ALL EQ PLUG-INS BEING EQUAL

This is regarding your article “Plugging Into EQ” by Nick Peck (October 2003). I found the article informative, especially in its explanation of FIR, IIR, and FFT filtering algorithms. However, I felt that the article was diminished by its comparison of plug-ins that run on dedicated hardware cards and calculate in 48-bit fixed point (such as Pro Tools plug-ins) with plug-ins that run on a CPU and calculate in 32-bit floating point. One exception was Anwida Soft's GEQ31V, which runs on 64-bit floating point.

The disparity between 48-bit fixed point and 32-bit floating point is greater than 16 calculation bits (see “Computer, Do the Math!” by Oliver Masciarotte in the February 2003 issue of Mix magazine). The real comparison should be made between 48-bit fixed-point plug-ins and 64-bit floating-point plug-ins.

I believe that Anwida's GEQ31V was a poor choice for this article because it's a graphic equalizer and not parametric. It would be more fair to line up the 48-bit fixed-point plug-ins from Bomb Factory, Focusrite, and Sony against 64-bit floating-point native plug-ins such as Sonic Timeworks' parametric equalizer (which comes in two flavors), DSound's Simple Audio Plug-in Pack parametric equalizer, iZotope's Ozone 2 parametric equalizer, Wave Arts' TrackPlug parametric equalizer, Voxengo's CurveEQ VST, and Sonalksis's SV-517 Stereo EQ.

I understand that you were looking to compare cross-platform products and this imposed the 48-bit fixed/32-bit float issue from the onset. But I feel it would have been in the better interest of your readers to go to the bottom line: sound-price performance. I think more information on CPU DSP as opposed to proprietary-card DSP serves the greater EM community. Though I would say that both EM and Mix are commendably informative for all things audio, I feel that Mix is more geared towards the established Pro Tools/Apple community and its more expensive counterparts while EM is more appealing to the, shall I say, independent entrepreneur community.

I think your readers would want to know whether in fact less expensive 64-bit floating-point DSP competes with more expensive 48-bit fixed-point DSP. Pro Tools has a clear advantage for now, but computers will always get more powerful. I predict that the next step for proprietary DSP will be to standardize at least 56-bit fixed point to stay competitive as 64-bit DSP becomes more common.

For a future article, I think it would be good if you compare 48-bit fixed and 64-bit floating formats head to head, potentially shedding light on a more cutting-edge issue.
Merrick Fleisher
Bellsong Recording Studio
Brooklyn, New York

PC POWER

I wish that EM would test Windows products on machines equipped with Pentium III or higher processors. Readers learn little when such products are tested on Celeron chips, which are not really suitable for serious audio work.

Of course the Digidesign Digi 002 [covered in EM's November 2003 cover story, “All Fired Up”] will not run well on a Celeron chip, but I want to know if it is right for my Pentium 4. The manufacturer's Web site states that the product needs at least Pentium 4 or Xeon CPUs running at 2.0 GHz. They probably know what they're talking about.

And in the case of your Native Instruments Vokator review (November 2003), your own sidebar (“Minimum System Requirements,” p. 96) clearly states that the manufacturer specifies Pentium III or higher.

So testing on a Celeron processor is not useful.
Mitch Greenhill
via e-mail

Mitch — The Vokator review and the Digi 002 feature coverage turned out to have two different problems. In the Vokator review, our error was in leaving out the complete system requirements. In fact, according to Native Instruments, a Celeron/800 MHz or better is adequate to run Vokator, although the program probably would run better given the Level 2 cache incorporated into the Pentium. So reviewer Brian Smithers's test system met the manufacturer's specifications.

The Digi 002 story is more problematic. Indeed, Digidesign does not recommend the Celeron for use with the 002, but the Digidesign staff could not explain the reasons to our author. We find it hard to understand how the use of a Celeron instead of a Pentium could affect the operation of the 002 hardware. (Obviously, we understand how it can affect the Pro Tools LE software bundled with the 002.) No other FireWire interfaces we covered had problems with the Celeron. Nevertheless, you are right, the manufacturer specified a higher-end CPU, and we should have caught this problem. We completed our tests on a Macintosh system that met Digidesign's specs, so our overall conclusions about 002's performance are valid.

We apologize for the errors and certainly feel we have learned a lesson. Thanks for pointing it out. — Steve O

STACKS OF SOFT SYNTHS

Hi, and thanks for the great magazine. Many subscriptions have come and gone but yours will continue coming to my mailbox due to the great info that you provide.

I am writing in response to a letter (“Letters: More Firepower, Please,” November 2003) regarding Cakewalk Sonar's inability to echo a single MIDI input to more than one channel (that is, for layering sounds from soft synths). I recently upgraded to Sonar 3 and have happily discovered that this is now possible.

This newest version is perfect for layering various soft synths, including those using the VST standard. It also allows for the rewiring of native soft synths (like Cakewalk's Project5 and Propellerhead's Reason) for additional layering within Sonar 3 if desired.

The resulting sound is nothing short of spectacular. I have been comparing sounds from the included soft synths (such as Speedsoft's Vsampler) to those of my Korg Triton ProX and can only say that the soft synths are at least as good as if not better than the Korg.

I had problems with previous versions of Sonar and Pro Audio and was a bit skeptical about upgrading. However, Sonar 3 is outstanding, and when coupled with a powerful PC, can totally replace standalone hardware-based synths and samplers. The latency I can achieve with my MOTU 828mkII audio interface is 2.2 ms with five layers of synths using 30 to 48 percent of the CPU.
Mark Krebs
via e-mail

EM ARCHIVES REVISITED

As an avid EM reader I have several large boxes of back issues. Even though I usually read them pretty thoroughly when I receive them, there are good reasons to reread the older ones (especially when current issues bring up topics that are covered in back issues). Thus my several large boxes.

I second the suggestion of Dave Quick (“Letters: EM Archives,” October 2003) — an EM archive on CD! That's something I'd pay good money for — that is, assuming I won't have to pay at a rate of $10 per back issue!
Jim Bates
via e-mail

I read that you would like to know whether your readers are interested in having back issues of EM archived on a CD and sold to us. Of course! As long as the price is not ridiculous, I could definitely see purchasing your archives, especially if you did it in volumes, for use in my personal studio — even if I had to scrape some extra pennies together. I've come to trust you guys, anyway, so I'm not worried.

Remember one thing: it needs to be searchable. Two o'clock in the morning is not the time for reading through an index!
C. Ross
Brooklyn, New York

G'DAY FROM OZ

To the crew at EM: I scoured the Internet for a magazine that gives practical, understandable advice for aspiring recording engineers — now, I have found it! Top work.

I get a lot out of reading the articles that you print, but at times I get lost in the technical details. Perhaps more articles like the one titled “Loud, Louder, Loudest” [“Square One,” August 2003] would be great. This article was superinformative and very readable, but still respectful of the audience.
Matt F.
Melbourne, Australia

WE WELCOME YOUR FEEDBACK.

Address correspondence and e-mail to “Letters,” Electronic Musician, 6400 Hollis Street, Suite 12, Emeryville, CA, 94608, or emeditorial@primediabusiness.com. Published letters may be edited for space and clarity.