Producing Your Own Classical Guitar CDs:
(An Ongoing Saga)
As much as I enjoy performing on classical guitar, my major focus in recent years has been on recording. This may seem a backward approach. Aren't we first supposed to make our name as concert artists and then hope major record labels will start calling us up offering contracts? I'm not waiting around. To quote a Phil Ochs song, "If I've got something to say, sir, I'm going to say it now!"
Apart from one CD which is on Centaur Records and distributed by them, I sell my CDs at concerts and on the Web. But at this time, rather than focus on how to get more people actually buying my recordings, I prefer to start the next recording project -- a process which has taught me more as a musician than any lessons I have taken.
To go at your own pace, on your own schedule and with a thorough knowledge of the equipment at hand, one needs a recording set-up at home. For Guitara Illuminata: The Music of Josquin des Préz, I knew I would need a multi-track unit (as I would be playing each of two guitar parts). I called a music store and the salesman told me I should check out the new Roland VS-880 8-track hard-disk recorder, which he had just seen at a trade show. He described it as "wicked stupid" which, in case you don't know, is a high compliment from someone born in the mid-1970s. Selling for $2000, it was one of the first such recorders to have adequate storage (or track minutes) for a whole guitar duet album. I bought the first one he could get in, which included the internal digital effects board (another $250). This earliest version of the operating system was still a little wet behind the ears. The newest version, the VS-880EX, now sells for $1750 and includes everything I had to subsequently pay extra for (which totaled $2750!). Upgrades for the VS-880 are issued on Zip disks. This caused me to buy a Zip drive which plugs into the VS-880's SCSI port at a computer store, giving me a reliable way to store audio data from off of the hard disk. (The alternative storage method, using a DAT recorder, took forever and was not reliable). In fact, you may record directly to a Zip disk (though its 100 MB capacity doesn't buy you much time).
At this time I also charged up a DAT (Digital Audio Tape) recorder (the Tascam DA-20) for mastering. DAT recorders can work with different "sampling rates", but CDs only use one: 44.1 KHz. In addition to recording the master tape, this unit came in handy for combining tracks created in separate "songs" (the basic level of organization on the VS-880) into one continuous album, without leaving the digital domain.
The more I made use of the VS-880's digital editing capability, the more I depended on it. It allowed me to record much as Glenn Gould is said to have, seamlessly splicing together the best of sometimes many takes. Some may say this is cheating, but the process did much to shape my concept for each piece which enhanced my live performance in a way that hours of simply practicing could not. I also found myself taking more risks as a player, my performance not becoming too "careful" out of fear of making mistakes.
There are now many other stand-alone hard-disk units to choose from. I would not recommend the Mini-Disc recorders for CD-quality classical music. There are also many hard-disk systems based around your computer, that function in more or less the same way but often without the dedicated hardware interface. If you hope to release CDs of your live unedited performance, however, you may be able to get by with a simple stereo DAT recorder like the portable Tascam unit that has mic inputs and one or two good microphones. But be advised that even a note-perfect live performance will require digital editing for professional results.
Unless you're going to do extensive soundproofing, you need to record in a space without exterior walls. I use the upstairs bathroom, adding ambience digitally (via the above mentioned internal effects) .It's easier to make clean edits when the ambience is not recorded with the guitar so you're not editing out the reverb "tail" of any note that you are grafting onto another.
Another advantage to digital reverb is that you can put off deciding how much of it to use until final mixdown. The president of Stedman (whose mic is described below) complemented me on the natural ambience heard on my "Clear Away" CD; he was surprised to learn that most of it was digital reverb. Eric of Earthworks (another microphone company) scoffed at digital reverb as being nothing more than a series of echoes, unable to match the complexity of the real thing. But if you listen carefully to the sound of a space that you like, you can emulate many of its characteristics with the digital reverb's parameters. In particular, I found it effective to cut the low end of the reverb's "wet" signal quite a bit. My ambition is to someday fool even Eric. Some engineers have captured natural reverb by playing the final mix over speakers in a space of their choice (at, say, 6 AM, before traffic begins), recording the reflections along with the original track. Of course, if you're the kind of guitarist who can walk into that same space and play a note-perfect performance before a plane passes or the garbage trucks make their rounds all the better.
