Taking the Mystery Out of Mastering

As featured in: Performing Songwriter Issue #70, June 2003, Visit performingsongwriter.com to order back issues or subscribe.

By Craig Anderton

Mastering is the stage between mixing and the pressing plant, where you add any final edits or enhancements (fade in or out, compression, etc.) to your collection of stereo mixes. As the final link in the CD production chain, mastering can make or break a project--which is why people often pay big bucks to a veteran mastering engineer. Unless you have much experience, hiring someone who really knows how to master could be the best move for your music.

However, sometimes mastering can be handled a bit more casually, like when you’re making one-off CDs as demos or to sell at gigs. And even if you do go to a professional mastering engineer, trying your hand at it first might help you to communicate your needs better. Besides, if you do enough mastering, hopefully at some point you’ll get good at it.

(Do remember that mastering is not intended to salvage a recording but enhance an already superb mix. If there’s a problem with the mix, remix the tune--don’t count on mastering to solve the problem.)

MASTERING-THEN AND NOW
Prior to the digital revolution, you brought your finished mixes on tape to a mastering engineer, who would use various signal processors to sweeten the sound. The tunes would then be assembled in the desired order, and vinyl test pressings made so the final product could be evaluated prior to mass production.

Mastering was rightly regarded as an arcane, mystifying art. Few musicians had access to the high-end, expensive tools needed to do mastering; nor did they have the experience of someone who had listened to thousands of recordings and knew how to prepare them for mass production.

Today mastering tools that are useable with personal computers are within the financial reach of just about everyone. However, experience is still not something you can buy or download, so it’s no wonder that even those with decent mastering tools and technique often seek out a professional’s objective viewpoint.

MASTERING PROGRAMS
Most project studio mastering is done within a two-track digital audio editing program--you import a mixed file into the program and start tweaking. Programs for Windows include Steinberg WaveLab (or its less expensive “lite” version, WaveLab Essential), Sonic Foundry Sound Forge, the very cost-effective Syntrillium Cool Edit Pro and Magix Sequoia. For the Mac, popular options are BIAS Peak, TC Electronic Spark (both Peak and Spark are available in “lite” versions at different price points) and Prosoniq sonicWORX Studio. These generally offer superb navigation facilities and the ability to host plugins.

Plugins, which range from small to ambitious, are programs that accessorize the main editor with a variety of functions, from the mundane to the esoteric. Note that there are multiple plugin formats; check for compatibility with your host of choice. Fortunately many plugins are available in different formats. For an overview of the subject, see the May, 2003 issue #69 of Performing Songwriter.

There are options other than standard digital audio editors. Some programs, like Emagic’s WaveBurner Pro and Sonic Foundry’s CD Architect, started as CD-burning programs. But they’ve since acquired signal processing and waveform manipulation tools that make them well suited for general mastering tasks.

Another program, IK Multimedia’s T-Racks 24, does not do waveform editing but is instead a software suite of vintagesounding processors. It’s also available as a plugin for host programs. 

Some multitrack recording programs have become sophisticated enough that a few people use them for mastering, too. The master bus will usually have virtual “slots” for plugins which you can tweak in real time, at which point you can apply the effects to your mix and export it as a finished WAV or AIFF file.

MASTERING TASKS
The mastering process addresses several areas:

Equalization. This balances out a track’s overall frequency response. EQ can even affect the mix by emphasizing the frequency range of particular instruments or vocals.

Fade ins and fade outs. Editing programs often allow for complex fade shapes.

Dynamics processing (compression/limiting/ level maximizing). These are all variations on the same basic process, which is to even out the dynamic range, resulting in more apparent loudness.

Occasional effects. Adding a bit of reverb may help give an additional feeling of space and stereo imaging.

Cleanup. This is your last chance to remove any remaining noise, hum, pops, clicks, crackles, etc.

Dither. This process promotes better sound quality when going from highresolution audio recordings (e.g., 24 or 32 bits) to lower-resolution audio (e.g., the 16 bits used in CDs). We’ll say more about this important process later.

