks Past, Present and Future (CobraNet and November 2011.
Dante), White Paper, www.audinate.com.
Sony-Oxford Technologies, “Oxford SuperMac/HyperMac AES50 Digital
———, “Dante: Digital Audio Networking Just Got Easy 2.0US-09A09. Audio Interconnection,” technology brief, 2005, www.philippe-lahaye.fr/
pdf, White Paper, www.audinate.com. IMG/pdf/xmac_brochure.pdf.
J. Berryman, “OCA Alliance Open Control Architecture: The OCA Alli- M. Johas Teener, A. Huotari, Y. Kim, R. Kreifeldt, and K. Stanton,
ance Overview,” www.oca-alliance.com.Audinate, “Evolving Networks “No Excuses Audio/Video Networking: The Technology Behind AVnu,”
to Audio Video Bridging (AVB)” White Paper, 2011, www.audinate.com. AVnu Alliance White Paper, 2009.
A. Hildebrandt, “Networked Audio: Aktuelle Entwicklungen & Tech- M. Vest, “High-Quality, Low-Latency IP-Based Audio Routing via
nologische Perspektiven für den Broadcast-Markt (Networked Audio: Ethernet,” NTP Technology Presentation, 2013, www.ntp.dk/media/
Current Developments & Perspectives for the Broadcast Market), VDT NTP_presentation_IP_Audio_2013.pdf.
International Convention, November 2010.
As loudness management becomes the required correctly integrated, it can become more than a
norm, the responsibility for compliance increasingly means of avoiding consumer complaints and poten-
falls on audio post-production engineers. But there’s tial fines. It can reintroduce creative freedoms lost
some good news. Since several world regions have in the old peak normalization paradigm and become
been delivering loudness-compliant broadcasts for a tool to improve broadcast audio quality.
two years or more, a methodology for best practice
is beginning to emerge. The solutions for loudness Loudness and Post-Production:
control are not only maturing, but also becoming A Creative Marriage
versatile tools in the engineer’s toolbox. At first glance, playout processing is an obvi-
Initially, loudness compliance may appear ous solution to loudness compliance. By definition,
to be just another task added to the seemingly playout processing achieves loudness normalization
ever-growing list of tickboxes engineers must check as an afterthought by adjusting station output to
before delivery. But if loudness normalization is ensure compliance after post-production. However,
experience has shown that playout processors are the mixing process, the right loudness tools provide Figure 2: A section of
not particularly good at solving loudness issues in a an engineer more creative freedom. Compressing heavily compressed audio
consumer-satisfying manner. They can even intro- a mix to achieve a consistently loud level under has been normalized to –23
duce their own new loudness problems. In regions loudness normalization will cause the audio to be LUFS (a). The same section
where loudness recommendations have been in turned down. So the best way to achieve a mix of audio is shown with light
compression normalized to
place for more than a year, there is a clear move- that stands out in the crowd is to creatively engage
–23 LUFS (b). A comparison
ment away from correction after the fact. with the content and make the most of the avail-
clearly shows that instead
A better solution would be to consider loudness able dynamics. of producing a “louder”
compliance as part of the creative process during Loudness normalization coupled with the use result, much of the heavily
pre-production and even at acquisition. In this sce- of true-peak maximum levels enables you to cre- compressed signal is below
nario, simply pushing the mix against the limiter is atively use dynamic range and increase headroom. that of the corresponding
no longer a viable option. Instead, incorporate loud- When experienced operators ensure compliance in lightly compressed
ness compliance into an experienced post-production post-production, the playout processor becomes alternative after loudness
professional’s creative decision-making process. In largely inactive. These are great developments, normalization.
general, noncompliant material that falls outside a especially for professionals involved in short-form
tolerance margin is rejected. Near-compliant audio production.
can be satisfactorily corrected with loudness batch Loudness can be considered in pre-production
processing that brings it into compliance with a gain and acquisition. Normalizing archive and library
offset and possibly some true-peak limiting. Here, the material and ensuring field recordings and outside
playout processor’s role shifts to that of an error-han- broadcasting (OB) sources are already compliant
dling stop-gap and is bypassed with the delivery of speeds the production process, enabling a faster
compliant material. turnaround while ensuring the audio remains faith-
Instead of adding another complicating factor to ful to its original context. This can be especially
S Loudness
I Loudness
M Loudness minimum
M Loudness maximum
Variance maximum
Loudness range
True peak maximum
True peak clip
Alert
Figure 3: The typical log
file output from a real-time
loudness meter (VisLM-H)
shows the variation of
loudness parameters over
time.
important during live sports events, breaking news even experts need to confirm that they’ve met target
broadcasts, event coverage, and studio interviews. values. Likewise, anyone looking to push creative
boundaries needs checks to ensure their work is
Building a Post-Production Loudness compliant. With visual meters, editors can keep an
Workflow eye on the meter and loudness profile while relying
New workflows call for new tools. Fortunately, on their trained ears to make most of their decisions.
companies are developing products designed spe- Another important factor in this workflow is a
cifically for post-production engineers. Intuitive high-quality true-peak limiter that can handle the
audio-editing tools (e.g., real-time metering, offline new standard’s intersample true-peak requirement,
correction, and loudness-compliant limiting) enable which is something traditional sample-peak limiters
post-production editors to put their creative exper- cannot do. It’s tempting for engineers to use their
tise to work while ensuring compliance. existing sample peak limiters with a setting that
Ears are the best tools when it comes to making would yield results “safe enough” to be compliant
creative decisions in audio post-production. That with the loudness standards’ true-peak measure.
rule also generally applies to loudness normalization. But, those who follow this practice do so at their
The new loudness standards hold the potential for peril. Simply put, it’s impossible to arrive at an
increased dynamic range and contrast. But com- accurate true-peak reading with a sample peak
puters also play a key role with their ability to take limiter because the measurements are different.
measurements and make smaller adjustments to What may seem like safe settings on a sample peak
get things exactly on target. Computers can work limiter would not guarantee compliance.
