Do-It-Yourself Mastering

Mastering is essentially the art of applying signal processing to finished mixes in order to enhance them or correct perceived problems, sequencing the
Image placeholder title

First Name: Last Name: Address: City: State: Zip: Email:

Mastering is essentially the art of applying signal processing to finished mixes in order to enhance them or correct perceived problems, sequencing the processed tracks in the desired order for playback, and exporting the resulting file in the correct format for delivery to the replication plant for mass production. Although mastering is a highly specialized process, so many affordable mastering-oriented products have been introduced to the market over the past several years that almost anyone can now master their own recording project. However,candoesn't always meanshould.

Image placeholder title

Many engineers look to mastering as a solution for fixing a mix that has problems they can hear but can't quite figure out. For example, they might feel their mix lacks punch or is fatiguing to listen to but don't know how to make it sound otherwise. In this case, indiscriminately piling on additional processing during the mastering stage won't solve anything and could, in fact, make the finished masters sound a lot worse. Here it makes a lot of sense to hire a great mastering engineer to add fresh ears and skill sets to the project.

If, on the other hand, your mix has problems that you can readily identify and know how to fix, it makes more sense to remix than it does to try to correct a bad mix in the mastering stage. Because processing applied during mastering typically affects all elements of a mix, correcting one thing might make another thing sound worse. For instance, boosting EQ in the mastering stage to make a dull-sounding snare drum sound brighter might make vocals too sibilant or the cymbals sound harsh. (There are other ways to fix such a problem, which I'll discuss later.) In this case, it would be best to remix the song and apply needed EQ only to the snare drum track before mastering the project.

All that said, mastering can work miracles on poorly wrought material that cannot be remixed for some reason, such as when the multitrack master is lost or damaged or exists only in an obsolete format for which there is no playback mechanism. In that case, the 2-track mix may be all that is available to work with, and mastering is the only recourse to improve the program.

Mastering yields the best results when used to make great mixes sound even better. It can also make a program flow better from beginning to end by lending more consistent spectral balance and dynamics to all the mixes on a project so that no jarring changes occur (unless intended) from song to song. Also implicit in the mastering process is preparation of the premaster (the file or disc used for replication), which involves sequencing of songs (with gaps and track offsets), documentation (including CD-Text), preparation of reference discs (for auditioning and final approval), and delivery of the master in an error-free format that the replication house can handle.

In this article, I'll discuss what's involved in mastering your own stereo project for CD release. I'll begin with a brief overview of requirements for accurate room response and monitoring setup. Then I'll dive into how to optimize your work flow and evaluate what types of processing may be needed for your program material. Along the way, I'll give a small sampling of some of the mastering products currently on the market. We'll finish with a discussion of preparation of the premaster.

The main focus will be on do-it-yourself projects and not on running a commercial mastering studio. (See the sidebar “Do unto Others” for a brief overview of things to consider when mastering other people's projects.) For the sake of simplicity, I'll talk about mastering an album project inside a DAW, but most of what I'll discuss also applies to other mastering scenarios.

Go to Your Room

It is absolutely vital that your mastering work be performed in a room that has highly accurate frequency, phase, and reverb responses. Trying to master your project in an inaccurate room makes as much sense as doing color photo touch-ups while wearing tinted sunglasses. Likewise, unless your monitoring setup is flat and can accurately reproduce the entire audible frequency spectrum, how can you confidently decide which frequencies need adjusting on your program? If you can't truly hear what is going on with your source material, you're just shooting in the dark.

While a full-blown examination of room acoustics, control room equalization, and monitoring setups is way beyond the scope of this article, a few points bear mentioning here. (For an in-depth look at tuning your control room, see “Truth or Consequences” in the November 2001 issue of EM, available online at First, learn what room modes (those narrow peaks or dips in frequency response) your room exhibits and keep them in mind while mastering. Before boosting or cutting EQ at or near any room-mode frequencies, listen to the same program from a position in the room where those same modes are not being reinforced to determine if corrective EQ is really needed. Mastering can be effectively performed in a room that has only minor imperfections in response if you know what those imperfections are and compensate for them.

Room tone is another matter. You should make sure the RT60 (essentially the reverb decay time) is not skewed in any one frequency band in your room. Otherwise, you may end up cutting bass frequencies, for example, simply because your room reverberates longer in that band than elsewhere throughout the spectrum and not because there is excess bottom end in your mix.