PLUGINS FOR MASTERING
Plugins designed for multitrack recording, where there might be several instances of a plugin on multiple tracks, sometimes sacrifice sound quality to avoid hogging the computer’s CPU power. Fortunately, processing a single track usually isn’t as crucial as processing a final two-track mix, but true mastering plugins need more horsepower for more refined algorithms.

Several companies sell “bundles” of plugins for mastering. Waves’ plugins are highly regarded by mastering professionals for their sound quality, but they’re not the only game in town. For example, Steinberg makes a Mastering Edition set of plugins; iZotope’s popular Ozone DirectX plugin contains an entire suite of integrated mastering tools, including multiband dynamics, EQ, level maximization, reverb and metering.

MASTERING-ORIENTED PLUGINS
Following are some of the most common plugins used for mastering:

Multiband Compression.With a traditional compressor, a strong signal that exceeds the threshold brings down the gain, but this affects all frequencies. For example, if a strong kick drum hits, it will bring down the level of the cymbals and other high-frequency sounds as well. Multiband compressors split an incoming signal into several bands (typically 3 to 5). A compressor follows each band, so that each compressor affects only a specific band of frequencies. When properly adjusted, this can give more effective, transparent dynamics control than single-band compressors.

Mastering EQ. This falls into two broad categories: multi-stage parametric EQs designed for precision signal shaping, and “vintage” EQ circuits designed to add some particular attribute or flavor rather than solve specific, detailed response problems.

Unlike the 4-band EQs typically found in multitrack hosts, mastering parametrics may have six, seven, eight or even more stages. At least some of these will probably have adjustable responses, such as bandpass, notch, high shelf, low shelf, etc. For example, Sonic Timeworks’ equalizer provides six parametric stages and separate high cut/low cut sections. It also includes a spectrum analyzer that can switch between monitoring the incoming signal’s spectrum so you can see where response peaks and valleys lie, or monitoring the output so you can see the results of equalization on the spectral response.

Vintage equalizers tend to have gentle response curves that sound more natural, like the Pultec equalizer, a studio mainstay during its heyday in the ‘60s and ‘70s (also see the sidebar, “What About Hardware?"). Software emulations of these hardware classics range in terms of quality, but some are exceptionally good (e.g., the Bomb Factory EQP-1A plugin and the Pultec module of the Mackie/Universal Audio UAD-1 Powered Plugins board).

Level Maximizers. With the trend toward ever-luder CDs, new and more potent
strains of dynamics control have appeared called level maximizers, loudness maximizers, level optimizers, etc. These restrict dynamic range by essentially changing the level of peaks to go no higher than a usersettable threshold; it’s then possible to increase the overall file’s level by whatever amount the peaks were reduced. Although not as flexible as traditional limiting or compression, for a given perceived increase in level, maximizers can produce fewer obvious audible side effects (but don’t go overboard--unrelenting loudness and lack of dynamics ultimately leads to listener fatigue). The industry standard is the Waves Ultramaximizer and its descendants.

"Warmers” and Other Distortion Devices. It may seem heretical to add distortion to a final mix, yet the concept has a precedent: in mixing to tape, engineers would frequently hit the levels really hard so the tape would distort a bit. Many companies have tried to emulate this effect with varying degrees of success. One popular option, PSP’s Vintage Warmer, combines “old-school” compression, overdrive, and high/low frequency equalization. It’s not supposed to be sonically neutral; instead, it gives you the option to “hype” the sound and give it a bit more “crunch.” Another is Steinberg’s QuadraFuzz multiband distortion unit (full disclosure: it’s based on a hardware box I designed that’s sold by PAiA Electronics).

You may want to add an equalizer before distortion to trim the highs a little bit, as distorted high frequencies can sound brittle. You have to be very careful--you don’t want your mix to turn into a muddy, distorted mess (well, maybe you do, if that’s your thing). But some engineers use light, subtle distortion as their “secret ingredient” for mastering.

Bass Enhancers. In addition to EQ, there are other ways to enhance bass. The Waves MaxxBass adds harmonics to the bass, so it can stand out better in systems that lack good bass response. When you hear the upper harmonics, your ear tends to “fill in” the bass’s fundamental. Other plugins take an opposite approach, adding sub-harmonics to give the bass more depth. Yet another approach is to add a stage of compression, optimized for low frequencies, to bring out the bass.

