LETTERS to Electronic Musician

OUT OF CONTROL Thanks for the excellent article about MIDI winds (In Control, May 2001). However, I take issue with the review of the Akai EWI3020. Akai
Publish date:
Social count:
OUT OF CONTROL Thanks for the excellent article about MIDI winds (In Control, May 2001). However, I take issue with the review of the Akai EWI3020. Akai


Thanks for the excellent article about MIDI winds (“InControl,” May 2001). However, I take issue with the review of theAkai EWI3020. Akai may have worked extensively on the EWI in the past,but it has done nothing to it in at least five years. Furthermore, themodule is unnecessarily huge — a full-depth 2U rackmount unit,which makes it unwieldy for road use.

Perhaps Akai bought too many large rackmount cases about six yearsago and needs to recoup its investment. On the other hand, the new MIDIEVI, custom-built by Nyle Steiner, offers significant improvements toAkai's product line, namely better MIDI compatibility and morefeatures. In addition to being MIDI compatible, the MIDI EVI has acontrol voltage (CV) output (a much more sensitive response thanMIDI).

Many EWI owners are waiting to get their hands on an improved MIDIEWI, if only Nyle Steiner would build one.
Ernie Mansfield
via e-mail

Ernie — I agree that the EWI sound modules are large andunwieldy. I also agree that the Steiner MIDI EVI is a big improvementto the Akai EWI in terms of features and MIDI capabilities, and it doeshave a CV output. However, not many synths are available that can takeadvantage of that feature, and it's difficult to access because it'scarried on an unused conductor within the MIDI cable. To access it, youneed to build a box that routes that conductor to a CV output. On theother hand, the Akai sound modules can process external analog signalsthrough their filters and amp sections, which the EWI controls. Thatcapability is not available on any other MIDI wind controller.
— Scott Wilkinson


I've been reading EM for almost 15 years. I've alwayslooked to it for information about what's happening in the MIDI, synth,and recording worlds, and usually I have been satisfied with thearticles.

However, I noticed that the article about alternative MIDIcontrollers (“In Control,” May 2001) didn't mention StarrLabs' Ztar line of alternative controllers. I own a Ztar Z2, with theingenious Ztar synth controller neck, and an electric guitar with a hexMIDI pickup. I've been following guitar synthesis trends since themid-'80s and have never encountered an interface like the Ztar's. Itnot only brings guitarists a fairly straightforward way to triggersynth sounds but also works within MIDI's limitations as a keyboardinterface protocol.

The Ztar deserves way more attention than it has received. Pleaseinclude it in the next relevant article about MIDI controllers or,better yet, dedicate an article to exploring it.
Michael LaMeyer
via e-mail

Michael — I agree that Starr Lab's controllers offer anoutstanding degree of control and MIDI implementation that isdifficult, if not impossible, to achieve with mainstream MIDI guitars.Nonetheless, the focus of my article was on mainstream controllers.Starr's instruments, however closely they resemble guitars, are notmainstream controllers.

Associate editor Gino Robair and I covered several Starrcontrollers in the August 2000 cover story, “The OuterLimits”; “What's New” in the July 2001 issue has awrite-up of the Starr Z6 and Z6-S; and a review of the Starr Z1 andZ1-S is in progress.
— Marty Cutler


Larry the O's comments in “Final Mix: Stuck in theMiddle” (June 2001) are right on the money.

I am the broadcast and recording engineer for Atlantic Records'Internet department, and I own a commercial studio in New York City. AtAtlantic, I use a Digidesign Pro Tools/24 system (hardware andsoftware), a Yamaha O2R digital mixer, and a pair of Genelec monitors.At my studio, I use a Mark of the Unicorn (MOTU) 2408mkII system(hardware and software); a Mackie 32-8 analog mixer; Mackie HR824monitors; Event 20/20 monitors; and a consumer monitoring setup, whichconsists of an Onkyo M-501 power amp, a pair of Allison bookshelfspeakers, and a Panasonic boom box, all for A/B comparisons. Bothstudios use Mac G3s. I have a clear understanding of what Larry the Ois talking about.

For the high-profile Atlantic recordings, the tracking gets done onthe Pro Tools system, recorded in 24-bit, 48 kHz format. I thentransfer the tracks to a FireWire hard drive and bring them to mystudio for post-production work. There was a time when that would havebeen impossible without having an expensive setup to handle thecritical applications of truly professional work — but notanymore. Although Atlantic's pro system is great, you can't beat themodern midrange audio systems. No matter what audio sequencing softwareyou use, the 2408's I/O can compete with the big dogs.