Get with the Program

You should have at least two pairs of monitors on which to evaluate the program. One pair should be full range (extending at least down to 35 Hz or so) or be paired with a subwoofer to extend bass response. The second pair should ideally be band limited (that is, bass deficient) to provide you with an idea of how the finished product will sound when played back on small consumer systems.

Full-range monitors suitable for use in mastering are too numerous to list here. But strangely, there are relatively few high-quality band-limited models on the market, so I can make some recommendations: the Yamaha HS50M and NS-10M Studio (the latter model is discontinued) and the Avant Electronics Avantone MixCubes are the best I've heard for this purpose. Most other tiny close-field monitors I've auditioned attempt to sound like a big speaker in a small box and have way too much bass response to serve as a proxy for small consumer-playback systems (not to mention that their bass response usually becomes highly inaccurate when placed on workstation shelves or a console meter bridge).

Both pairs of monitors (full range and consumer proxy) and the subwoofer should be wired to a switch box (or patched to separate control-room outputs on your mixer) so that you can select playback on either pair — alternately with and without the subwoofer engaged — at the push of one or two buttons. Depending on the capabilities of your gear, you may need to wire up a custom setup to allow simultaneous playback on one pair of satellites and sub. By switching among different playback references while working, you'll be able to tell if your mastering changes will sound good across a variety of audiophile and low-end consumer-playback systems.

Equally important to having high-quality monitors is knowing their inherent frequency limits (bass and high-frequency rolloffs). For example, I know my NS-10M Studio monitors don't reproduce bass frequencies below 60 Hz very well. If I hear boomy bass guitar frequencies on my full-range monitors but the boominess disappears when I switch to my NS-10M Studios, I know the problematic frequencies lie below 60 Hz. And if I listen to playback on my subwoofer only (with satellites muted) and a boomy acoustic guitar track completely disappears, I know the boomy frequencies lie above the 110 Hz high-frequency cutoff of my sub.

During your mastering sessions, you'll occasionally and very briefly want to check your work at a loud volume to make sure extreme low and high frequencies are in proper balance with the rest of the spectrum. (Compared with the midrange band, the human ear is less sensitive to the extreme frequency ranges at low listening levels.) But avoid listening fatigue by resisting the urge to listen at loud levels for more than just a few minutes spread out over each day. And as the day progresses, take more frequent breaks to rest your ears and gain a fresh perspective.


At the beginning of the mastering session, your first task is to import all your mixes into your digital audio workstation (unless they are on tape, in which case any desired analog processing should be applied before you bring those tracks into your DAW). This is usually accomplished by simply dragging-and-dropping each file into the appropriate folder or window in your DAW and, from there, into a blank track. If at all possible, you should be working with 24-bit files for your mixes.

After importing all the audio files for an album project, I will listen through virtually the entire project and make notes for each song as to what types of signal processing may be needed to correct problems or further enhance what's already good. (I'll discuss what some of those treatments might be in a bit.) This gives me a plan of attack for each song while my ears are their very freshest. It also reveals at a glance whether the same problems repeatedly crop up in most or all of the mixes. For example, I often hear unnecessary boosts or cuts in the same bass band for every song when a project was mixed in an inaccurate control room. In such a situation, the engineer compensated for a problem with their mixes that simply didn't exist, and part of the mastering process is to undo that equalization and restore spectral balance.

As I'm listening through and evaluating the project's mastering needs, I'll place markers at the beginning of each song. That will later allow me to quickly jump from songs for which I've already rendered signal processing and level changes to the one I'm currently working on, to check for consistency or artistic compatibility of spectral balance, dynamic range, and so on.

I will also place markers where at least one significant peak for each song occurs. This allows me to make comparisons between songs at those points to ensure that no song is much louder or softer, or more dynamic or compressed, than the others in the finished product. More >>>

Fill in the form below and click Order Now! to get two years (26 issues) for just $23.97 - the regular price of one year. But HURRY - this offer won't last forever! (U.S. orders only please)

This data will be sent directly to Electronic Musician Magazine and will not be used for any other purposes.

