You are reading the older HTML site

Positive Feedback ISSUE 70
november/december 2013


AES 2013
by Scott Dorsey


The annual AES conference is a mixture of a professional audio trade show, combined with a scientific conference downstairs. Increasingly there have been program items and workshops attempting to straddle the two. What I love about the show is that it has engineers and customers interacting together, and you can see the people who actually designed a product if you have a question for them.

I enjoy writing up the show for a high end audio audience because there is important research going on that really is advancing the technology, and at the same time I am very interested in the confluence of the high end world with audio production.

A decade or so ago, the booklet for the papers sessions was maybe a quarter inch thick while the booklet for the trade show exhibitors was twice that size. Things have changed and this year they are nearly reversed. There are fewer exhibitors every year, but at the same time the number of really great workshops, papers, panels and programs increases dramatically every year. I don't get a chance to see anything but a tiny fraction of the show, I tend to spend more time upstairs with the trade show, and I missed almost all the really interesting workshops and panels. Don't consider this an exhaustive review, it's just a list of the things I looked into and happened to have seen. Because of the way the show is structured, you could go there and see a totally different set of things.

Another thing is that the big companies seem to be downsizing their booths while smaller companies with more esoteric products are expanding. The days of the two-storey Ampex and 3M booths are gone, but this year it was odd just to notice that Radial Engineering, a maker of DI boxes and related gadgets, had a larger booth than the huge Shure Microphones company.

There were quite a few papers that either weren't presented at all or were presented by someone other than the first author due to the US government furlough that was taking place. I know some of the Library of Congress folks went on their own dime in order to present.

I didn't get a chance to see more than a tiny fraction of the show, being only one human being. Some of the papers I describe, I only briefly saw part of the presentation and relied on the preprints in order to summarize them because I was zipping through so quickly. So, if I made an error, or I missed something amazing and wonderful, please forgive me because after four days of talking to people it's a wonder I can even still stand up.


Probably the most interesting thing is that Schoeps was introducing a new studio vocal microphone, something they have not made in many many years. Although it looks like the old Schoeps 201M of the 1960s, it has a modern small diaphragm capsule with a baffle around the capsule in order to make the off-axis response more like that of a larger capsule, built into an odd palm-shaped cast grille with modern electronics that are very quiet. And yes, the front end FET is mounted right at the capsule so the odd size and shape does not make for problems with long wiring from the capsule. A quick test on the show floor never is all that enlightening but it sounded good there and seemed to have good rejection which I think is an underappreciated attribute of a vocal mike that has to be used in the real world.

Neumann was showing another new low-cost addition to their line, the MKH107,a large diaphragm multi-pattern microphone that is a completely new design and seemed very clean. Definitely worth checking out, it had a figure-8pattern with a real null to it.

Microtech Gefell, formerly a division of Neumann until the Berlin Wall was established and it found itself in the Eastern Bloc, has long made a sort of line-array microphone called the KEM 970 which is intended for applications where extreme directionality in one plane is called for. They now have an improved version, the KEM 975 which has an additional rear capsule which improves directionality at low frequencies. It's a shame they did not have one of these to actually audition because the show floor is a great place to show off the directionality of this sort of product.

The guys from TAB Funkenwerk were showing off a number of microphone parts, clones of older Neumann designs for repair of old Neumann microphones.

Another interesting new microphone is the AEA N22 ribbon microphone, a completely new design built to a much lower price point than any of AEA's previous designs. The motor assembly is new but they are still using an RCA-style ribbon and aiming toward the classic RCA sound. It's designed with active electronics because it's intended for the project studio market where quiet preamplification is not common.

On a totally different note but still with a conventional ribbon design are two new microphones by Mark Fouxman of Samar Audio Design. So new they don't even have names yet, he's built some lower cost ribbon designs with a very clean and accurate top end, quite the opposite of the old RCA sound. The test ones had amazingly good and clean nulls, better than I have heard from any of the more recent ribbon designs.