The power has been brought back to the people by the outstandingefforts of companies such as MOTU, Echo, and so many others. Larry theO is right when he says that rough times are ahead with new operatingsystems, protocols, and software. If you can ride out the storm, it'sonly going to get better.
A. J. Tissian
via e-mail


As I have been considering the purchase of a new computer for audio,I sent an e-mail inquiring about Windows ME versus 2000. Thanks to themagazine and associate editor Dennis Miller for the timely andimmensely helpful response and for the pleasant surprise of seeing myletter in the June 2001 issue.

EM is the highly valued center of my music-referencelibrary. As a reader for more than ten years, I have many issues torevisit, all of which contain good, trustworthy advice. My projectscontinue to improve not only with practice but also from reading eachissue.

Thanks to the entire staff, which consistently does an excellentjob. You should sleep well knowing that you help many people pursuetheir passion with confidence and satisfaction. You really do make adifference.
Brian Strines
via e-mail


Dan Phillips's article (“Debunking Digital-Audio Myths,”May 2001) covered many interesting points. I found it frustrating,however, when he failed to provide at least basic definitions fortopics such as fixed- versus floating-point calculations. I know that'san important topic because I've heard of it so many times indescriptions of computer-music gear. However, Phillips basically saidthat there is a distinction between the two — they result indifferent calculations by the computer — but gave no explanationof what they are. I still don't know whether one is better than theother.

Likewise, he wasn't clear about how plug-in compressors and limitershave certain “frequency-related issues.” What does he meanby saying compressors and limiters “work by modulating oneaudio-rate signal with another audio-rate signal”? Does audiorate refer to the sampling rate? That section, which I reallyneeded to understand, wasn't written in clear terms — unlessyou're a total tech head.

I'm making the same observation that a reader(“Letters,” May 2001) recently made. Please havenontechnical editors edit the articles to make sure at least basicdefinitions are provided for the lexicon used in more technicalsections. Take a look at the articles in your sister publicationRemix, in which almost everything is explained clearly.
Michael Arana
New York

Michael — The question of fixed point versus floatingpoint was mentioned as a side point, so I could spare only a fewparagraphs. It's a highly technical subject, and a complete discussioncould easily fill an entire article.

The two formats have varying ways of approaching the possiblerange of values. For audio, think of the difference between silence andfull-scale signal. Floating point has finer resolution near zero thanat larger values, resulting in greater dynamic range; fixed point hasconstant resolution throughout its range, culminating in finerresolution for larger values.

Bit resolution is often not the same when comparing fixed- andfloating-point hardware, which makes apples-to-apples comparisonsdifficult. Some tasks may be easier or more processor-efficient,depending on the choice of fixed or floating point; even chips capableof both may have fairly contrasting capabilities when doing one or theother.

As for which calculation is better, I don't have an easy answer.Highly regarded digital signal processing programmers have differencesof opinion on the matter. Some point out disparities in ease ofprogramming; others point to algorithms that seem to work better in onecase or the other. Because the experts disagree, it's safe to say thatyou don't need to worry about it.

My point was not that one is better but simply that, in somecases, they may not produce identical results, and if a developerimplements the same synth or effect on floating-point and fixed-pointsystems, the two may sound slightly dissimilar.

Regarding compressors and audio-rate signals, audio rate means asignal with its frequency in the audio range, approximately 20 Hz to 20kHz. A compressor's gain control may respond to incoming audio soquickly that it produces signals within that range.

With 1 ms attack and release times, for instance, a compressorcan act like a 500 Hz oscillator. That may produce frequenciessignificantly higher than 500 Hz, depending on the waveform shapecreated by the attack and release characteristics.

The frequency produced by the gain control interacts with thefrequency of the signal being processed and produces aliasing noise ifthe combined frequencies exceed half the sampling rate (the Nyquistfrequency). To diminish aliasing, use slower attack and release timesand, if possible, use processors or plug-ins with higher internalsampling rates.
— Dan Phillips

Michael — We do indeed have nontechnical copy editorsworking on the stories. However, EM has never been anentry-level magazine with respect to the articles' technical level. Weconstantly do a balancing act. On one hand, we don't want to losereaders who lack an extensive technical background, so we try toexplain the underlying technical issues when we can. On the other hand,if we never assume a certain level of technical knowledge, we can'tdiscuss more advanced issues that are of interest to our manytechnically savvy readers.

Fixed- versus floating-point calculation is a good example of ahighly technical issue about which we had to assume readers already hadsome knowledge.
— Steve O


Address correspondence and e-mail to “Letters,”Electronic Musician, 6400 Hollis Street, Suite 12, Emeryville, CA94608 oremeditorial@primediabusiness.com. Publishedletters may be edited for space and clarity.