quickly, saving considerable time near the end of Therefore, the best true-peak limiters offer a
the process. With these tools in mind, a post-pro- true brick-wall solution, measuring inter-sample
duction workflow begins to emerge. peaks and enabling the user to define the audio
Clear, intuitive loudness metering is the key to output’s true-peak limit (rather than the more tra-
delivering high-quality, loudness-compliant audio. ditional threshold control at which limiting begins to
Because the new loudness measurements are take effect). Based on ITU-R BS.1770’s standardized
designed to correspond to the human ear, a good true-peak algorithms, these tools are suitable for
engineer can almost mix sound without a meter. It’s controlling audio for post-production and broadcast
possible for skilled engineers in a calibrated room applications. True-peak limiting can also be used to
to simply occasionally glance at the meter during ensure that downstream codecs (e.g., MP3, AAC, and
the creative process to maintain their bearings or others) do not introduce distortion into the signal.
to check something in particular. But in the end, Once a mix is more or less loudness compliant,
Figure 5: Production to
loudness standards are
shown within a typical
nonlinear editor (NLE).
editors can use offline tools to fine tune the mix acquisition is a significant timesaving pre-production
and speed up the last part of the normalization technique that brings audio into the editing suite
process. These timesaving tools can be plugged at the right time. Another area in which loudness
into the editing environment to bring a mix into can play an important role is dialog clarity. Even
line quickly, correcting any true-peak overshoots today, mixes are occasionally broadcast with the
along the way. background music too loud, which makes dialog
Batch analysis is another highly useful tool for indistinct and results in viewer complaints. Using
busy post-production operations, enabling them to a meter to preserve loudness separation for dia-
automate part of their loudness processing. Acting log above other mix components can help guard
as a rapid fail-safe system and internal QA compo- against these mistakes. Measuring the loudness of
nent, a batch processor can automatically assess background music beds and FX spots can also help
files for compliance and correct or reject as needed. maintain consistency from section to section. In
sessions that require significant complex editing,
An Expanded Role for Loudness loudness normalization can quickly match dialog
Loudness measurement doesn’t have to end levels to a far more useful control than 0 dBFS (i.e.,
when broadcast criteria are met. In regions where the maximum possible digital level).
loudness compliance is an accepted part of the Audio libraries can also benefit from loudness
audio workflow, the same normalization tools can normalization, ensuring that audio is always inter-
also be employed in several areas that go beyond nally consistent and available at an expected level.
technical compliance to support new and improved Consideration of the loudness range (LRA) param-
Figure 7: A sample-by-
production techniques. eter can also be useful when mixing material for a
sample analysis shows a
True-peak over. As I mentioned, loudness consideration during specific target device.
By Although there are many transducer technologies trade shows have also displayed some “erector-set”
Mike Klasco and for sound reproduction, generating a lot of deep bass frames for one-off concept speakers.
Steve Tatarunis requires several woofers, a huge horn, or one really big For these large woofers, a 4” diameter voice
woofer. Of course, high excursion and a lot of power coil is barely enough, and the tooling is limited for
(United States)
handling and sensitivity does not hurt. And while we voice coils beyond 4”. Large-diameter voice coils
don’t want to disparage some of these efforts, several are not just needed for thermal power handling. If
of these designs are flawed and never saw much com- the voice coil diameter remains constant and the
mercial success. The failures were due to size, and in cone’s diameter increases, then the cone’s unsup-
some cases, questionable engineering. Yet, these giant ported area increases. This puts increased strain
subwoofers are entertaining to revisit and a couple of on the cone’s strength.
them do contain inspired engineering. For example, take a large powerful speaker
While good engineering does not always scale, motor and a tight box then add a lot of excursion.
loudspeaker physics may. No matter what the dia- As the autosound subwoofer aftermarket discovered,
phragm’s diameter in these giant woofers, they all it results in a cone-crunching machine.
require cabinet volume to maintain efficiency at Paper cones rule the woofer world, at least for
very low frequencies. Speaker manufacturers that the majority of units. For these subwoofer giants,
make 15” and 18” woofers say the sales ratio is many of the cone processing machines have size and
approximately 10:1 in favor of the smaller woofer material limitations. Fostex circumvented this chal-
because larger piston areas need larger internal lenge by using pie slices pieced together on its 31”
box volumes. Anything larger than an 18” woofer paper cone. Paper can be strong, especially with the
becomes abnormally sized. right additives such as aramids (Kevlar), carbon fiber,
or hemp. Some woofers use woven and non-woven
How Do the Big Woofers Differ? high-performance composites. Others have metal
There are no standards for speaker frames or cones. Electro-Voice even had some success using
cones beyond 18”; however, 21” is the logical next Styrofoam on its 30W, a 30” woofer. And, recently
size increase. Frame fabrication and construction developed high-performance ultra-high density poly-
is limited by projected sales vs. tooling costs, so ethylene materials (e.g., Teijin Endumax) promise
aluminum gravity-cast or sand-cast frames are significantly enhanced strength at lighter weights.
more common than aluminum, zinc, or magne- Large-size speakers do not always have high
sium alloy injection castings. Stamped steel tooling excursion, but when they do there are some seri-
is expensive and the likelihood of a heavy magnetic ous design issues with the coil centering spiders.
structure warping the thinner gauge steel frames The large cones with big-diameter voice coils usually
requires some sort of metal casting. However, some result in high-moving mass, which means the spider
History
Historically, large speakers appeared relatively
early because amplifiers were limited in power,
which made loudspeaker efficiency crucial. One
way to provide higher bass efficiency was to use a
large radiating area (i.e., a large cone).
