10 Alesis Masterlink Tips

The Masterlink Has Been a Workhorse in Studios - Here's How to Get More Out of It

By Dan Tinen and Craig Anderton

The Alesis MasterLink was one of the first stand-alone CD burners designed specifically for mastering; it incorporates a host of mastering-oriented DSP, and can record/play back CD-Rs in 24-bit as well as 16-bit resolution. These 10 tips will show you to get even more out of this versatile device.

The Multiple-Disc, Multiple-Playlist Mix

Because standard CD-Rs can hold a maximum of 650 MB (or 72 minutes of 16/44.1 audio), if you want to prepare an hour-long DVD that uses 24-bit resolution, you'll need to split the "album" into at least two CD-24 format CDs worth of material (24-bit data storage uses up 50% more capacity than 16-bit material). Also prepare a single, 16 bit/44.1kHz Red Book CD to show the DVD mastering lab how you want the final project assembled. You'll end up with three or more CDs to send:

• CD24 (high resolution) with the DVD's first half
• CD24 (high resolution) with the DVD's second half
• Red Book CD with all songs at CD-resolution, in order, with spacing and fades
  
  

Check with your DVD mastering engineer to find out requirements for received material; note that CD24 disks do not require a MasterLink for playback, as they store AIFF files that can be opened by most digital audio editors.

Incidentally, mastering engineers may request that the DVD tracks be submitted to them "raw," with no DSP or fades, so they can do that work on their (very expensive) workstation. If you add the types of DSP and fades you want to the Red Book CD, it can help get across what type of sound you'd like to hear on the final DVD.

Fade Curve Options

MasterLink includes three fade curve options. Each one is useful with different types of material.

• Lg1 (concave logarithmic) is best for most music. The logarithmic curve fades abruptly at first, then decays slowly.
• Lg2 (convex logarithmic) keeps the volume up as long as possible, then does a quick fade. For example, suppose you want to fade out a 30-second commercial, but the narrator goes almost to the end. Add a short (500 ms) fade with the Lg2 curve.
• Lin (linear) can seem abrupt compared to a logarithmic fade. However, if a song already contains a fade made during the original mix, the Lin curve can shorten the fade but preserve the existing fade's curve.
  
  

Audio/Metering Mismatches

Because of the MasterLink's "look ahead" feature (i.e., it analyzes the signal before applying compression to catch any transients), when auditioning DSP it takes about 500 msec to hear any changes made to the EQ or compression settings. Therefore, the display will change before you hear the sound that corresponds to that change.

Using the Compressor's Key Function

The Key option can compare either the right, left, or both signals to the specified threshold when determining whether to apply compression. L & R is the most common choice, as linking the left and right channels together allows changes in one channel to affect the other, which preserves stereo imaging.

However, suppose the channels are out of balance because the left channel has higher peaks than the right. Set the threshold just above the right channel's highest peaks, and key to left channel only. When the left channel signal exceeds the threshold, compression kicks in on the left channel. As the right channel signal does not exceed the threshold, it will not be compressed.

The Importance of Limiter Release

With short release times, the limiter tracks every little change in level, producing a poten¬tially uneven or "choppy" effect. However, overly-long release times can result in dynamics that sound unfocused and mushy. If the dynamics are crisp and well-defined, and the transitions as the MasterLink goes in and out of limiting sound smooth, then you've found the optimum release time.

Note that the rate at which the compressor's gain reduction meter decays provides useful information about the compressor's release time setting. If the gain reduction meter darts rapidly among segments, the release time is relatively fast. If the meter transitions more slowly between segments, the release time is slower.

Don't Add Fade Ins and Outs While Mixing

Fades are best left for the MasterLink's mastering process, because with digital fades, the change in level is not continuous, but goes through a series of very tiny steps. Inexpensive digital gear often has coarser steps than the MasterLink, which uses a high-resolution fade algorithm to minimize the gap between steps. This produces smoother, "sweeter" fades.