Probably the most exciting new microphone product at the show was an old favorite redone. The folks at Granelli Audio Labs took a Shure SM-57and put a fitting in the middle of it that puts a right-angle bend in the middle of the microphone without changing the internal volume. It's simple, kind of silly looking, but looks to be a lifesaver trying to shoehorn spot mikes into drum sets.

Triad-Orbit was showing off microphone stands that were very, very cool, with highly articulated ball joints that locked and unlocked very easily and quickly. They look like they could be a lifesaver at a festival or any other application where quick setup and change of spot microphones is important.

Sure, you've seen baffled omni microphone configurations like the Jecklin Disc and the Schoeps Sphere in which foam-like or rubberlike material is used as a barrier to make omni-microphones directional at high frequencies. But what if you wanted source material to play back on a huge array of 22speakers? You'd probably want to use a smaller array of microphones with baffles, and some matrixing stuff to direct them. In "Portable Spherical Microphone for Super Hi-Vision 22.2 Multichannel Audio", Kazuho Ono and others from NHK do just that. Preprint 8922.

In "Sound Field Visualization using Optical Wave Microphone Coupled with Computerized Tomography", Toshiyuki Nakmiya and other from Tokai and Kumamoto Universities in Japan set up a microphone with a semi-transparent diaphragm and run a laser through it. The laser is reflected both from the diaphragm and an external fixed reflector, and the two reflections beat against one another creating a signal that varies with diaphragm position. They are then scanning the laser to map the position of the diaphragm at all points. Their intention is to visualize a sound-field in 2-space but I think diaphragm resonances and the small size of the diaphragm used makes this impossible in this case. Still, it is a very interesting exercise and could be used for studying diaphragm behaviour if nothing else. Preprint 8923.

Although it is not really a microphone paper, I am going to mention "Reproducing real-life listening situations in the laboratory for testing hearing aids" by Minaar, et al. from Otikon in Denmark here. They made environmental recordings with a similar arrangement, a spherical microphone with 32 omni capsules around the surface, and then used a matrixing arrangement to reroute them to an array of 29 loudspeakers, employing deconvolution of measured transfer functions to route the signal. The ultimate idea here was to reproduce an enveloping sound environment in which to test hearing aids but that same notion of reproduction of an enveloping environment is important elsewhere. Preprint 8951.

At the last European show, Juha Backman from Nokia talked a bit about models of fluid flow around microphones for determining wind noise characteristics. In "Numerical Simulation of Microphone Wind NoisePart 2: Internal Flow" he talks about flow across the grille and into the microphone itself. He shows a couple simple electret capsule designs and how the shape can make air flow smoothly or turbulently around them, that affecting wind noise pickup considerably. Preprint 8925.

How much headroom does a vocal microphone really need? In "Maximum Averaged and Peak Levels of Vocal Sound Pressure", Braxton Boren, Agnieszka Roginska, and Brian Gill from NYU attempted to measure vocal levels as part of an attempt to validate estimated levels from an experiment conducted by Benjamin Franklin in 1739. The details of that experiment recreation are in a different paper, but in THIS paper they measured singers with peak values as high as 110-114 dBA at one meter. Is your microphone linear up there? Should it be? If levels are that high atone meter, just imagine how high they can be right up front of the mouth for a PA mike. Some of this duplicates existing work but Preprint 8958is worthwhile for both their new work and a summary of existing studies in the field.

In "Controlling Drum Bleed with Laser Vibrometry", Andrew Greenwood and Sebastian Chafe from Sennheiser use a laser sensor to detect when drums have been struck and use that in order to control gates on drum microphones for PA applications. This gives more reliable gate operation in high noise and bleed environments like music festivals, compared with using the microphone levels to key the gates. Useful? I don't know, but still very ingenious. E-Brief 109.


