Hardcover
-
SHIP THIS ITEMTemporarily Out of Stock Online
-
PICK UP IN STORECheck Availability at Nearby Stores
Available within 2 business hours
Related collections and offers
Overview
At the core of the book are three original audio projects, drawing from rock & roll, jazz, and traditional African music, through which Savage is able to target areas of contemporary practice that are particularly significant in the cultural evolution of the musical experience. Each audio project includes a studio study providing context for the social and cultural analysis that follows. This work stems from Savage's experience as a professional recording engineer and record producer.
Product Details
| ISBN-13: | 9780472117857 |
|---|---|
| Publisher: | University of Michigan Press |
| Publication date: | 09/26/2011 |
| Series: | Tracking Pop |
| Pages: | 268 |
| Product dimensions: | 6.30(w) x 9.10(h) x 1.10(d) |
About the Author
Read an Excerpt
Bytes and Backbeats
Repurposing Music in the Digital AgeBy STEVE SAVAGE
THE UNIVERSITY OF MICHIGAN PRESS
Copyright © 2011 University of MichiganAll right reserved.
ISBN: 978-0-472-11785-7
Chapter One
Application Study Rock Band
"In Tune and in Time"
One of the first responsibilities that a producer of popular music takes on is the requirement that the final product delivered to the record company be "in tune and in time." That is to say, the musical performances are to realize a certain standard of technical proficiency in pitch and rhythm. The legacy of this central role for a producer may be found in the many rough performances that were a part of the early history of rock and roll. Along with a heavy reliance on attitude came some rather oblique relationships to musicianship on the part of some of the musicians. Thus, especially in the "band era" of the 1960s and 1970s, came the need for some QC (quality control) and the centrality of the producer's role as the arbiter of traditional musical standards. The long-dreamed-of tools for relatively easy pitch and rhythm "fixing" have now arrived along with the DAW. In regards to the direction of technological influence ("top-down" versus "bottom-up"), the realization of this desire suggests that the pitch and rhythm tools of computer-based audio may be seen as a striking example of agency driving technology (we needed to fix stuff, and now we can). However, as we shall see through this volume, there are always elements working in both directions. Perhaps some of what is generally considered to be excessive "fixing" of rhythm and pitch may be considered to be examples of technology driving agency (we can fix stuff, so we do). In any case, the ease and degree of control over pitch and rhythm have dramatically changed, so the new paradigm of music construction is in full bloom when it comes to realizing the producer's dictum that performances must be "in tune and in time."
Prior to the current computer technology the producer had relatively few options in the control of intonation and rhythmic accuracy in musical performances. The primary tool was, after a studio take that wasn't up to the desired standard, to get on the "talk-back" to the performing musician and say something like: "That was great, but can you do it one more time for me: It was a little pitchy" (meaning either too sharp or flat for use) or "It felt a little awkward" (meaning not good enough rhythmically for use). Here the final recorded performances were created through selective repetition. Bits of performances would be captured to allow a complete, musically acceptable performance to be pieced together. If the musician was very capable, then little or no such repetition would be necessary, but in the case of the relatively inexperienced rock-and-roll band member, this could be a long and tedious process. Over time certain techniques and technologies developed that could be applied after the performance, and these aided this process in small ways—tape editing, "flying in," and later judicious use of a harmonizer could correct problems in certain instances—but these options were time-consuming and only successful in a very limited number of circumstances. For the most part getting the required performance out of the musician, sometimes one arduous bit at a time, was the only viable option.
To explore the changes in the application of control over tuning and timing I undertook a project involving a song that I had recorded in 1994 for the band "The Jenny Thing." I had made the original recording using the dominant professional recording technology at the time, which was a 24-track analog tape recorder. This meant I had twenty-four individual tracks for recording on which to build the music for each song. The original sessions were carried out in the typical studio production style of the time. We recorded the initial "basic" tracks of drums, bass, guitar, and vocal together, but all of the performances other than the drum track were considered "scratch" tracks—that is, they were played as guide tracks to be later "scratched" or discarded in favor of new takes of these performances. All of the instruments, as well as each element of the drum set, were recorded on their own individual tracks. All of the instruments and the lead vocal were isolated from each other so that there wasn't "bleed" from one sound into the recording of the other, facilitating the replacement of parts later. Additional parts such as lead guitar tracks, harmony vocals, and percussion tracks were added later. By recording each part at different times I was able to focus the attention and the process of revision on the execution of each individual performance until it was considered acceptable. This was a relatively low-budget record, so the standard for "acceptable" had to take into consideration the capabilities of the musician along with the overall time that the budget allowed for the entire recording and mixing process. This remains the dominant procedure in pop music today, though the weight of performance control has shifted from being almost completely a part of the original recording to a balance between recording and the kind of postrecording manipulation that I was now going to apply to this production. My goal in this study is to apply to this older recording the process of rhythm and pitch "fixing" currently used in pop production, allowing me to compare the original master as it was released on record to what would probably comprise the master recording if this song were produced using contemporary technology.