Avoiding Blind Alleys

Before we dive into signal processing, it's important to set up a good monitoring scheme to facilitate work flow and maintain critical perspective throughout the session. During the course of the session, I may elect to add analog or digital processing or both. I'll want to refer back often to how the audio file sounded before I started messing with it, just in case I'm making things sound worse instead of better. To facilitate these comparisons, I'll set up multiple monitor-source points for listening to the song I'm currently working on at various stages of processing.

Image placeholder title

FIG. 1: A multichannel mastering setup in MOTU Digital Performer 5. Each track is routed to a different stereo digital output feeding the same downstream monitoring path so the engineer can monitor each one separately in turn as they all play back. Note the markers placed at the beginning and in peak sections (the latter denoted by the word loud) of each file.

One monitor-source point will be for the original audio files playing back in the track they were first dragged into, with no processing added. A second monitor-source point will be for any of the same audio files that have already been processed in the analog domain and “printed” (recorded with the analog processing rendered to new files) to a new track (see Fig. 1). Each of these two tracks gets routed to a separate stereo digital input on my mixer feeding the exact same downstream monitoring path. I also set up a third monitor-source point (feeding another stereo digital input on my mixer) for the audio file currently being processed in the analog domain (monitored post A/D converter so that I'm making an apples-to-apples comparison with already-printed files). With this setup, I can compare — by selecting with a simple button push each monitor-source point in turn — the sound of the audio file for which I'm currently tweaking analog signal processing settings to both the sound of the original, unprocessed file and the sound of the files for other songs that have already had their analog signal processing imparted.

I now have two tracks in my DAW, one containing the original, unprocessed audio files for all songs and another containing the same songs with analog processing added (if they needed it). Typically, if one song in the batch needed some sort of analog processing, all others recorded and mixed in the same studio will need it too, though often to varying degrees. However, if I elect to forgo analog processing on one or more mixes, I can simply copy and paste their original audio files over to the track containing mixes that have had analog processing applied so that all songs at this stage of the game are contained in the same track. I keep notes along the way as to which mixes had analog processing applied and which didn't, in case I need to go back and tweak prior settings. And I either document for each mix the exact settings for any analog gear used or, if possible, save those settings as a custom preset in gear that offers digital storage and recall.

Once analog processing has been rendered for any song files that needed it, I'll trim the head and tail of each processed mix to remove any unnecessary pre- and postroll. Then I'll bounce those files to a third track in my DAW while adding any needed digital processing (which, in some cases, might be limited to only dither).

Digital processing may be provided by plug-ins, outboard devices (such as digital compressors and equalizers), or both. As I add digital processing, I'll once again use multiple monitoring paths so I can compare my work in progress to the sound of mixes already printed with digital processing and also to the same mixes with only their analog processing (if any) applied.

Image placeholder title

FIG. 2: In the rightmost column of the List view for BIAS Peak Pro''s Playlist, Vbox plug-ins can be turned on and off as a group with one mouse-click to facilitate A/B comparisons between dry and processed mixes.

I'll also store the settings used for digital outboard processors and plug-ins as custom presets named after the song they were used on. This makes it easy for me to revisit those settings in case I need to later rework the processing for any particular song. When mastering in MOTU Digital Performer, as I regularly do, I'll save the settings for all the plug-ins used on a particular song as a group in a Clipping; that allows me to later restore the whole kit and caboodle with one simple drag-and-drop operation into Digital Performer's mixer. BIAS Peak Pro users can do essentially the same thing by saving and recalling a Vbox matrix of plug-ins for each Playlist Event; this has the advantage that a Vbox and all the plug-ins it contains can be bypassed or activated with one mouse-click for A/B comparisons between original and processed mixes (see Fig. 2).

The above process has been oversimplified for the sake of clarity. In reality, I may sometimes elect to add corrective digital processing (to fix obvious problems with the current mix) before the D/A conversion on the first pass through analog processing if it's obvious I'm going to need it, and if failing to do so would make it harder for me to hear the best settings for the analog gear. Additionally, I prefer to bounce my fully mastered mixes (those rendered with any needed analog or digital processing) to disk and then import the finished file into the third track mentioned earlier, rather than bouncing directly to the track, as the results sound a little more transparent to my ear when working in Digital Performer.