There seems to be a real renaissance these days in the old Heil AMT tweeter, with adaptations of that classic design appearing in a number of studio monitors, notably the Adam ones. They are controlled and yet very detailed on top which is very good for that application. In addition to Adam, a company called Eve Audio from Germany is now selling a line of 2-way and 3-way studio monitors with integral amplifiers and dsp-based crossovers in them. It's hard to tell what anything really sounds like on the show floor but this looks like another design to watch. Unity Audio was showing a couple of smaller 2-way and 3-way active monitors using "ribbon tweeter" designs which also looked like AMT derivatives, combined with built-in amplifiers designed by Tim de Paravicini. They seemed solid and with controlled low end although of course it is hard to tell on the floor. On the papers sessions, there were a lot of folks talking about speakers. For example, in "Digitally Steered Columns: Comparison of Different Products by Measurement and Simulation", Stefan Feistel and Anselm Goertz set up eight different steered array speakers in a fairly reverberant church, set them all to the same patterns, and tested intelligibility at a number of listening positions with the STI method (which estimates intelligibility from impulse response). Then they used vendor's provided pattern information (as supplied in DLL or GLL format files) and used EASE ray-tracing software to estimate intelligibility. They found that, aside from two outlying measurements, all the loudspeakers performed very close to the simulations and they all performed very close to one another in terms of intelligibility. The authors note that under more severe conditions they may diverge more but remark on how amazing it is that we have come to this point of consistency. Preprint 8935.

Adam J. Hill and Malcolm Hawksford talked "On the perceptual advantage of stereo subwoofer systems in live sound reinforcement". They tested large auditorium rigs in appropriate rooms with stereo source material and recorded test samples at various points in the room with a binaural head. They found that although there was no real stereo imaging, the coverage at the sides of the room were improved by the speaker configuration, which was unexpected. Preprint 8970.

There were crossover papers too. In "Linear Phase implementation in loudspeaker systems; measurements, processing methods, and application benefits," Remi Vaucher from NEXO talked about building mixes of FIR and IIR digital filters that allowed control of the phase response of the crossover in such a way that they could match the reverse of the speaker response and make a speaker system that has flatter and more controlled phase response, very close to acoustic linear phase. He points out that folks have used FIR filters for crossovers before, but blindly and not really taking the speaker responses themselves into account. Preprint 8926.


In the high end community there is almost an obsession with imaging, but yet nobody really agrees on what good imaging is or whether a given system really images well. In "Perception testing: Spatial Acuity", P. Nigel Brown from the Ex'pression College for Digital Arts proposes a listening test procedure using synthesized test sounds, primarily with the goal of detecting hearing problems. However, such a test also has great applications to determining positional accuracy of playback systems with a test listener panel. Preprint8972.

In "Listener Preferences for In-Room Loudspeaker and Headphone Target Responses", Sean Olive, Todd Welti and Elisabeth McMullin from Harman International presented listeners with a reasonably flat playback system and asked them to equalize it with simple tone controls to their liking, then the actual system response was measured. It was found that people on the average tended to turn the bass up, that the headphone response they preferred had about 2dB lower bass and treble than the speaker response they preferred (for all listeners) and interestingly enough that the source material did not seem to make much difference (which makes us assume that all of the source material selected must have been well-balanced subjectively). Very interesting for anyone doing listening tests especially of headphones. Preprint 8994.

Now, that same Sean Olive twenty years ago wrote a paper in which he tested listener's ability to adapt to changes in playback timbre. That paper was validated by "Auditory adaptation to loudspeaker and listening room acoustics" by Pike, Brookes, and Mason of the University of Surry. They had listeners compare headphone playback to speaker playback and doing so they were able to judge timbre very well, but when they had a few minutes difference between references they were unable to because their ears adapted to the new sound quality. In this paper, they verified Olive's original work and established that it wasn't due to response bias. This sort of work is surprisingly important because so many of the standard experiments we rely on for understanding psychoacoustics have not been exhaustively repeated; more testing on different populations strengthens them. Preprint 8971.