My first task was to transfer from the analog tape to digital audio in the DAW. For the software needed to control the digital audio stored on the computer's hard drive I used today's dominant professional recording software, Avid's Pro Tools. While making the transfer and listening to the original audio I noted a slight tempo fluctuation during the song's introduction. I remembered being continually aggravated by this when I produced the original track. This was a case where I had deemed the inconsistency to be slight enough to be acceptable, though it was significant enough to have bothered me throughout the process. I smiled to myself knowing that now I would be able to "fix" this slight problem, and then thought of the countless number of examples of such occurrences in other recordings made before the current capabilities were available. How many slight problems in recorded performances have haunted musicians and producers before there was a means of correcting them as a part of the production process? But before I explore the meaning of such musical "fixing," I provide a narrative of the process I undertook in applying contemporary production practices to this particular piece of music.
First, a caveat: I use the terms fix, correct, consistent, and the like as technical terms, while recognizing that these also carry significant implications about the value of the alterations being made. The reality is much more complex, for value in musical performance is most often ascribed to deviations from the standard to which we are "fixing." The implied values of such words as fixing are not necessarily a reflection of how one might value the actual effect of this process. In fact they may be completely at odds with such implications (supposedly "fixed" performances may be considered inferior to the original). I will be addressing questions of value in this more general sense later in this chapter, but for the moment I ask the reader to temporarily indulge the use of the language for the sake of the narrative.
Fixing in Pro Tools
As with most popular music, this song was recorded to a click track generated by a metronome. That is to say, the drummer listened to a click when laying his initial track while the other musicians played to the (click-informed) drummer's performance. Using contemporary production techniques where we would be recording directly into the computer, with the click generated by the computer, the bar and beat information would already be an integral part of the recording and established before any music was played. With this historical recording I had an individual audio track with the click from the metronome recorded separately, but this was not integrated into the computer clocking function—Pro Tools wasn't able to give a readout of the metronomic bar and beat information. In fact, because it was recorded using analog gear, the original metronomic timing was not perfect. Slight variations in the creation of the click by any analog clocking device (metronome), combined with minute variations in speed from an analog tape recorder, mean that it is not possible to simply assign the correct bpm (beats per minute) reading to the audio now in Pro Tools and have the music line up correctly with the bar and beat information. Fortunately there are tools to assist us in adjusting for these inconsistencies so that we can work within the traditional music organization of bars and beats along with the clocking precision of a computer.
In the Pro Tools program there is a plug-in tool called Beat Detective. As the name suggests, this tool investigates rhythmic qualities of audio data. It distinguishes beat information by identifying transients (high-frequency leading sound elements) that are likely candidates for marking the beginning of each beat. In this case, because I had the click track recorded on a separate audio track, it was an easy matter for Beat Detective to create a tempo map from the position of each click and thus organize the file into bars and beats. To do this Beat Detective assigns an exact tempo for each beat, to within three decimal points of bpms, thus yielding a bar and beat map that remains consistent with the original click. Beat Detective does not alter the placement of the beats, but it identifies and organizes them in a way that makes them conform to a bar and beat structure. By doing this I had a tempo map that represented the "ideal time" when the performances were made. This was the "correct" beat structure that the drummer was conforming to when laying the initial drum track.
I then used another feature of Beat Detective to slice all of the various drum tracks into separate regions, setting a variety of parameters to help it make "intelligent" decisions about how to read the transients and divide the performance into various beat-related elements. As is typical in contemporary drum set recording, there were individual tracks for bass drum, snare drum, tom-toms, and hi-hat cymbals, as well as separate stereo recordings of overhead microphones to capture the cymbals, and room mics to capture the overall sound of the drums in the room. Beat Detective processed each track separately. Using Beat Detective on overhead and room tracks is difficult because of the complexity of the information. As sophisticated as Beat Detective is, it has trouble determining beat divisions when the audio consists of all of the drum instruments mixed together.
Once Beat Detective had created individual regions from each track of the drum performance, I used the "quantize" function to correct the timing of the drum performance. Quantizing takes the beginning of each separated region of audio and moves it along the musical timeline to the beginning of the nearest user-defined beat subdivision (in this case the smallest subdivision was eighth-note triplets, as this song used a "shuffle" or triplet subdivision of the beat). Quantizing each track individually yielded the most accurate results, but it also meant that where there were inconsistencies in the quantizing process between individual tracks I would have to make manual changes for the parts to conform to each other. While this process created a much more accurate version of the original drum performance in terms of note placement relative to the "ideal" of metronomic time, there was still considerable variation in the volume and timbre of each individual sound, as well as internal variations within the larger segments that were quantized into position. Thus the resulting performance was not the same as a performance coming from a drum machine, where every note may be metronomically placed and there is generally little or no variation in dynamics or timbre. Quantizing these tracks took about two hours, but this is remarkably efficient considering the literally thousands of edits, adjustments of beat placements and extension of regions to close gaps, creation of crossfades in order to smooth transitions, and about twenty manual adjustments at places where the automated process produced slightly anomalous results.