It must be said that the multichannel paradigm I've presented here isn't the only way to master. If you're mastering in a 2-bus application such as Peak, for instance, you can listen over the same output bus to multiple documents in turn (each representing a different audio file, such as processed and unprocessed versions of the same mix) simply by pressing a number key on your computer keyboard. It all comes down to how you want to work. The point is that it's important to set up a monitoring scheme that allows quick A/B/C comparisons.

Put It in Gear

When building your arsenal of mastering gear, it's crucial to know what makes a product specifically tailored to this application. The most distinguishing and vital aspect — beyond superior sound quality, which is a given for mastering applications — is repeatability. Mastering gear must provide a way to recall your exact settings so that you can try different things and always return to prior settings if you take the wrong fork in the road.

Image placeholder title

FIG. 3: URS is one of many manufacturers whose parametric EQ plug-ins can adjust boost and cut settings in exacting 0.1 dB steps, suggesting their use in mastering applications. The 7-band URS S Mix EQ plug-in is shown here.

This repeatability can be attained either through detented or switched control settings (which you must remember in the short term and document for the long term) or by including facilities for digital storage and recall (as do, for example, digitally controlled analog processors). Obviously, DAW plug-ins easily meet this criterion, as all their settings can be stored in user presets. An increasing number of plug-ins are going one step further by providing two or more work spaces (alternate A and B control setups), allowing the user to try a couple of different tacks and toggle back and forth between them to see which sounds best.

Another requirement for mastering equipment is that all controls have very small incremental steps. For instance, equalizers should ideally have no greater than 0.5 dB steps between adjacent boost or cut settings if they are to be used in mastering. The most important rule of mastering is “First, do no harm,” and handling a project with kid gloves requires that the gear used be capable of making very subtle adjustments.

Analog Demagogue

Despite having many suitable plug-ins at my disposal, I often send mixes that have been created entirely inside a DAW out to the analog domain for processing. I've yet to hear plug-ins that can produce the warm, nuanced sound of, for example, a Millennia NSEQ-2 parametric equalizer or Tube Tech SMC 2BM tube multiband compressor. When cold, digital tracks need warming up, some mastering engineers record DAW mixes to an analog half-track machine before they even touch them with EQ or compressors.

Image placeholder title

FIG. 4: Almost any equalization curve imaginable can be drawn using a mouse in the RND Frequal-izer plug-in''s frequency graph. An overzealous mastering engineer got fired after applying this particular curve to the client''s mix.

That said, digital processing — both hardware- and software-based — has a lot to offer. For instance, if you're looking for a convincing emulation of tube and tape saturation without leaving the digital domain, the Crane Song HEDD 192 (a hardware device) has found favor with mastering engineers. And there are a number of plug-ins and bundles that excel at mastering, not just because they sound great, but also because they have capabilities their analog counterparts can only dream of.

For example, most parametric equalizer plug-ins can adjust gain boost or cut in surgical 0.1 dB steps (see Fig. 3). And with the Roger Nichols Digital (RND) Frequal-izer plug-in, you can draw any equalization curve you want with your mouse (see Fig. 4). Try doing that with an analog equalizer.

Image placeholder title

FIG. 5: iZotope Ozone 3 incorporates six mastering processors that alternately share the same window. The graphic display for the Paragraphic Equalizer module is shown here.

There's also your budget to consider — a high-end analog mastering equalizer or compressor might cost you several thousand dollars, whereas a bundle of mastering plug-ins such as the excellent-sounding iZotope Ozone 3 will set you back less than $250 (see Fig. 5). In the end, both analog and digital processors can yield excellent results; it's what you do with them that counts. That's what we'll examine next.

In the Trenches

Let's take a look at some examples of what you should listen for when mastering and how to evaluate what types of processing might be helpful to make the changes you want to hear. Of course, an in-depth discussion of signal processing techniques is beyond the scope of this article, but the following mastering tips will give you some useful ideas to pursue further.

First, switching among your different pairs of monitors during playback should give you a good idea of where the spectral balance may be skewed. For instance, if the current mix lacks presence on your full-range monitors but sounds balanced on your band-limited speakers, the problem probably isn't a lack of upper midrange frequencies but too much bottom end.

Image placeholder title

FIG. 6: Here, the 7-band PSP MasterQ linear-phase equalization plug-in is being used on only the left channel of a stereo mix. MasterQ also includes various limiting and saturation algorithms, which may be disabled.