In "On the Influence of Headphones on Localisation of Loudspeaker Sources, Darius Satongar, Chris Pike, Yiu Lam and Anthony Tew talk about doing various testing in which people compare loudspeaker and headphone playback, and in order to do this they use open-backed headphones and listen to the speaker systems with the headphones on so there is no delay between the two. How does this affect the sound of the speaker playback? It turns out, less than you would think. They make measurements and then do subjective listening testing, both of which show some headphones (notably Stax Earspeakers) can be quite transparent to outside sounds. Preprint 8953.

In "Listener Adaptation in the Control Room: The effect of varying acoustics on listener familiarization", Richard King from McGill University and others set up a room with variable acoustics; absorptive and reflective surfaces hidden behind baffles that could be rotated to change the sound of the room. They gave subjects a series of simple balancing and mixing tasks and they found that people adapted very quickly to changes in the hall acoustics. How fast? Order Preprint 8959 and find out.

Everyone complains about the explosion in file sharing causing the death of commercial recording, but in "Music Consumption Behavior Of Generation Y and the Reinvention of the Recording Industry", Barry Marshall from NEIA at least tried to quantify it. He gave surveys to young students asking them about their behaviours, attitude and motivation and while he didn't find anything new or particularly insightful, the summary is worthwhile. Preprint 8956.

And finally, I was on a panel called "Lies, Damn Lies, and Audio Specs" in which various people discussed both the importance of measuring audio systems and the way that bad measurements have caused such harm in the industry and fostered skepticism about measuring anything at all.


Lots of folks were there, from the RealTraps people to the Soundproof Windows people, to folks like Munro Acoustics who specialize in room design. But what was new this year was a company from Japan demonstrating the "Shuzuka Stillness Panel," a broadband panel absorber made of an aluminum honeycomb mixed with foam, structurally stronger than foam alone and because it's intended for use like a gobo spaced in front of a surface, it is more effective than simple foam mounted directly on the surface.


There were a bunch of events downstairs about production. All my life, I have been hearing people talk about how pop music has been progressively getting brighter, but the reality isn't quite that simple. In Spectral Characteristics of Popular Commercial Recordings, 1950-2010", Pedro Duarte Pestano and others averaged spectra of all the Billboard Number One pop hits and plotted them, showing a very interesting shift with time that shows more of a change in level than anything else but shows curious shifts in the upper midrange as styles have changed. Preprint 8960.

Paul "Willie Green" Womack gave a talk as an E-Brief called "Respect Due: The Urban Mix Engineer", basically talking about practices and procedures for rap music production and how they differ from typical production methods and why. He seemed a bit Rodney Dangerfield at times but was very enlightening as he explained to a bunch of traditional folks what rap music was actually about. This really wasn't appropriate to be an E-Brief but it was still a really interesting talk where I learned something useful. Convention E-Brief 110.


Fred Schukert, one of the principals at ATR Magnetics who has really worked hard to bring back analogue tape, is introducing a line of various processing hardware under the name Fredenstein. Including a large number of API 500-series modules, as well as rackmount mike preamp and a limiter, they are mostly based on an op-amp module with a discrete JFET input stage, which is compatible with the classic Jensen JE990 designs. I want to commend them for being the first company so far that I have seen putting usable and readable metering on 500-series modules.

An Italian company called Sknote was showing a few new boxes, including their Vastaso stereo compressor and distortion device, and their Rame stereo passive EQ device (well, passive networks with FET make-up gain). Looks like another new addition to the market.

ATM Bettermaker from Australia was showing off their EQ232P, a digitally controlled analogue equalizer. These days with digital consoles all over, people have got to expect the ability to store and recall settings on parametric filters, and this device allows you to do it with a conventional analogue equalizer. Very, very cool idea, and it even allows you to integrate the equalizer with an automation system.

