The Renaissance artist Michelangelo, when asked how he sculpted his statue of David, is famously reported to have answered "I just took a block of marble and chipped away everything that didn't look like David."

The process of assembling an anime dub is very much like that. To do a proper English adaptation requires a sort of craft, with everyone involved from the adapter to the director to the actors carefully working with and molding the material, paring away anything that doesn't belong.

Headline Studios in New York -- nestled in the small town of Irvington (named for Washington Irving, the Sleepy Hollow guy), right on the bank of the Hudson river -- has used this almost painstaking process to produce some of the best English anime dubs on the market. Click here for a brief history of Headline as told by the studio's producer, Joe DiGiorgi.

I spent a few days with Headline and even helped dub an anime myself, the otaku slice-of-life comedy Genshiken. I was the Sailor Moon Hentai Fanboy, the Pervert With Camera, the Boy on the Beach, and the teacher in the Kujibiki Unbalance extra on the DVD.

Faith and the Funny

Everything begins, of course, with the raw materials. Headline gets the original Japanese version of the show, plus a version that has been stripped of dialogue, leaving just the footage and the music and effects (or M&E) audio track.

They are also given an English translation of the script, in Genshiken's case done by Duane Johnson. Sometimes the translation of the script is just a raw, literal version, and sometimes it's been given a slight bit of polish by the translator.

In either case, the adapter's job is not just to provide a final polish to the script but to ensure that the words the actors actually have to speak match up with the lip flap in the original footage. All, of course, while staying as faithful as possible to what's actually being said in the original dialog.

There is an old Russian aphorism that a translation is like a mistress: If it's beautiful it's probably not faithful, and if it's faithful it's probably not beautiful. The adapter for Genshiken, Bill Timoney, erred on the side of beauty -- or, rather, on the side of comedy. If it got the point and tone of the original version across, and it was funny in English, then a strict adherence to literalism was of secondary concern. This not only avoids the old dub stereotype of rushed, stilted, Speed Racer-style dialogue, but it still lets the plot and characters come through in English. The specifics may be a bit different between versions, but the overall tone must be the same.

Of course, even after the first adaptation script is finished lines can change, often drastically. The licensor of the anime (the guys paying Headline to adapt and dub the show in the first place) get final approvals before anything gets recorded, and they can make changes to the script. Often this is motivated by translation concerns (sometimes faithful is better, after all). But even after final approval is given, things can still change during the directing and recording process.

Locked to the Lip Flap

An anime dub is carefully assembled from bits and pieces. The director (for Genshiken, also Bill Timoney) tailors each line and vocalism uttered by the actor in the recording booth to the scene, the emotional context, and the aforementioned lip flap.

The actors in Genshiken are all professionals (with the exception of myself in bit parts and Headline producer Joe DiGiorgi, who had an amusing cameo role as, appropriately, an anime director), and know their work. But the demands of voiceover acting are quite different from stage or even TV and movie acting.

Most of the time everything just meshes together in the recording studio. The director sets up the scene, the actor delivers the line and nails it dead-on, the engineer at the big, elaborate Star Trek recording console saves the line to hard disk, and it's on to the next cue. Sometimes it takes a little more tweaking.

Bill would usually give the actor in the booth some basic scene context and emotional direction. ("Okay, you're in a classroom, and the teacher is yelling at you for being a slacker.") Often, everyone would listen to the original Japanese dialogue to see how they handled the emotional tone and content of a scene. Then, if necessary, he'd modify the direction a bit to suit the characters, the tone and the lip flap. ("All right, more rebellious. Wait, try it without the 'you' in that sentence.")

Even so, the actor rarely had to do more than two or three takes of a single line, and not every line had to be redone this way.

Tweaking and Mixing

On a few occasions the engineer also aided in the tweaking process. The timing of a recorded line of a dialogue can be changed a bit, to start sooner or later to better match the footage. A line can also be chopped up and moved around, and a finished line of dialogue can even be composed of parts of multiple takes, though this is not often done since it doesn't always sound right when compared to a single continuously-recorded take.

And then, of course, there are various sound-manipulation tricks the engineer can do to make dialogue sound like an internal monologue, or sound like it was spoken in a giant lead pipe, or even sound like it was recorded by Alvin, Simon and Theodore. The engineer is also able to duplicate and layer voices, turning a handful of actors into a whole crowd full of enthusiastic fanboys, for instance.

Actors would usually record independently, doing their lines alone in the booth. Most of the time the actors never met in the studio at all, even when their respective characters were engaging in an intense back-and- forth dialogue exchange. For an experienced voiceover artist this is not much of an issue, though it seems pretty counterintuitive to a layman. Obviously, it's easier to perform a role when you have something or someone to play against, and as the week of recording progressed, the actors who came in later were able to take advantage of the fact that they had previously-recorded lines to work with.

After that comes the final step, the mix. In the mix stage, lines are further tweaked, shifted in time if necessary, and sometimes further "effected", such as to make dialogue spoken over the radio sound staticky. Then they are placed back with the M&E background.

Sometimes there are panning perspectives (making audio move from one speaker to the next to match the movement of whatever is on screen), and in a 5.1 mix (like Headline's Boogiepop Phantom), some effects elements are shifted from front to rear speakers.

The mixes can be "film perspective", thus having more dynamic range and spatial effect like the audio track of a theatrical film, or they may be more "compressed" and evened out in volume. The latter often plays better on television. The mix is yet another fine line to ride; if the dynamic range is too broad, the soft dialog can be lost to the ears.

The most remarkable thing about the Genshiken dubbing process was sheer speed and professionalism of everyone involved. Lines would be worked out and recorded with an almost rapid-fire pace, and yet the director was willing to work with a line, both in the script and as read by the actor, until it was right. Headline may work fast, but they don't work half-assed. Given the quality of the dubs this studio has produced, this procedure certainly seems to work well for them.

To be continued.