As you evaluate what equalization may be needed to restore the proper balance, keep in mind that the problem frequencies may be affecting only one side of the mix. For example, a guitar part with excess energy in the upper-bass band may be panned more or less to the left side of the stereo field, in which case cutting in that band on both channels would cause the right channel to sound thin. The PSP MasterQ equalizer plug-in is highly useful here, as it allows you to equalize only one side of a stereo file while leaving the other side untouched (see Fig. 6).

The RND Uniquel-izer and Frequal-izer plug-ins both go one better, allowing you to apply different EQ treatments to the left and right sides of a mix simultaneously. With Uniquel-izer, you can add and delete as many bands of EQ as your CPU can handle, choosing from 11 different filter types (see Fig. 7).

Image placeholder title

FIG. 7: As long as your CPU can handle it, you can add as many bands of equalization as you want in the RND Uniquel-izer plug-in. Choose from 11 different filter types for each band.

As you work, listen also to the current mix's dynamics. Does the mix sound full enough? If not, some light stereo-linked compression with moderately fast attack and release times and a low ratio should add some pleasing “glue” to the performance. Just be careful not to go overboard, or depth will go out the window and your mix will start to sound two-dimensional, squashed, and lifeless.

Perhaps guitars and keys are blanketing the trap drums too much in your mix. Instead of EQ'ing the traps so they sound brighter, try applying stereo-linked compression with slow attack and fast release times. That should make percussive elements such as kick and snare drums “pop” more.

Boom Town

Say you've got a mix in which the acoustic guitar gets boomy every time it plays the D string, but it sounds fine at all other times. (Make sure this is really the case and isn't just a room mode blossoming at that frequency.) A static equalization cut around 160 to 200 Hz will tame fundamental frequencies for that D string (if fretted below the sixth fret), but it will also thin out all other guitar passages, as well as the entire mix, when that string isn't played. A dynamic approach is called for here.

Image placeholder title

FIG. 8: This screen shot shows the Waves Linear Phase Multiband plug-in being used to reduce boominess in the 79 to 281 Hz band. The other bands are bypassed, leaving their frequencies untreated.

You can use a split-band compressor such as the Waves Linear Phase Multiband (part of the Waves Masters bundle) or iZotope Ozone 3 Multiband Dynamics plug-ins to put a lid on the 160 to 200 Hz band so that its energy is dynamically cut every time the boominess surpasses the threshold set for that band. (You might actually need to set the bandwidth wider to also tame boomy formants; see Fig. 8.) Both of these plug-ins can also execute upward expansion (sometimes called bootstrap compression) that will make quiet song passages such as a solo guitar intro louder without compressing the top end of the dynamic range and squashing peaks.

Of course, squashing peaks is where it's at if you want your mixes to be competitively loud. Plug-ins such as Universal Audio UAD Precision Limiter (see Fig. 9), iZotope Ozone 3 Loudness Maximizer, and Waves L3 Multimaximizer and L2 and L1 Ultramaximizer can really pump up the volume of your mixes by clamping down on peaks and bringing up the average level. The danger is in going too far, killing any punchiness and making your project sound like a fatiguing onslaught nobody can stand to listen to for more than five or ten minutes at a time. In fact, for some musical styles, such as classical, using any amount of limiting or maximizer processing would be inappropriate. If used, maximizing and dithering (which I'll discuss momentarily) should be the very last processes you apply to your mixes.

Image placeholder title

FIG. 9: Universal Audio UAD Precision Limiter is an excellent mastering tool for cranking up the volume and power of your mixes.

Waves L3 Multimaximizer is an especially powerful mastering limiter because it can condition different frequency bands to alter the limiter's response. With L3, you can subtly favor average levels in one frequency band and peak levels in another by changing the plug-in's Gain and Priority settings (see Fig. 10). For example, you can increase the energy in the bass band and enhance the upper-midrange-frequency component of snare drum hits while keeping a firm lid on other bands. L3 includes the Waves IDR word-length-reduction algorithm, a highly transparent quantization and dithering process useful for rendering your files to 16-bit format for CD release.

Image placeholder title

FIG. 10: Users can condition the response of the Waves L3 Multimaximizer''s limiter in five different frequency bands. With the settings shown, the average level of a cello section and, to a lesser degree, the crack of snare drum hits are enhanced in a mix while other elements are more tightly controlled.