In addition to a quality pair of headphones, one needs a pair of near-field (you're meant to sit close to them) "reference monitor" speakers. I bought the Alesis Point Sevens, being the least expensive option; I also figured I didn't need the "low end" of larger speakers for mixing a guitar CD. In retrospect, I would say this is almost true as I do hear more of my recorded guitar since adding a powered subwoofer.
To get the signal from any quality microphone into the recorder one needs a mic pre-amp with phantom power. This is not a good place to scrimp, as the first one I ordered ended up being too noisy though it took my ears a few months to fully perceive it.
I started out with an Audio-Technica 4031 small-diaphragm condenser mic. Though "recommended" for guitar I would say in retrospect that purchasing this $300 mic for recording was a mistake. After a period of time I noticed it was boomy in the bass, harsh in the treble and put out too much of its own hiss. It has redeemed itself as a live mic, however. [When I need amplification in concert, by the way, I use a combination of a saddle pick-up and external mic. Internal mics don't cut it to my ears].
My next microphone was an Earthworks. Advertised as the only mic for under $1000 that doesn't "smear time", the Earthworks OM-1 uses a very small diaphragm that comes to rest for more quickly than many more expensive mics, making its reproduction far more accurate particularly noticeable to me in the upper range. It being omnidirectional (Hears in All Directions) means that it doesn't get boomy when placed close to the guitar (our ears are also omnidirectional, by the way). I got mine drop-shipped direct via a retailer in Wellesley for a total of $400. Although it meant re-recording the whole project, I've never regretted in the long run having to do this (even when I lost most of a subsequent project through a system error), as many improvements will suggest themselves.
While the Earthworks mic made a significant improvement in the sound, after a while I began to realize that its self-noise was just as bad as that of the AT4031. I sought advice from Centaur's president and concluded that the least I could spend on a quieter mic that sounded as true as the OM-1 would be $1200. My testing of a number of well-known studio mics (costing between $350 and $950) confirmed this assessment; they all (including the famous AKG-414) made my fingernails sound like plastic picks. Feeling I couldn't afford the $1200 (it's somewhat inconsistent the role that "feelings" play here), I tried other pre-amps that might minimize the OM-1's noise settling on the dbx 286A single-channel preamp/processor for $220. Its Expander/Gate feature cut the noise in gradations such that it became far less noticeable. I also found its built-in Compressor and De-esser useful the latter attenuating shift squeaks somewhat. It is easy as well, on the VS-880, to edit out any noise between pieces with precision. I used this set-up to record the Voice of Creation CD. To simulate stereo miking, I used the "distance" effect available on the VS-880, panning its "return" away from the original signal. [See diagram in "Evolution" article]. Another method for simulating stereo miking of guitar has been described by author/musician Craig Anderton, involving high and low cut EQ and two 5 ms (millisecond) delay lines.
I sent a copy of the CD (how it became a CD so quickly, I'll get to later) to Earthworks. They called to say they were impressed with the playing and asked if I would do some test recordings for them in return for one of their new mics, the QTC-1 "quiet omni". So I drove up to their southern NH factory to pick up the fancy equipment they wanted used in the recordings.