KuSh Audio was showing off a line of equalizers designed more as effects boxes than as traditional equalizers. Not neutral sounding, not intended just to alter frequency response like a conventional equalizer, they can be a useful addition to a producer's palette of sounds.

Serpent Audio was showing their "Chimera" optoelectronic compressor that fits into an API 500-series package.

And a number of classic companies were showing off preamplifiers and equalizers, folks like Gordon, Great River, D. W. Fearn and Daking who have long been fixtures at the show.

There were a few papers as well on processing hardware. In "Evolution of Dynamics Processors' Effects Using Signal Statistics", Timothy Shuttleworth points out that most manufacturers talk about compressors and limiters in terms of a simple transfer function (maybe with or without a knee) but don't talk at all about ballistics or any time-related effects. He gives a number of measures to statistically characterize the action of such a device. While a single stage compressor might be possible to characterize fully as transfer function and a mass-spring system, anything beyond that becomes quickly difficult so it only becomes meaningful to view more complex devices in terms of statistics. I think his measures are incomplete and miss some things like timbre effects but could be extremely useful to describe more complex dynamics processors. Preprint 8938.

There was a LOT of new research being done in digitally modeling existing older processing hardware. I think this important work both in that it helps folks understand what is really going on with existing hardware and to allow emulation of the hardware with DSP.

In "Digital model of the passive James/Baxandall tonestack", Christopher Bennett, Jonathan Toft-Nielsen, and Connor McCullough do a simple circuit analysis on the traditional Baxandall tone control circuit and reduce it to a simple z-transform that can be easily implemented digitally. What is interesting about this presentation is less the end result but that it is a well-explained demonstration of simple z-transform implementation of a real-world filter circuit. Nothing earth shattering but worth looking at for the elegant analysis. Convention E-Brief 124.

On the other hand, Priyanka Shekar and Julius Smith from Stanford CCRMA attempted to model the Aphex Aural Exciter in "Modeling the Harmonic Exciter". They used the FAUST package to develop a model of a filter and an asymmetric clipping device as used in the Aphex, but they didn't make any attempt to verify that the clipping was in any way like that of the original (which is not abrupt clipping due to the imperfection of the diode used and therefore will have a different spectrum). Then, on top of that, they didn't have an Aphex to compare the actual results of it, so they compared their results against someone else's digital model. This is really kind of shameful, I think. Convention E-Brief 104.

In "Simulation of an analog circuit of a wah pedal: a port-Hamiltonian approach", Antoine Falaize-Skrzek and Thomas Helie use the new Port Hamiltonian Systems approach to model the circuit of a Cry Baby wah-wah pedal in terms of a set of 2-port and 3-port devices in a graph, then implement that as a big matrix, and then use that matrix to solve the system. Again this is not a practical simulation in any possible way because they make no attempt to deal with any nonlinearities and they do not attempt to verify the model against a real-world device, only against a fairly low-fidelity SPICE model of the circuit. Still, the talk was a good introduction to a new modeling method that may have more useful applications in the future. Preprint 8981.

Guitar amplifiers are odd sort of effects devices in and of themselves, because they have multiple electronic stages each intended to operate in a nonlinear regime. In "Gain Stage Management in Classic Guitar Amplifier Circuits", Bryan Martin of McGill demonstrates how two different classic amplifier designs behave very differently due to changes in the gain structure between them affecting the parts of the circuit which clip and how they clip. I find this very interesting coming from a world where we work hard to set gain structures to avoid clipping, so the notion of doing FFT analysis in order to deliberately characterize it was an eye-opener. Again not a sufficiently complete model to actually simulate an amp, but an interesting view to how guitar amps make different sounds. Preprint 8986.