As a young man, Rudy Bozak worked at a com-
pany called Cinaudagraph. While there, he helped
prepare a tower topped with a cluster of eight 27”
Cinaudagraph model PE-27 loudspeakers in 30”
frames with large 450-lb field coil magnets for the
1939 New York World’s Fair.
This package provided the low-frequency signals
for a two-way sound system at Flushing Meadows in
Queens, NY. An ad for the PE-27 loudspeaker claimed
a 20-to-10,000-Hz response and a 600-lb shipping
weight! The loudspeakers were mounted into bass
horns with 14’ wide mouths and were each driven
by a 500-W amplifier derived from a high-power Photo 1: Hartley Products
radio broadcast tube. Bozak went on to found Bozak continues to produce its 24”
Loudspeakers in Stamford, CT, which was famous woofer. (Photo courtesy of
for a strong bass signature sound. Hartley Products)
Hartley Products—founded in England in 1927 and
still operating today in North Carolina—was another
early large woofer contributor. The company began
developing a unique polymer cone for its own 21”
in 1956. The suspension uses a silicon rubber and
the spider shape was the old style “45-RPM” type
made from trilaminate fiberglass.
Hartley Products’s magnetic damping suspension
transformed the driver’s voice coil into a part-time
electromagnet. In the presence of an AC signal, the
coil and cone could move freely. In the absence of
a signal, a strong restorative force pulled the coil
back to its starting position. In doing so, it created
an electrodynamic damping effect. The Hartley 220
MSG is built on a sand-cast aluminum frame with
a 6-lb ferrite magnet (see Photo 1).
Electro-Voice first introduced its 30W with a 30” Photo 2: This is an early
Styrofoam cone about 50 years ago (see Photo 2). ad for Electro-Voice’s 30W
The loudspeakers matching enclosures were the loudspeaker. (Photo courtesy
corner-loaded Patrician 700 and 800. The enclosures of Electro-Voice)
There will always be reflections, even in a imply the reflected sound’s time-of-flight is about
so-called “anechoic chamber” (although the reflec- 24’ longer than the direct path.
tions in a chamber will be very low in amplitude,
especially at high frequencies). The time gap between The Haas Effect
the onset of the initial sound and the first reflection In 1949, Helmut Haas published research indi-
is called the “initial time-delay gap.” This gap cor- cating that reflections occurring within 50 to 80 ms
responds to the distance from the sound source to of each other are merged by the ear-brain system
the nearest reflecting surface, and from there to the into a single acoustical event. This phenomenon is
listener’s ears, compared to the sound’s direct path. called the “Haas Effect.”
Figure 1 shows the early part of a decaying sound Actually, the critical time separating sound into
wave with the gap marked. The brain concludes that multiple acoustical events depends on the type of
the acoustical space is large if the initial time-delay sound. Well-defined percussive sounds, such as
gap is large. As an example, Leo Beranek studied snare drum rim shots or single bongo strikes, can be
listener preferences and concluded that listeners distinguished at intervals much closer than 50 ms.
prefer halls with a gap of about 21 ms. This would The commonly mentioned 50-ms criterion applies
to speech and some types of music.
Many types of music have an integration time
closer to 80 ms. Some very legato types have much
longer times. The integration of multiple sound
direct sound arrivals into a single aural event does not mean
initial time-delay gap that later arrivals within that 50-to-80-ms window
first reflection have no aural effect.
If early reflections after the initial time-delay
gap (including early “artificial echoes” produced by
loudspeakers close to the listener) are sufficiently
strong, they can confuse the sense of aural localiza-
tion, making the sound source seem to be located
somewhere besides its actual location. Such strong
early off-axis reflections can also cause music and
speech to sound unnatural. These detrimental effects
only occur with discrete early echoes. If the off-axis
echoes are diffuse (i.e., spread out in time), they add
to the sense of spaciousness and apparent source
Figure 1: There is an important initial time-delay gap between the direct sound and the width. Classical-music audiences consider spacious-
first reflected sound a listener hears. ness desirable, although it must be balanced so the
ratio of direct to reverberant sound is high enough Less architecturally complex (and also less Figure 2: Scattering and
diffusion are not the same.
to yield good clarity, definition, and intimacy. expensive) means to provide diffusion include geo-
Variations include specular
The bottom line is that, while Professor Wallace metrical diffusers of various kinds. Among them are
reflection from a plane
Clement Sabine of the Harvard University physics barrel diffusers and pyramidal diffusers, which have surface (a), diffusion from
faculty was correct in identifying the importance of been available since the early 20 th century. Barrel a rough surface (b), and
reverberation time (RT) in determining the sound of diffusers look like sections sliced from a barrel and redirection or scattering of
an auditorium, it is also true that not only the late provide 1-D diffusion in the plane perpendicular to incident rays (c).
reverberation (which determines the RT) but also the axis (see Photo 1). Pyramidal diffusers provide
the early reverberant sound play important roles 2-D diffusion (see Photo 2). Some diffusers have
in concert or lecture hall acoustics. The direction a gentle curvature in either a cloud pattern or in
from which the early reflections arrive, as well as repeating suspended-ceiling tiles.
their diffusiveness, significantly affect the listening
experience. Quadratic Residue Diffusers
There is nothing magical about cylindrical or
Scattering and Diffusion pyramidal shapes for diffusers. Many different
Two technical terms have come into use in shapes have been utilized, including the one shown
acoustics in recent years: “scattering” and “diffu- in Photo 3. This product is one of many that pro-
sion.” These terms describe similar but not iden- vide absorption as well as diffusion.
tical phenomena. In 1983, RPG introduced a new type of diffuser
Scattering is the redirection of sound by specular based on slots (for 1-D diffusers) or square wells
(light-like) reflection from the surface of the scat- (for 2-D diffusers) with the depth of each slot or
terer. It is often considered as though it involved well chosen according to the branch of number
random or nonspecular reflections, since the many theory called “quadratic residues.” These diffusers
tiny scattering surfaces of a material make for a are also called refraction phase gratings (RPGs).
situation too complex to analyze except as an aggre- These quadratic residue diffusers (QRDs) operate
gate effect. But in fact, each reflection is specular. on the basis of constructive and destructive inter-
Diffusion involves wave interactions (diffraction) to ference between sound reflecting from adjacent
spread the reflected sound. wells. The relative phase of adjacent reflections
Figure 2 shows the difference between scattering depends on the well depth.
and diffusion. Figure 2a shows specular reflection QRD variations are available from several
from a plane surface. Figure 2b shows diffusion from
a rough surface. Notice that the rays representing
reflected waves are spread out. Figure 2c shows
redirection or scattering of incident rays striking
the angled surfaces of a pyramidal diffuser and
diffusion of rays striking the peak of the pyramid.