In working with the drum track I made several other typical alterations to the files in order to create cleaner and more consistent performances. There were a few weak or bad-sounding bass drum or snare drum hits that had come from inconsistent striking of the drum, and I replaced those with better-sounding hits using a basic cut-and-paste function. There is also a tool in Pro Tools called Strip Silence that allows one to create silence below a user-definable amplitude threshold. In this way it is possible to quickly eliminate leakage sound from adjacent instruments, and for drums this can create a much cleaner overall sound. For example, the tom-tom tracks had substantial off-axis leakage from the snare drum, bass drum, and hi-hat. By stripping away all parts of the audio file other than the actual tom-tom hits I could remove the clouding effect of this leaked audio. Strip Silence provides a very efficient means of eliminating these off-axis sounds.
After "fixing" the drum part I proceeded to work on the timing of the bass guitar part. With the first two-thirds of the performance I was able to capture and separate regions into beat-oriented sections using Beat Detective. I then quantized to eighth-note triplets and smoothed the transitions using the automated fill and crossfade function. This extends audio regions where necessary to fill in the gaps created by moving the regions to their corrected beat placement. It then creates a short crossfade between adjacent audio regions to create smooth transitions. In a few places I used the copy-and-paste function to replace a poorly played part with the same part from a different section of the song. The last one-third of the bass part is continuous legato triplets with little dynamics and enough sustain to make it impossible for the computer to discern the break points needed to create the individual beats. This portion of the music consisted of a repeated two-bar pattern, so I found the best iteration, massaged it into shape, making slight adjustments to both rhythm and dynamics, and then pasted that "fixed" two-bar phrase throughout. I had to requantize each two-bar section to the appropriate downbeat because of the slight tempo fluctuations of the click. There was one musical variation at one transition point (the same pattern played up an octave) and I left that from the original performance. Then I used the automated smoothing function to close gaps and create crossfades on this final section. The result of this work was a more rhythmically stable bass track that sounded more accurately played to the drum track.
The Wonderful World of Auto-Tune
Having started to work with audio that contained pitch information—in this case the bass guitar track—it was now time to employ some pitch correction. The most frequently used tool for pitch correction in the digital domain is a plug-in called Auto-Tune. Auto-Tune and its various successors employ pitch detection algorithms that are capable of reading very small variations in pitch in real time (single-voice only, it cannot read multiple notes played at the same time). Once the software has determined the continuous pitch information for a segment of audio, it creates a graphic representation of that pitch on a grid where the vertical axis is pitch and the horizontal axis is time. The user can then redraw the pitch representation on the graph to alter the pitch. When the original audio is played back through the plug-in, it adjusts the pitch to the redrawn graphic information. In this way variations in pitch deemed incorrect can be "corrected" in exactly the way the user desires. This may be gentle correction to move pitch variations closer to the actual note, or aggressive changes that lock the pitch to the desired note. There is also an automatic mode that corrects pitch in real time as the audio is fed through the plug-in. In this mode the audio is gently moved toward whatever note the original audio is closest to, though you can dictate which notes are "valid" by indicating scale function or even by designating your own custom scale.
(Continues...)
Excerpted from Bytes and Backbeats by STEVE SAVAGE Copyright © 2011 by University of Michigan . Excerpted by permission of THE UNIVERSITY OF MICHIGAN PRESS. All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
Table of Contents
Contents
List of Audio Clips....................xvIntroduction: Reproduction and New Paradigms....................1
part one: Repurposing Presentation....................21
Introduction to Part I....................21
one. Application Study: Rock Band....................25
two. Studio Study: Lipsmacks, Mouth Noises, and Heavy Breathing....................48
three. Art or Artifice?....................61
part two: Repurposing Performance....................79
Introduction to Part II....................79
four. Application Study: Jazz Piano Trio....................81
five. Studio Study: Capturing the Unintentional Performance....................99
six. Artist or Artisan?....................107
part three: Repurposing Participation....................127
Introduction to Part III....................127
seven. Application Study: African Folklore and Music Communities....................129
eight. Studio Study: From iPod to GarageBand....................150
nine. Integration or (Dis)integration?....................174
Conclusions: Reflections on the Future....................193
Notes....................197
References....................219
Index....................229