In the meantime I was contacted by a local recording engineer who wanted experience recording classical guitar; he was happy with a copy of each CD as payment for recording my next local performance. From his arsenal of high-end mics we settled on a pair of Neumann 130s. They sounded very good but I still couldn't see going into debt over them. At a subsequent session, however, he pulled out a pair of Stedman C-15s. My head told me they shouldn't have been good for guitar: large diaphragm (which are most commonly used for vocals) and cardioid (Hears in One Direction generally too boomy for close miking of guitar). But my ears told me otherwise; and I was floored to learn that these mics went for only $400 each. He loaned me the pair to compare with the pair of QTC-1s Earthworks had loaned me. While I would recommend either of these pairs of mics to a classical guitarist (the Earthworks particularly when you want to capture all the ambience of your recording space), I concluded that the Stedmans sounded warmer. The QTC-1 costing twice as much (as either the Stedman or Earthwork's earlier OM-1) made the C-15, which is also a bit quieter, a better value. I ordered a matched pair of the Stedmans from the same Wellesley retailer. (After all, I had a free QTC-1 coming anyway). My engineer friend has since informed me that a new mic by Shure (KSM32) sounds comparable to the Stedman (to his ears).
One problem with a sensitive large-diaphragm condenser mic is its vulnerability to moisture; and a spell of hot humid summer days soon caused one of the Stedmans to malfunction. The president of that company was very responsive. At first he replaced the electronics of the mic, which probably wasn't necessary. He explained to me that excessive humidity would cause problems not just for his microphones, but eventually for disk drives and potentiometers (the knobs and sliders on your stereo) as well. Large diaphragm condensers are just "the canary in the coal mine", as it were. A few weeks after I had brought down my humidity (alternating between a dehumidifier and an air conditioner) the mic returned to normal functioning. The lower humidity also enhanced my ability to concentrate.
Mic position (mono)
I decided the best position for the mic was pointing at the soundboard two inches north of the fifteenth fret, about six inches out. When you record up close with sensitive equipment, you often find out for the first time about extra-musical noises you make: squeaks from chair or footstool, breathing, subtle tongue/saliva swishes, even the cracking of joints.
At the same time, every nuance you give the music will also be rendered with detail; and you can count on better sustain of voices than when recording from a greater distance. Of course, this method is not about capturing the audience's perspective, which is quite a bit further from the instrument. But the further back you go, the more you must increase the gain, and your pianissimo will be competing with mic hiss and environment noise. Most classical recordings try to capture the feeling of Being There, but let's face it: they're not "there" and neither are you.
If you record with a lot of ambience (natural or digital), those listening in ambient spaces are hearing the effect compounded. I'm reminded of that whenever I see my music is being choreographed in the dance studio. Artur Rubinstein wrote, on the back of his album "The Brahms I Love", that he preferred hearing the pieces through a stereo at home -- a more intimate place in listeners' lives -- over attending a performance in a concert hall. You're trying to convey the depth of your art through someone's car stereo, clock radio, living room speakers or Walkman headphones. I feel it's our choice what works best to that end even though many well-qualified people would argue against close-miking.
You may need to re-finger certain passages to minimize squeaks on the wound strings. I discovered "back buzz" on one guitar, requiring the insertion of a thin plastic shim under the sixth string at the nut, and a third string buzz on another guitar requiring a similar solution.
One also must focus in on whether a cut in EQ is required at any frequency. After determining that the mic's position is not the problem (you're generally asking for trouble if you point a mic at the soundhole, for example), listen for any notes that may pop out and calculate their frequency in cycles-per-second (a.k.a. Hz). In standard tuning, the open strings have the following frequencies: 1 - 330 Hz, 2 - 248 Hz, 3 - 202 Hz, 4 - 146 Hz, 5 - 110 Hz, 6 - 83 Hz. The VS-880 has a pretty capable EQ section that can be applied either before or after recording a track. You also may want to record through a Compressor/Limiter, so you can keep the signal pretty hot without distorting on a big chord but some purists prefer to simply keep the overall level set lower.
Even for just a solo guitar recording, levels between pieces, and often within pieces, will need to be evened out. Hard as you might try, there will always be takes when you sat in a slightly different position relative to the mic(s) -- for which you will need to compensate. The VS-880 offers automated mixing, either within itself or via MIDI from a computer sequencer. I use the latter approach -- since I have the interface and the software -- as it allows me to basically replicate the VS-880 internal mixer on the computer screen and program fades, level changes or effects parameters. Should you record the audio itself onto a computer-based system, mixdown and editing are even easier -- as every parameter may be varied from one screen rather than from two units operating in synch. In the big leagues, specialized Mastering Engineers are utilized at this point to optimize the recording so that it sounds its best on the various playback devices it will be heard through.