The redirection provided by a diffuser changes
the average direction of the reflected rays, while
diffusion only adds rays in more or less the same
direction as the main specular reflection. An excel-
lent illustrated discussion of scattering and diffu-
sion is available from RPG Diffusor Systems (see
Resources).
Historically, excellent concert halls had good
scattering and diffusion provided by architectural
features such as deeply coffered and ornamented
ceilings, crystal chandeliers, niches, cornices, side-
Photo 1: This barrel diffuser
wall balconies, columns, and statues. The Boston
redirects sound that strikes
Symphony Hall, designed by Sabine, benefits from it, reducing flutter echo.
diffusion provided by 18 statues placed in niches (Photo courtesy of Acoustic
above the second balcony. Surfaces, Inc.)
Removable
Capsule
Titanium
Diaphragm
Acoustical Diffusion
In addition to providing a desirable quality of
early reflections, acoustical diffusion is often used
to eliminate flutter echo in rectangular rooms, and
slap echo from rear walls or balcony faces. When Photo 4: This quadratic residue diffuser is constructed of wood. (Photo courtesy of RPG, Inc.)
using diffusion for these purposes, the acoustician
should remember that diffusers do not get rid of
sound; they only redirect it.
For example, in an auditorium or worship center Resource
in which echoes from the balcony face inhibit speech RPG Diffusor Systems, Inc., www.rpginc.com/pdfs/news/ScatteringvsDiffusion.pdf.
intelligibility for those on the stage or platform, dif-
fusion can be used to redirect the reflected sound
to another place. The question becomes, “which
other place?”
If the reflected sound is redirected upward, it
can then reflect from the wall behind the stage or
platform, producing objectionable echoes for people
seated in the front balcony rows. If the reflections
are redirected downward, the result may be that
those in the front rows of the main floor will be
affected by objectionable echoes or comb-filtering
effects. If the sound is redirected to the side walls,
the result can be “round-robin” reflections causing
late echoes at numerous places in the room. This
is not to say that diffusion should not be used to
deal with balcony-face reflections, just that such a
treatment must be carefully engineered.
Another use for diffusion is to help projection
of unamplified sound from the stage to the seat-
ing areas. In many new auditoriums, the ceiling is
carefully shaped to provide helpful early reflections.
While this approach works well in new construction,
it is difficult to retrofit to existing rooms. Appropriate
use of diffusion can redirect sound from the stage
out to the audience. The diffusive materials can be
directly attached to the ceiling, or applied as clouds.
Although the engineering use of scattering and
diffusion for acoustical design is a fairly young sci-
ence, it provides an important tool for improving
the listening experience and speech intelligibility in
a variety of acoustical spaces. ax
DAN: In grade school, I remember making up a program on tape. Something historical, but I can’t
remember what it was about.
SHANNON: Describe some of the jobs you had prior to inventing the automatic microphone mixer.
DAN: After doing all the lighting for four years at the University of San Francisco (USF) College Players
and for concerts in the USF Gym, I did sound for the Globe Theatre in 1964 and lighting and sound in
1965, and lighting for the first production of the San Diego Opera in 1965. In 1967, I switched to doing
theater sound, working for the San Diego National Shakespeare Festival and the American Conservatory
Theatre in San Francisco.
[The title “Sound Designer” was created in 1968 to describe what Dan was doing. He provided sound
services for many seasons of the Mondavi Jazz Festival, and engineered several independent record
albums, including Kate Wolf’s first two albums which are still in print, now as CDs].
Photo 1: Dan Dugan was DAN: In theaters a “sound designer” supervises the sound from the microphones to the audience’s ears.
the first person in regional In motion picture production there are two meanings. The first is the same as in theater, also called
theater to be called a “sound
supervising sound editor, and the second usage is for a person who creates novel sounds like monsters.
designer.” He also developed
the first effective automatic
microphone mixer—the
SHANNON: How did you come up with the idea for the automatic microphone mixer?
automixer. He is shown here
with his museum rack of DAN: In 1968, I did sound design for the resident companies of Hair in Chicago, Las Vegas, and Toronto.
Dugan automatic mixers. There were 36 microphones and one operator working rotary-knob mixers in a rack. I thought there
In some ways, my living room home theater rooms suffer more from bass peaks and nulls that
looks more high-tech than some of the million-dollar cause music to sound boomy or thin and individual
home theaters you see in high-end designer pho- “early” reflections that mask musical details and
tographs (see Photo 1). And, it definitely sounds make movie dialogue difficult to understand. Both
like a million bucks. To replicate the sound, you of these problems can be solved using absorbers
need to understand how I have acoustically treated made from rigid fiberglass or similar materials.
this room and why having good acoustics affects I’ll address bass problems first because they’re
the sound quality more than the type of receiver more difficult. Excess ambience and reflections
or even which loudspeakers you use. Understand at mid and high frequencies are easily tamed
that all acoustic problems are caused by reflections using relatively thin absorbers, but low
from the walls, floor, and ceiling. However, there frequencies require bass traps that
are several types of acoustic problems, and each are much larger and thicker.
one requires a different solution. F i g u r e 1 s h ow s t h e
low-frequency response
Acoustic Reverberation measured in a small
In a large space (e.g., a gymnasium) the main listening room
acoustic problem is excess reverberation. If you clap before and
your hands or yell, the sound after
may continue for 5 s
Photo 1: The home theater or longer. But,
system boasts a 159” screen home-
and plenty of acoustic sized
treatment to let music and
movie soundtracks appear
crystal clear.
Figure 1: I measured the
low-frequency response in a
16’ × 11.5’ × 8’ room, with
adding bass traps. This highly skewed response is and without bass traps. It
not only common, but typical. If you tried to sell should be obvious which
trace is with the traps and
an amplifier or loudspeaker with a response this
which is without!
skewed, you would be laughed out of business.