One common way of modeling time-varying nonlinear systems as a single matrix is the Generalized Hammerstein model, which is effective and useful and can be derived from a set of swept-sine tests at varying levels, but which is computationally difficult. In "A Computationally Efficient Behavioral Model of the Nonlinear Devices", Cho, Kim, Yu, Park, and Yang develop a simplified calculation method that is close to the Hammerstein model in accuracy but requires only a third as much CPU power to compute. More efficient models mean more faithful models can be implemented with the same hardware and that is always good. Preprint 8930.

One of the weirder filter designs out there could be found in the Korg synthesizers, which used a filter whose corner frequency could be adjusted by an external voltage supply. In "Modeling the Korg35 Lowpass and Highpass Filters", Will Pinkle from the University of Miami generates a digital model of these filters. He models them as Sallen-Key filters and ignores all nonlinear effects save for that of an internal clamping stage (which would be the primary source of nonlinearity in the original). When compared with the original, the response plots look the same and the model goes into uncontrolled oscillation at the exact same gain level as the original which is a good proof of fidelity. Convention E-Brief 103.


As existing wideband telecom infrastructure goes away and media folks are forced to rely on internet services, the ability to get reliable, low latency audio from distant locations is being affected. In "Low-Latency Replacement of ISDN and 4-wire for Remote Broadcasts", Anthony Faust from Atlantic post describes a method whereby several different internet paths are used and the same data sent through multiple sources in an attempt to minimize transmission errors and timing issues. It's kind of inelegant but the low cost and ubiquity of the internet makes it a clear win for remote broadcast applications if something like this can be employed. Convention E-Brief 120.

And, every year there are one or two papers on audio for communication systems. This year an interesting one was "Portable Speech Encryption base Anti-Tapping Device" by C.R. Suthikshn Kumar from the Defense Institute of Advanced Technology in Pune, India. He describes a system that takes audio, digitizes and encrypts it, then modulates it onto a tone sequence that can be encoded by cell phones and transmitted. He did a lot of hand waving about the encoding and the ability to get data through a lossy encoder at reasonable rates, but the discussion of encryption and DSP throughput was interesting. Preprint 8999.

Another one of those was "Pilot Workload and Speech Analysis: A Preliminary Investigation" by Rachel Bittner and others. This team put pilots on a test where they performed air traffic control instruction read backs under various different conditions where they were performing other tasks, and compared the workload with measured parameters of the subject's voice. They found tone changed, but not the rate of speech. Preprint 8985.


Lots of new stuff in the higher-end mid-sized console range. Tree Audio was showing off a hybrid tube-based console called the Roots, based on their own channel strip modules and with dynamics processing on each channel.

API had another small modular console, called The Box, using their modular system but with some stuff built into the frame. It's an odd configuration to my mind, but designed for project studios based around a DAW and it's a very elegant thing.

Ocean Audio was showing off an entry level side-by-side console that consisted just of summing and routing, faders and master section. You could populate the channel strips as you like with any API 500-series modules. At $20k, it's more expensive than non-modular consoles but it's a lot cheaper than a similarly-configured API. Build quality looked excellent.


DirectOut was showing their line of 32-channel A/D and D/A converters, in their new Andiamo series. They have put a big investment into the MADI interface which has actually been around for some time but which is only now beginning to take off. It has become a real industry standard for multichannel devices and the Direct Out people have really spent time getting it right. In the same line, also they had a device for converting multiple (64!) AES/EBU streams to and from MADI in order to integrate MADI gear into existing systems and vice-versa.

Also from Germany, Mutec was showing off a line of various devices for generating, distributing, and controlling data clocks for synchronization and reducing jitter. I spent some time looking at their MC-7 which is a 1->8 signal distribution amplifier for word clock which squares up and stabilizes the clock waveform before passing it on. Very big deal for use in a larger facility where a lot of digital gear needs to be kept in lock-step.