"Burning" a CD
The next upgrade available for the VS-880 was the capability to physically manifest CDs, via new software and compatibly designed CD-R recorder. I remastered two earlier albums of mine (from DAT), thrilled to finally hear them reproduced through CD players with the clarity they had lost in their previous life as cassettes.
We call it "burning" a CD because -- unlike tape that has its magnetic particles temporarily aligned in accordance with the audio signal it records -- a CD is etched permanently with tiny indentations that are read by a laser beam on playback -- an irreversible process. Having the ability to burn CDs directly from the VS-880 obviates the need for a DAT recorder -- as most CD replicators prefer a CD master to a DAT and will charge you less when you can provide it. Digital Audio Tape is a somewhat fallible medium, one glitch in the tape and you’ve got to remaster on a fresh one.
The great thing about burning your own is how you can make changes -- after listening to the CD in a variety of settings or after getting feedback from others -- before shelling out for a package of 1000 from one of the many companies offering this service. The disadvantage is in the time it takes. For an hour CD, the VS-880 spends two hours writing an "image file" and then half an hour on each CD (so long as you’re present to keep feeding CD-R blanks into the CD recorder). Occasionally, you get a defective disc, or accidentally jostle the sensitive unit, which requires you to start the whole process again. Add to that the time its takes to print and cut the inserts and stick on the donut-shapes labels. Ironically, you don’t save any money doing it yourself when compared with the per-CD rate available when you commit to a factory run of 1000. But this also depends on the price you can get the CD-R blanks for.
Even without a professional sound card for your computer, you can take advantage of a computer-based system by importing tracks from a CD you’ve burned on the VS-880, mastering them there with digital audio software (saving them as WAV or AIF files), and burning the finished CD directly from the computer. This last process is much faster on a computer than with the VS-880 as you don’t have to wait for an image file to be created. You can also use audio editing software, once your sound files are in the computer, to even out levels, reorder tracks, etc. I use Peak LE, which I purchased for only $50.
Stereo Miking & A Bargain
Now, with the larger drive, I had plenty of track-minutes to record a true stereo guitar album. Needing at least two quiet pre-amps, I ended up purchasing a small Mackie mixer, which -- in addition to other important features -- offered four. In fact, now I had what would be needed to record a CD of Modern Dance drama "Clear Away: A Fisherman’s Farewell", which I had scored for 11 musicians.
In recent months I came across a microphone bargain that you should all know about: the Oktava 012 small-diaphragm condenser from Russia. I found them new for only $150 each (rubles, anyone?) and they sound indistinguishable from the Neumann 130s [$1200 each — referred to in Part 2]! Humidity doesn’t bother them in the least; and they are quiet.
In stereo miking, one runs the risk of phase cancellation -- making you sound brittle and sometimes even out of tune. My understanding is that so long as you keep the two mics panned hard right and left, you’re OK. But when you bring the image toward the center of the stereo field -- or make it monaural, as will happen on a mono unit such as a TV -- frequencies common to the left and right can cancel each other out. One can avoid phase cancellation by placing the diaphragms directly over one another, at an angle between 90° and 120°. Doing this with clunky large-diaphragm mics requires the top mic to be positioned upside-down. [See further discussion of stereo miking in the 8/98 issue of Recording]. Unlike the mono omni mic position I referred to in Part 2, I place my pair of cardioids at least a foot away from the instrument, making sure that neither points directly at the soundhole. Many professionals go further back, to pick up room ambience.
I hope you’ve found most of the above useful in some way. I intend to continue experimenting with many of the possibilities I have referred to -- including the note-perfect first take in an ambient space at 6 AM. May your own projects be blessed.