Yet, many people have no idea their rooms impart
similar responses—regardless of the sound. In an
attempt to document the responses, I used Room
EQ Wizard (REW) acoustics analysis software. This
Figure 2: Derived from
excellent freeware program is available for Win-
the same “before” data as
dows, Mac OS, and Linux.
Figure 1, this plot shows
Generally speaking, the more bass traps you each peak’s decay time as
have, the closer you’ll get to a flat response. It’s well as its amplitude. The
impossible to make any domestic-size room per- numbers along the upper
fectly flat, but reducing the span between peaks right show the decay times
and nulls to 10 dB or less greatly improves bass in milliseconds.
fullness and clarity.
The best locations for bass traps are a room’s
corners because that’s where bass waves tend to
collect. Note that a rectangle room has 12 corners, Figure 3: After adding
not just four where each wall meets another wall. bass traps, the response is
Bass traps can also be placed in the corners at the much flatter and the peak
wall’s top where it meets the ceiling or in the floor decay times are also greatly
corners at the bottom of each wall. Photo 1 shows reduced making the bass
instruments sound tighter
traps in the wall-ceiling corners. There is also a
and clearer.
smaller trap in the wall-floor corner to the right of
the right speaker. Another floor trap is just visible
behind the receiver and satellite TV box.
All rooms have peaks and nulls, so a room can
sound boomy and thin at the same time, depend-
ing on what musical notes happen to be playing. a distinct echo. Most loudspeakers are designed
Low-frequency peaks in a room make some bass to disperse sound horizontally over a wide range
notes sound too loud, and nulls make other notes to fill the entire room with sound. Therefore, a
sound too soft. fair amount of sound reaches the side walls and
Another problem, “modal ringing” is just as dam- reflects toward the listening spot. Since the reflec-
aging. If you’ve ever clapped your hands in an empty tions travel farther, they arrive at your ears a few
room, you have probably heard the resonant “boing” milliseconds after the direct sound. These reflec-
sound as the waves bounce repeatedly between tions reduce clarity and harm imaging because the
opposing surfaces until they diminish after a few same sound arrives from two different locations
seconds. The same thing happens at low frequen- at different times. The ceiling is another source of
cies, although you won’t hear bass resonance with early reflections. The three panels shown at the top
hand claps. Low frequencies, whose wavelengths of Photo 1 avoid reflections from the left, center,
correspond to the room’s dimensions, resonate and and right loudspeakers.
linger after the source sound has stopped. Figure 2 The rear of my living room has yet more bass
and Figure 3 are “waterfall plots” that demonstrate traps, as well as diffusers to scatter sound rather
where the peak decays come forward over time. than reflect it straight back to the seating area About the Author
(see Photo 2). Most people make all the acoustic Ethan Winer is co-owner
Bass Traps treatment in a room the same color. However, my of RealTraps, an acoustic
Bass traps work well in any room corner, though living room also occasionally serves as a product treatment company based
damaging reflections at mid and high frequen- showroom, so I have a mix of colors in this room, in New Milford, CT. His
late s t b o o k is titled
cies occur at specific locations. Referring again to as well as off-white and gray (not shown) in my The Audio Expert. You
Photo 1, the panels in the foreground at left and home recording studio, which is located in another can communicate with
right are placed to absorb early reflections from part of the house. Ethan Winer by visiting
the side walls. Reflections that arrive within about There are a total of 55 acoustic panels in the www.realtraps.com and
20 ms of the direct sound are considered “early” living room. However, many people are not willing www.ethanwiner.com.
because the delay is too short to be perceived as to install that many acoustic panels in one room.
Resource
Real Traps, LLC, http://realtraps.com.
Source
Room EQ Wizard (REW) acoustics analysis
Photo 4: Additional software
absorbers under the ceiling
peak avoid the focusing Home Theater Shack | www.hometheatershack.com
effect that otherwise occurs.
musicdirect
®
ax You Can DIY!
Vacuum Tube
Low-Frequency
Oscillator
Why build a vacuum tube oscillator when, with a few
hundred bucks, you can buy a semi-professional piece of
gear that performs better and contains a larger frequency
span? The answer is simple. It is fun and challenging to
replicate state-of-the-art technology from the 1950s. The
project is also a great way to really appreciate those who
made historic instrumentation technology.
By
Atto Rinaldo
(Italy)
The oscillator described here resembles a 1950 Design Changes
HP-200CD, which is a Hewlett-Packard (H-P) design. When building my own vacuum tube oscillator,
I have implemented several modifications to make I made a few changes and added some features
it easier to build with modern components. I did so to the original design. I replaced the frequen-
while maintaining the vacuum tube technology that cy-controlled oscillator components with vari-
characterizes most of my projects. able resistors and a fixed capacitor. The original
I had the opportunity to use the HP-200AB at used a variable capacitor and fixed resistors.
school. At the time, it was an outstanding piece of I eliminated the output transformers, mostly
hardware and superior to any other test instrument because they are impossible to find. I replaced
available on the market. While the HP-200AB was the tubes with ones that are currently available.
limited to 20 Hz to 40 kHz, the HP-200CD had a fre- I changed the output to make it unbalanced as
quency range from 5 Hz to 600 kHz in five ranges. opposed to balanced. I added an output voltage
measurement device, a 10-step attenuator, and
a frequency counter. I also used solid-state reg-
ulated power supplies.
Theory of Operation
The circuit includes a frequency-controlling
bridge and a balanced push-pull amplifier, which
form the oscillator circuit (see Figure 1).
The frequency and amplitude controlling cir-
cuits are arranged as a floating bridge. They are
symmetrical with respect to ground. This feature
ensures frequency stability, constant amplitude,
and high reliability. The bridge is fed by the bal-
anced voltage developed at the cathode of V2 and
V4 (see Figure 2).
The output of the bridge’s frequency-controlling
branch is applied to the grid of V3 and the output
Figure 1: Here is the block diagram of my modified Hewlett-Packard 200CD oscillator. of the amplitude branch is applied to the grid of
V1. Table 1 shows the oscillator’s frequency range The first condition is relatively easy to achieve.
with the components value’s formula. However, the second condition is slightly more dif-
Potentiometer P3 adjusts the bridge’s ampli- ficult to meet. It requires an automated control
tude-stabilizing branch, while lamp L1 stabilizes the mechanism on the feedback network by the lamp
oscillation amplitude across the entire frequency L1, which acts as a gain stabilizer.