What's with the Multiple Levels

There are three principal ways to change levels:

• The overall level that's set for the entire track
• Gain applied in the compressor
• Gain applied by normalization
  
  

Choosing the correct place to change level is like gain-staging in a mixer, with the three MasterLink controls above corresponding respectively to a mixer's input gain, channel volume, and master volume.

The track level shown on the top line of Playlist Edit mode is the first in the chain. Changing this affects what the compressor "sees," which lowers or raises the threshold where the compressor starts to work. The normalize setting is the last in the chain, and makes up for any gains or losses in the compressor, EQ, and limiter before it.

Want a Hot Sound? Try the Least Obtrusive DSP First

Let's assume a CD doesn't sound "loud" enough, and you want to make it louder with dynamics processing. Try the DSP effects in the following order:

1. Normalize. This doesn't affect dynamics at all, it just raises the overall level to use up the maximum available headroom. If this delivers the results you want, there's no need to go further.
2. Limiter. This affects only the peaks of the signal, leaving the rest pretty much intact. Light amounts of limiting (e.g., a threshold between --6.00 and 0.00) can increase the overall loudness without perceptibly altering the song.
3. Compression. Compression affects the signal the most. Use this if you want a really "hot" CD (e.g., dance music), or if the dynamics need to be restricted more drastically than what you can obtain with limiting.

Use a Safety Net

If mix from your multitrack directly into the MasterLink hard disk, until you burn a CD, there's no backup for that mix. If the hard drive fails, a power surge takes out your gear, or an errant pet knocks over the MasterLink onto a concrete floor, you may lose that mix forever. To be safe, mix to the MasterLink and back up to CD as soon as the mix is complete.

Choosing Sample Rate Rate and Word Length When Mixing

Many engineers use MasterLink as a "drop-in" replacement for DAT, and mix directly to it from a mixing console. As MasterLink can record at different resolutions and sampling rates, here are some guidelines on how to set the MasterLink's parameters, depending on the type of source material that feeds it.

If your mixing console is analog:

Use a 20 bit word length. The noise floor that results from the console itself, microphones, preamps, and all the components connected to it is much higher than the 20-bit encoding of a digital recorder. You can of course use 24-bit, but you're just wasting disc space to store random numbers well below audio thresholds.

Use a sampling rate of 88.2 kHz. Recording at 88.2 kHz will give you and future listeners on DVD the chance to hear any high frequencies in the 20-40 kHz range that may have been recorded by analog equipment (although very few microphones actually produce frequencies up there, and even fewer speakers can reproduce them). Professional mastering engineers like to receive 88.2 material, because it allows them to use EQ in the top audible octave (10-20 kHz) without distorting the anti-aliasing filter required for 44.1 kHz operation. An additional advantage is compatibility with today's CDs, as sample rate conversion from 88.2 to 44.1 causes no sonic artifacts. This is because conversion simply involves taking every other sample, as if the source material was originally recorded at the lower rate. However, if saving space is important, then you can always go with 44.1 kHz.

If your mixing console is digital:

Use a 24 bit word length. While the noise level of analog input signals to the digital console is (at best) at the 18 to 19-bit level, the digital signal processing used by the console itself will generate usable information down to the 21-22-bit level. (When you lower a fader of a digital mixer below unity gain, it is mathematically dividing the numbers and creating significant bits of resolution: a 19-bit signal running through a digital fader set to --12 dB becomes a 21-bit signal.)

Use a sampling rate of 44.1 kHz. If your digital console has 88.2 kHz capability, you'll of course use that. But for the 99% of digital consoles that are limited to 44.1 or 48 kHz, you may as well use the CD-compatible 44.1 kHz rate to avoid the sample rate conversion process to go from 48 to 44.1 kHz. Although modern sample rate conversion algorithms don't degrade the sound like earlier algorithms did, it's always good practice not to add unneeded processing.