Weiss is from Switzerland and not Germany, and they are now making not only some of the finest A/D and D/A converters out there but also now lower-priced unit called the DAC-2 with S-PDIF, AES/EBU, and Firewire inputs which would seem a great thing for a home installation. They also were showing off sampling rate conversion software, done right and carefully by people who know numeric analysis and know where the precision issues are. I didn't realize until this show that Daniel Weiss was the guy behind the early Harmonia Mundi equipment that first made digital recordings tolerable.

With the increase in the use of floating point for internal processing, work on operations precision becomes important. In "Experiments with Dither in Level-Calibrated Floating Point Audio Processing, Douglas "Tad" Rollow from Sennheiser attempts dithering floating point audio data by applying unmodulated noise to the mantissa, which turns out not to work. Sometimes knowing what not to do is important knowledge, and this was interesting in great part because the end result of the process was not what I would have expected. Convention E-Brief 106.


Zaxcom was showing off their digital wireless systems but they also had some great little digital recorders in a very wide variety of configurations, mostly in the 10 or 12 track range and with wireless microphone recorders built into the box for film sound work. Very clean and convenient field recorders.

Tascam was showing a DSD master recorder, their DA-3000, designed for two channel recording but with the ability to chain multiple units together


On that very subject one of the most interesting things is that John Hardy Co. has upgraded their Jensen 990 Discrete Op-Amp Module so that it is internally built with surface mount components. The tighter packaging improves performance and they have upgraded a few components as well in the process. The 990 is a very handy module for building all sorts of projects, as well as the integral element around which the John Hardy microphone preamplifiers are built.

NKK and Electroswitch were there with very good lines of mechanical switches.

Lundahl was showing off their excellent line of transformers. This year they had a couple of new products including some silver-wire stepup transformers for moving coil phono cartridges. The whole new line, the LL1943Ag, LL1933Ag and LL1941Ag are specifically designed for the needs of the high end MC user. They also had a new 1:110 ribbon microphone transformer, the LL2916, and a 1:4 splitter, the LL1944.

BeStar is a Chinese company that has been at the show before, and they were showing off various very tiny speakers and microphone elements intended for use in appliances from cellphones on down to hearing aids. Not really intended for hi-fi applications but with a lot of interesting uses.

Anaview was showing off some OEM Class-D amplifier modules for integration into power loudspeakers as well as standalone amplifiers and receivers.

Down in the paper sessions, there were a lot of folks talking about class-D amplification who didn't show up. It seems like half of the amplifiers talks were cancelled. But in "The Audio Performance Comparison and Method of Designing Switching Amplifiers using GaN FET", Jaechol Lee and two others from Samsung compared a cascoded output circuit with a gallium nitride transistor and a small silicon transistor. Although the cascoded pair has a higher turn-on voltage than the gallium nitride device alone, it is faster, and so losses are reduced. In addition, the cascode inverts the signal as the depletion mode gallium nitride FETs are normally turned on until turned off by gate voltage. Preprint 9008.


Fifteen or twenty years ago you could go out on the show floor and see companies selling lacquer blanks, cutting heads, and various record playback gear. Al Grundy from International Cutting Head could be seen in the lobby holding court. These days Transco and Audio Devices aren't showing lacquers and Al Grundy is sadly deceased, but at the same time there is renewed interest in record cutting and playback just as this generation of people experienced in the technology is beginning to pass on.

Timestep from the UK was showing off their Timestep RA turntable, actually a Technics SL-1200 with a very fancy drive device which can keep it at controlled speeds from 14 to 170 rpm in 0.01 rpm increments. They claim reduced cogging and plenty of torque and that all of the sonic issues with the Technics have been solved with their outboard control box. Definitely of interest to anyone working with 78s.

Tom Fine gave a presentation on the brief history of 35mm magnetic film recordings for music. During the sixties there was a short era when Everest, then Mercury, Command and a few smaller labels were using 35mm magfilm for music recording and the results were spectacular but the costs were high and the equipment cumbersome. The technology fell to the wayside but the recordings remain. I missed his presentation but got to see his slides and talk to him about it and I wish I'd had a chance to see the thing.