About the Author
span. Potentiometers P1 and P2 vary the frequency To simplify the concept, if the signal voltage Atto Rinaldo retired
from IBM after 32 years.
within the ranges set by the value of capacitors across lamp L1 increases, (the oscillator tends to
While there, he held var-
C5 to C12. go toward saturation mode) the resistance of its ious managerial respon-
The balanced push-pull circuit includes V1 and filament will increase. In turn, this will make the sibilities in Italy, the
V3 as the true amplifier and V2 and V4 as the output V1 grid more negative decreasing its gain and vice- US, and other parts of
cathode follower. A criss-cross positive feedback is versa if the oscillator signal tends to decrease. For the world. He received
his Radio and Television
applied to keep output impedance as low as pos- more information about the concept, visit the Clifton
Te c h n i q u e s d i p l o m a
sible as seen by the cathode-to-cathode load. The Laboratories website (see Resources). in 1957 when vacuum
feedback paths travel from the plate of V2 to the tubes were the primary
grids (i.e., the control and the screen) of V4 and technology. Atto wrote
Frequency Range three books in Italian
from the plate of V4 to the grids of V2.
Valvole e dintorni (which
The oscillation is maintained through a positive 10 to 100 Hz According to: translates to “The Rise
feedback from its output to the input stage. Two a n d Fa l l o f Va c u u m
1
conditions are required to maintain this oscilla- 100 Hz to 1 kHz f = Tube Technology”) and
2π RC
tion. First, the feedback signal must be in phase Valvole e dintorni HI-
with the input signal to sustain oscillations. Second, 1 to 10 kHz With: FI (which translates to
R in Ohms “High Fidelity at Home”)
the system’s net gain at the oscillation frequency C in Farads
10 to 100 kHz and Il Fascino del Violi-
must be equal to one. If it drops below one, the no (which translates to
oscillation stops. If it is higher, the oscillator goes Table 1: The oscillator’s frequency range with the “The Fascinating Violin.”
into saturation. component value’s formula are shown.
Construction Tips
The oscillator’s construction is not recommended
for people who lack electronic assembly experience.
During a debug operation it may be necessary to
perform some troubleshooting, which is best done
by someone with electronics experience.
Photo 1 shows the oscillator’s front panel.
Photo 2 shows the internal layout and the point-
to-point wiring.
I built my oscillator on a Ballantine vacuum-tube
voltmeter enclosure. Its construction is rather com-
plex and requires the use of some critical compo-
nents to achieve maximum performance. Specif-
ically, potentiometers P1 and P2 must have a 1%
tolerance or better. This is necessary to get an
undistorted and stable waveform. I chose a ganged
10 + 10 kΩ linear type from Burns. Ideally, they
Figure 4: The high-voltage power supply details and components are shown. should be of antilogarithmic curve for better scale
a) b)
Photo 2: The components layout and point-to-point wiring are visible in this top view (a) and bottom
view (b).
Frequency Range 10 Hz to 100 kHz in four ranges range, particularly when switching from one range
to the next, select C8 and C12 with 2,100 pf and
Accuracy ±1 digit (frequency counter tolerance) parallel them with a 0-to-50-pf trim cap. You must
use caution during the components layout pro-
Signal level Monitored by meter
cess to avoid performance degradation over the
Output Maximum 10 V 70-to-100-kHz range due to stray capacitance.
Lamp L1 is critical. I used a 220-V, 3-W filament
3 to 10 V (on >10-kΩ load) lamp, but you could experiment with other lamps.
1 V–300–100–30–10-3–1 mV (Internal 600-Ω load)
Attenuator (±5%) Just adjust P3 for minimum distortion every time
300–100 µV (600-Ω load)
From 1 V down, adjustable by P6, uncalibrated lamp L1 is changed.
c)
a)
b)
Photo 3: The Hewlett-Packard 3585A spectrum analyzer waveform is
shown at 100 Hz (a), 1,000 Hz (b), and 10 kHz (c).
Resources
Clifton Laboratories, ”Bill Hewlett and His Magic Lamp,“
www.cliftonlaboratories.com/Bill%20Hewlett%20and%20
his%20Magic%20Lamp.htm.
Linecard:
• JJ-Tubes
• Genalex Gold Lion
• Sovtek
• Electro Harmonix
• Svetlana ‚S‘
• Tung-Sol
• SED Svetlana winged =C=
• Shuguang/Sino China
• PSVANE
• Full Music
• Hammond Transformers,
Chokes and Enclosures
www.btb-elektronik.de
ELEKTRONIK VERTRIEBS GMBH
Postal Adress: BTB Elektronik Vertriebs GmbH
Keplerstr. 6 · 90766 Fuerth · Germany
Phone +49-911-288585 · Fax +49-911-289191
info@btb-elektronik.de · www.btb-elektronik.de
audioxpress.com | February 2014 | 53
ax Audio Electronics
Tube Amplifier
High-Voltage
Delay
Tube amplifiers need some time for
the heaters to warm before a tube
starts conducting. Depending on
the tube type, this can range from
a dozen seconds to a minute. This
delay is not typically a problem—you
just wait a little bit before your music
starts playing.
But normally the tube’s high-voltage supply comes on at the same time as the heater
voltage, often supplied by the same transformer. That means the tube sits there for up to
a minute with high-voltage applied and no conduction. This can appreciably decrease tube
life. Again, it depends on the tube type and the actual high voltage, but the effect is there.
By It is compounded by the fact that the high voltage in the initial, unloaded situation can have a
Jan Didden much higher value than when the amplifier is operating and the high-voltage supply is loaded down. To
preserve the tube life, a device should delay the high voltage until after the tube heaters have warmed.