Stereo Image Enhancers and Harmonic Enhancers. Image enhancers widen the stereo image, particularly in the high frequency range where the ear is most sensitive to directionality. Some of these devices use phase changes or delay to enhance the stereo image, and therefore may give unpredictable results if played back in mono.

Harmonic enhancers add a high-frequency “sheen” without the use of EQ. As with bass enhancement, there are many ways to do this. One approach is to add a very slight amount of distortion, use a steep highpass filter to remove everything but the highest harmonics of the distortion, then mix this back in with the main signal. If used sparingly, the minute amounts of high-frequency distortion can give increased clarity and stereo separation, without the “tinniness” you might encounter with standard EQ.

As one real-world example, iZotope’s Ozone mastering suite plugin includes a multiband harmonic exciter that increases the sharpness and “definition” of audio within the selected frequency ranges, as well as a delay-based image enhancer. As with most mastering effects, though, be subtle--a little goes a long way.

Reverb. Reverb is not used a lot in mastering, but if material was mostly recorded direct (i.e., instruments plugged directly into the console rather than being miked), reverberation can add a little space and fill out the stereo image.

DITHERING
This is the last stop of the mastering process. It’s important because most modern digital audio hosts process signals with a much higher bit resolution than a CD’s 16 bits, such as 24-, 32- or even 64-bit resolution. Although it’s possible to just cut off the extra bits (a process called truncation), this produces an abrupt decay that, when listening to subtle musical material at high levels, can produce an audible “buzzing” type of sound.

The dithering process adds random noise to the lowest-level bits. It may seem crazy that adding noise can improve the sound, but adding noise to the lower-level bits increases their amplitude and pushes that information into the 16th bit. Okay, this is getting way technical, so let’s cut to the chase: With some types of program material, dithering produces a smoother, more detailed sound at lower levels.

However, there are different types of dither. Apogee is well-known for its dithering algorithms, which are not only highly regarded in the industry, but are included in Steinberg’s Cubase SX and SL. Dithering is also built into programs like WaveLab. Experiment with different dither settings to see which works best with the program material.

Note that dithering is always the last process in the chain so there are no subsequent processes to “undo” the effects of dithering. Also, do not dither more than once; it should happen only when converting your high bit-rate source format to its final 16-bit, mixed-for-CD format. Always keep an undithered copy of your unmastered audio. If you need to make changes later, work on the original and re-dither rather than work with dithered audio.

MASTERING FOR DATA COMPRESSED FORMATS
I used to do separate masters for CD and cassette; now I do separate ones for CD and MP3 or WMA. Mastering for a datacompressed format requires a somewhat different skill set compared to mastering for standard, uncompressed formats such as WAV or AIFF. This is because many tradeoffs are involved; when you save a file in a compressed format, you need to make various choices about its characteristics such as the target bit rate (will it be for dial up or broadband?), stereo or mono (a mono MP3 file is the same size as a stereo one but has better fidelity), and the like.

True data compression restores a file to its exact original contents when uncompressed. Audio compression schemes are really data omission--a process like MP3 throws away “unneeded” data. For example, if there’s a lot of high-level sound going on, the algorithm might assume it’s masking lower-level material and decide that reproducing a high level of detail is not needed in that particular section.

In any case, data compression is not about subtlety; I often use loudness maximizingtype tools to create a sound that doesn’t require as much dynamic range, thus allowing the data compression algorithm to work more efficiently.

A FEW FINAL TIPS
Mastering requires good ears and acoustics, so don’t invest in a ton of mastering plugins and a hot computer if your studio acoustics and monitors aren’t up to the task--fix them first.

As you master, compare the results to a well-mastered CD. Are the levels and spectral balance about the same? If not, you may need to re-evaluate some of your mastering judgments.

Most importantly, keep in mind that the adjustments you make in mastering are far less dramatic than for individual tracks. Even a dB of EQ boost can make a significant difference. Listen carefully, and good luck!