There are many other types of signal processors available for use in mastering, including harmonic exciters, reverb, and stereo imagers. The Waves S1 and iZotope Ozone 3 Multiband Stereo Imaging plug-ins can alter the perceived width of your mix, but not without making it sound more diffuse (which may or may not be appropriate). Mid-side (M-S) processing is an advanced mastering tool that can also be used to adjust your mix's width. By encoding a stereo mix into an M-S matrix, audio common to both channels such as center-panned tracks (the mid component) and audio that's exclusive to both channels (side) can be separated and independently equalized and compressed. For example, you can use M-S processing to beef up the kick and snare drums while raising the level of an uncompressed stereo string pad to widen the mix. (For more information on mid-side processing, see “Front and Center” in the March 2006 issue of EM, available at

As mentioned earlier, before you maximize and dither your files to 16 bits for CD release, you should trim the head and tail of each file to remove unnecessary noise at the start and end of each mix. Then create fades at the beginning and end of each trimmed file to avoid any pops or clicks that would otherwise result from potential DC offset or low-level noise instantaneously slewing up or down at butt-splice margins. Finally, after any last-minute fader adjustments or other gain changes are made so that perceived loudness flows smoothly from song to song throughout the entire program, maximize and dither all the files to 16 bits. To preserve the depth of your mixes, make sure you apply no further signal processing — including gain changes — after you dither your files. And avoid dithering more than once if possible, in order to prevent potentially audible artifacts from polluting your files.

Premaster Prep

After all of your mixes have been trimmed, faded, processed as needed, and dithered to 16 bits for CD replication, import them into whatever software application you've chosen to prepare your premaster. (The premaster is the disc or file from which a glass master is made at the replication house in preparation for mass-producing your CD.) If you've been using, for example, BIAS Peak Pro XT to master, your files are already ready for premastering in Peak's Playlist and don't need to be imported. (I'll talk about Peak and other premastering solutions in greater detail in a bit.)

Due to space constraints, I can't completely discuss premastering, but I'll hit on some of the major points. First of all, resist the temptation to make any further gain changes once signal processing has been rendered and your mixes have been dithered; doing so would degrade the quality of the audio portion of your CD.

Your premastering software should automatically assign a different CD-track number to each audio file you import, but you can rearrange the song order and their track numbers if you want. After the song sequence is set the way you like, set the duration of any gaps (silent portions) you want to have between CD tracks. One CD track is typically composed of only one song but may be fashioned to include a medley of several songs or a segue.

Image placeholder title

FIG. 11: Track offsets should be set for your premaster to prevent CD players from muting audio at the beginning or end of each CD track. In this example, global track offsets are entered in the Delivery window (Windows‘Preferences‘Delivery) for Sonic Studio PreMaster CD, a premastering application.

CD players take a fraction of a second to fill their play buffers with the current track's audio, during which time the player's output is muted. So if you keep your CD track's default start time, or track index, positioned at the onset of the audio program, playback of the very beginning of the track may be cut off. The solution is to program track start offsets, which are silent gaps that occur between each track index and the beginning of audio for the respective CD track (see Fig. 11). The first CD track typically requires a longer start offset than that for all the other tracks on the CD. It's also important to add track end offsets to prevent the end of a CD track from getting muted before its audio program fades completely out. (I've been told by a reliable source that this is an issue only when a CD player is in shuffle or random-play mode.)

Start and end offsets are usually programmed globally so that they are the same duration for every track after track 1 (which we noted gets its own special offset applied). That said, Peak can set different start offsets for each song (in the application's Playlist), which is useful when you want to arrange a specific start point during a crossfade between two tracks. As consumer CD players don't all mute their outputs for the same amount of time while filling their play buffers, play it safe and set the premaster's global offsets to be a tad longer than you think you might need. Safe numbers are 75 CD frames for the first track, 25 CD frames for track start offsets, and 15 CD frames for track end offsets. An offset of 75 CD frames is equal to a 1-second duration, so each 25-frame track start offset is ⅓ second in duration and each 15-frame track end offset is 1/5 second long. When calculating how long the silence will last in between each song on your CD, it's important to add the track start and end offsets to the gap time you programmed, to arrive at the real silent-gap length.