(The Netherlands)
There are many ways this can be done. The simplest method is probably to use a delayed relay that
switches on the high voltage after a preset time delay. This can be achieved with a discrete or integrated
timer circuit to activate the relay. You can also buy a timer unit that includes a relay and an integrated
delay you can set with a dial. But I don’t like relatively bulky and mechanical systems with their power
dissipation and mechanical failure modes. So, I thought there must be a more elegant and more reli-
able way to do this.
To solve my dilemma, I built a power supply unit (PSU) delay. However, I had a few additional require-
ments for the device including a solid-state switch used for noiseless and long-life operation, low power
dissipation, a programmable delay time, easy integration even in existing tube amplifiers, and no impact
on amplifier audio performance.
Figure 1 shows the topology I selected for the last two requirements. By inserting the switch in the
return line to the power transformer, I left everything after the first reservoir cap unchanged so there is
no impact on the power supply quality and no changes in any carefully laid out ground circuitry. It can
be used either with a bridge-type rectifier or a double-phase rectifier as shown.
The actual switch is a MOSFET connected between D and S. The unit also contains a time-delay cir-
cuit. The switch and the delay circuits are powered by a spare heater winding on the power transformer.
Figure 2 shows the complete circuit.
The spare heater voltage on J3 and J4 is followed by a voltage doubler to sufficiently raise the volt-
age for the control circuit and MOSFET gate drive. The control circuit is simple but you need to examine
it a bit closer to see how it works.
Let’s start with the switch Q1, which is an N-channel high-voltage MOSFET. The drain is connected to
the transformer return line at J1 (see Figure 1). The source at J2 is connected to ground. If the switch
is closed, the transformer’s secondary return is connected to ground and the high voltage is applied to
the amplifier.
FREE
Software
On the right you see microcontroller U1, which controls the switching. U1 is a small
eight-pin dual in-line (DIL) chip that is programmed to turn the MOSFET on after a set
delay. I will discuss that later. R3 provides a sample of the AC mains voltage to the
controller. This enables the controller to detect the mains zero crossing and switch
the supply on at the right moment. The result makes for a smoother increase in the
high voltage and avoids any high in-rush currents. You design it
Initially, the controller holds the MOSFET in the Off position by shorting the gate to to your specifications using
the drain via the activated transistor Q2. R2 and C4 provide that function for the first our FREE CAD software,
Front Panel Designer
few milliseconds until the controller is operational. Following the delay, the controller
releases Q2 by pulling down its GP10 Pin 7. R2 increases the gate voltage. The MOSFET
switches on at the next mains zero crossing and the high-voltage supply circuit is
enabled. The controller also drives a bi-color LED to show a delay or an active state.
The controller has an internal shunt regulator, which is supplied via R7. VSS is
the controller ground pin, and VDD is its supply pin (not shown on the chip drawing).
Since I had one pin left at the controller chip, I connected that to J5. Depending
on whether J5 is jumpered to ground, the controller selects one of two delay times:
low or high. We machine it
and ship to you a
Controller professionally finished product,
You may wonder what business a digital (there, I said it!) controller has in audio. no minimum quantity required
The answer is none, of course. That’s why it isn’t in the audio. And if you are afraid
of “digital hash” or whatever, rest assured: After the initial delay and the switching
●
Cost effective prototypes
on of the high voltage, the controller goes to sleep and switches itself off until the
next time you switch on the amplifier’s mains voltage. and production runs with
If you are new to controllers or are unfamiliar with them, it may seem like a bit of no setup charges
magic, but it is not rocket science. The controller program is simply several instruc- ● Powder-coated and anodized
tions executed by the chip after it is powered on. For example (in order of execution):
finishes in various colors
• Set Pin 7 to VDD to keep the MOSFET switched off ●
Select from aluminum,
• Read the voltage at Pin 5 to determine whether the jumper is placed and choose
acrylic or provide your
either the Low or the High delay value from memory
own material
●
Standard lead time in
5 days or express
manufacturing in 3
or 1 days
FrontPanelExpress.com
1(800)FPE-9060
Figure 2: The complete circuit for the high-voltage delay unit contains a time delay.
2 D2,D3 1N4002
DISCUSS
design tips tutorial software
mobile
engage
contests
engineering tools
audio business
system
data
networkingmedia
COMMUNITY
social media
talk
Want to talk to us directly?
information
Share your interests and opinions!
projects
product news
Check out our New Social Media
Outlets for direct engagement!
The fundamental frequency range of a guitar Figure 4 shows the response of such a speaker.
using standard tuning extends from 80 to 640 Hz, Later amplifiers often used 12” speakers. Begin-
assuming 24 frets. Figure 1 shows the waveform ning in the 1960s, the speakers were classified as
produced by an electric guitar with two humbucking having either an “American” or a “British” tone (see
pickups. The low E string was open and the upper Figure 5 and Figure 6).
three strings were fretted at fret 24. Some guitarists prefer premium speakers such as
Figure 2 shows significant harmonic content is the instrument speakers produced by Electro-Voice,
present up to about 2 kHz (roughly the third har- JBL, and (in the past) Altec-Lansing. These premium
monic of the highest note). If the waveform is heav- speakers have much flatter frequency responses and,
ily clipped, the frequency range is much wider (see therefore, do not affect the tone as much as the
Figure 3). To accurately hear the guitar, including the more common speakers. Since the speaker’s sound
distortion components, we would need the amplifier is inextricably a part of the guitar amplifier’s overall
and speaker to have a response from almost DC to sound, you must keep in mind the characteristics of
over 10,000 Hz. Actually, the upper harmonics are the speaker when considering the tone of an amplifier.
unpleasant, so most guitar speakers roll off well The tone of the guitar amplifier itself depends
below 10 kHz. Also, few speakers, even bass speak- on the low- and high-frequency cutoffs and the
ers, have responses below 40 Hz. design of the tone controls. Most amplifiers (not
including the speaker) have responses that are
The Speaker Helps Set the Tone reasonably flat from about 60 Hz to 15 kHz, if you
The speaker is an important ingredient in deter- ignore the tone control circuit. The speaker takes
mining a guitar amplifier’s tone. Many of the earliest care of the rolloff of undesirable high-frequency
amplifiers used 10” speakers with seamed paper distortion products.
cones and surrounds (see Photo 1). There are two notable exceptions. One is the
110
100
90
80
SPL
(dB) 70
60
50
Photo 1: Early guitar amplifiers used 10” speakers such as this one.