There were a number of other presentations on playback of discs, including Marcos Sueiro Bal from WNYC who gave a talk on "Disc Playback Stylus Selection" in which he discusses how to choose the correct stylus for playback of 78s and other records which do not follow the modern microgroove standard. Sadly that talk was later after I had already left so I can say nothing about it except that this sort of thing is essential to keep information on older formats alive.


In "Recording History in Audio Education", Jeffrey Ratterman from the Front Range Community College made a very fine appeal stating that teaching history in audio production programs was essential. Without knowing how we got here, it's hard to tell where we actually are, and many of the standard practices we use today come from methods and technologies that have fallen by the wayside but still have valid uses occasionally. The only problem with this paper is that he was preaching to the choir; I wish every audio educator out there could have seen this. Preprint 8933.


Neutrik was showing off their excellent line of connectors, including some outdoor fibre optic connectors for remote broadcast use and the like.

Ghielmetti is a company from Switzerland that I'd never heard of, and they are making patch bays for analogue audio and AES/EBU using tiny little gold-plated connectors. Very ingenious design with the ability to stick normal blocks into each slot as well as individual connectors, and with far greater density than 1/4" or Bantam plugs. Also they were giving out excellent Swiss chocolate which attracted a lot of folks to their booth.


The war against (and for) loudness goes on and on. This year, Bob Katz has announced the end of the war, and while I doubt he's right I surely hope he is. He recently performed loudness measurements of iTunes radio using the new version of iTunes in which levels are regulated using Apple's Sound Check algorithm, which is a good and important thing. This system works using metadata provided by the streaming server to adjust levels. He reports that average loudness levels remain fairly constant, and points out that producers can turn Sound Check on using their own iTunes instance to get an accurate notion of how it will sound over the "air." He considers this a huge improvement in the loudness war situation and I agree with him although I am curious what percentage of online listeners are using iTunes Radio.

I don't know where else to put this paper, but in "Music to our ears: the effect of background music in higher education learning environments", Adam Hill from the University of Derby talks about background music and some studies showing it improved learning among elementary school children. He points out that the music needs to be fast-paced enough to be noticed by extroverted people who are hard to distract, while not fast-paced enough to excessively distract introverts. He shows one study at Princeton but points out that there has really been no long-term extensive work on education with background music and proposes some possibilities. More of a proposal than an experimental paper, this can be found in Preprint 8932.


One of the more exciting things to my mind is that every year the AES has competitions for students, both for audio production and electronics and software design. I feel really bad about missing these competitions this year, especially seeing that I was supposed to be a judge for the student design competition and missed it due to fall-out from the government furlough issues.

If you are a student, or you know a student, or you have students, I want to strongly suggest you take a look at the AES student website at and look at the student competitions because I think these are some of the most exciting and severely underrated events and a great way for students to get real feedback on their work, see what others are doing, and meet people in the industry doing that work.


The audio press was well-represented with magazines like Recording, Sound On Sound, Tape Op and the like there, but there were also people from the Circuit Cellar, and the AudioXPress, Voice Coil and Elektor group. The British publication Resolution had folks there and I want to commend them for being one of the few folks delivering actual measurements with their reviews.

Linear Audio is sort of a magazine and sort of a paperback book, it's like the Evergreen Review of the audio industry and it is filled with some of the greatest writing on audio technology. Much more technical than the mainstream press magazine, it has accumulated authors like Doug Self and Erno Borbely. They have just published the latest two volumes which they had available at the show.


There was too much stuff at the show to see, but I saw as much as I could. This year it seemed like there was less on the trade show floor, but what was there was better than usual. The increase in workshops and crossover presentations was great. You should have gone. You can still go to the next one in Berlin, or the one after that in Los Angeles.