If your computer's operating system and drive can write CD-Text data (you'll need Mac OS X 10.4.3 or later if you're using a Mac), you may want to enter this information, which includes album and song titles, the names of the performer or band and songwriter, and ISRC data. (Go to for more information on obtaining ISRC codes for your CD tracks.) Once all text data is entered, use your software to print out your premaster's contents so that you can give a text readout of all PQ codes (which generate the disc's table of contents, or TOC), track offsets, and other data to the replicator.

If your software enables it and your replicator can accept it, export a DDP (Disc Description Protocol) file set and burn it to a CD-ROM. DDP file sets contain error-protected audio data and all the metadata (or “data about the data” on your premaster) needed by the pressing plant. DDP file sets are a superior delivery format compared with Red Book CD-DA discs, which are prone to errors, but not all replicators can accept them.

Learn to Burn

Several premastering programs are currently on the market, some of which are incorporated into mastering software. Here is a quick look at what's available.

With a $200 list price, Roxio Jam 6 is an inexpensive disc-burning application with a limited premastering feature set and no plug-in support or DDP-export capability. BIAS Peak LE ($99) includes audio file editing, plug-in support, and disc burning, but it can't export DDP file sets either. BIAS Peak Pro 5 ($599) offers plug-in support and advanced editing, mastering, and premastering features, including, with an optional extension costing $399, DDP file set export. BIAS Peak Pro XT 5 ($1,199) adds the company's Master Perfection Suite plug-in bundle and SoundSoap and SoundSoap Pro restoration software to Peak Pro 5's feature set and can also utilize the company's optional DDP extension.

Sonic Studio PreMaster CD ($475 for the download) offers intuitive editing and premastering capabilities but no plug-in support. It is the least expensive solution for DDP file export. However, PreMaster CD also forces you to set the same gap between every song, while even the rock-bottom-priced Peak LE and Jam 6 allow the gaps between songs to be set to different lengths. The 4-track, 4-bus Sonic Studio soundBlade (native, $1,495; with accelerated DSP, $3,995) does everything PreMaster CD can do but also adds powerful plug-in architecture and advanced editing, mastering, and premastering capabilities. SoundBlade offers expansion options for applications such as restoration as well.

Job Export

There's a lot more to mastering than simply slapping a maximizer on your project and calling it good. (In fact, we've just scratched the surface in this article.) And while many studios offer mastering services these days, there is a wide divergence in know-how and quality between the best and worst of the crop. If you can only afford the mastering services of a bottom-rung studio, you may find that you can obtain better results by doing the work yourself. With an accurate room and monitors, quality mastering gear, good ears, and technical chops, there's no reason not to give it a shot.

EM contributing editor Michael Cooper is the owner of Michael Cooper Recording in beautiful Sisters, Oregon. Examples of his mastering work are posted


If you're considering opening a commercial mastering studio, one of the first decisions you'll need to make is which formats to support. The best-equipped facilities support a wide variety of analog and digital formats, both for compatibility with clients' masters and, in some cases, for use as signal processors in their own right (as is the case when transferring program material to certain analog tape machines). The commercial operator must weigh the expense of supporting multiple formats against what the market requires and competitors offer.

Mastering other people's projects also demands that extra weight be given to considerations beyond your own creative leanings. First of all, it's important to ask your clients what their likes and dislikes are. That huge bottom end on their mix might be an annoyance to them — something that needs fixing — and not an intentional production value. How do they feel about competitive loudness? Do they want their record to be louder than anyone else's at any cost to sound quality? And if it's a band's project that you're mastering, which member(s) call the creative shots?

Before you begin working on a project, give a quick listen to short sections of a few tracks and offer your opinion of what needs to be done to fix problems (such as a murky midrange masking vocals and guitars) and also what can but doesn't necessarily have to be done in the realm of creative enhancement (such as making the drums slam more or widening the stereo image). Write down your client's responses and direction and use those notes throughout the mastering process as your homing beacon to keep true to the artist's vision for the project.

The need to work fast must be weighed against all potential engineering considerations. Unless you're working for experienced artists or companies, only a very small portion (if any) of the project's overall budget will likely be allocated toward mastering. Whether you're working for an hourly or per-song rate or a flat fee, you'll need to stay aware of the time you're spending, focus on the major issues, and realize when the law of diminishing returns is kicking in and it's time to wrap things up.