40
110 500 30
20 50 100 200 500 1,000 2,000 5,000 10,000 20,000
105
200 Frequency (Hz)
100
100
95
75 10
70
6
65
3
60
20 50 100 200 500 1,000 2,000 5,000 10,000 20,000
90
Frequency (Hz)
Figure 6: This “British” speaker does not have the midrange dip so it has
a different tone.
Figure 7: The Gibson EH-150 included a tone switch located between the Figure 8: This “losser” 1-potentiometer tone control has been popular,
preamplifier and the driver. especially in lower-cost amplifiers.
Figure 11: The bass control in the circuit shown in Figure 10 is boost-only. The green curve Midrange Control
shows full boost of both controls. Red is the full cut of both controls. To provide even more control over the amplifi-
er’s tone, some manufacturers added a midrange
control. The so-called “bass-mid-treble tone stacks”
reception has been mixed. Some guitarists like them
very much, while others cannot find any setting of
these controls that suits them.
The circuitry and curves are shown in Figure 12
and Figure 13. This amplifier also had “deep” and
Figure 12: This bass-mid- “bright” switches. The curves shown in Figure 13
treble tone stack was used were made with deep switch on.
by Fender in the AB-864 A few amplifier manufacturers (e.g., Orange)
guitar amplifier. experimented with design to provide a unique tone.
www.hometheatershack.com
Audio/Video Discussion Forum
FO R SE R IO U S AUDIO PHILES !
There are two aspects of a technical book that readers like to see, but which are often poorly
presented. These two aspects are accuracy and a pleasing readability. Readers less skilled
in mathematics and interested only in formulae that provide a valid ‘recipe’ like to skip the
Theory and practice with design
derivation of that ‘recipe’. In this book this is possible without prejudice to the readability of
the story. These ‘recipes’ are framed at the end of formulae derivations. methods for self construction
There are readers who are not satisfied with formulae that ‘fall from the sky’ without
derivation.
For these readers, a ‘recipe’ without substantiation is unsatisfactory and the application of
■ COMPONENTS the formulae will be accepted only when the derivation is shown. The formulae are proved in
such a way that the reader can easily follow them through, thus a deep knowledge of mathe-
■ HEADPHONES
in the application of audio amplifiers, including their power supplies, for the design and DIY
construction of these electron tube amplifiers. This is much more than just building an
electron tube amplifier from a schematic made from the design from someone else: not
■ INTEGRATED AMPLIFIED
only academic theory for scientific evidence, but also a theoretical explanation of how the
SPEAKER
tions, then SOLUTIONS
practice works. No modern simulations, but because you first understand the circuit calcula-
you can work with your hands to build the circuit and last, but not least, if you
have a multimeter, a signal generator and an oscilloscope, you can measure the circuit
■ MEASUREMENT
parameters yourself to see that theory and practice are very close. That is the aim, and
MICROPHONES
makes this book a unique reference source.
■ MICROSPEAKERS
Rudolf Moers
■ SPEAKER PARTS
ISBN 978-0-905705-93-4
■ TEST EQUIPMENT
Elektor International Media
www.elektor.com
Cover Sponsor
1 2 3
Read, learn, and do it yourself! See the niques. Then you can build the circuitis your-
• The impact of feedback on internal over-
This Masterclassload;
• Stability of feedback amplifiers
presented by Jan Didden
website for details on all seven editions. self. This is a must-have resource!
•
•
Feedback ‘speed’ and the difference
between delay and phase shift;
Positive feedback and its effects;
Jan Didden has spent most of his
working life with the Netherlands Air
• Simultaneous application of positive and Force and NATO, and in that period
negative feedback in a 1950s tube am- wrote many audio design articles for
the audio press. After his retirement,
SKU #BK-ELNL-978-0-905705-93-4
plifier
• Error Correction according to professor he finally found time to finish al
Malcolm Hawksford; those accumulated projects.
• Error Correction as used in the paX Am- At present his main interest is on er-
plifier published by Elektor in April and ror correction in power amplifiers,
May 2008; which leads to unusual designs. His
2 LOUDSPEAKER INDUSTRY
• Feedback or Error Correction - a case latest design, the paX power amplifi-
study: QUAD Current Dumping; er, was published by Elektor in 2008.
Since a few years Jan also publishes
SOURCEBOOK 2013
This Masterclass provides a clear overview of Linear Audio, a unique, six-monthly
the benefits that can be obtained by feedback ‘bookzine’ dedicated to technical On this DVD:
and its sibling, error correction; but also of audio, with articles written by inter-
• 140 min. video presentation
4 MASTERCLASS FEEDBACK
its limitations and disadvantages. Recom- nationally recognized audio design
T he 2013 Loudsp eaker Industr y mended to audio designers and serious audio
hobbyists.
experts. • All material as PDF documents
Bonus:
Sourcebook is the most comprehen- IN AUDIO AMPLIFIERS DVD
J. Didden
140 min.
video pres
looking for design software, drivers, view of the benefits that can be obtained
enclosures, manufacturing equipment, from feedback and error correction. It 4
microspeakers, microphones, test and also details their limitations and disad-
measurement equipment, or voice coils, vantages. It is recommended to all audio
this book has you covered. The new 2014 designers and serious audio hobbyists!
Loudspeaker Industry Sourcebook is com-
Further information and ordering
ing soon!
SKU #DVD-ELNL-978907920-16-5
www.cc-webshop.com
SKU #BK-ELNL-978-94-90929-01-5 CONTACT US:
audioXpress
111 Founders Plaza, Suite 300
East Hartford, CT 06108
USA
Phone: 860.289.0800
Fax: 860.461.0450
E-mail: custserv@audioxpress.com
JAMES THOMPSON
Lead Systems Engineer
The NEW Dayton Audio DTA-120
High Fidelity Power Amplifier