header image header image

Sound Advice

Staying In Synch – Part I: Word Clock Explained by Bob Snelgrove

December 18th, 2002

Everyone using more than one piece of digital audio equipment should be concerned about the quality of their studio’s Word Clock. This article will explain the critical role that Word Clock quality and distribution plays in the digital audio environment and the audible effects that poor clocking has on digital music systems. Examples of common mistakes and suggestions for proper hook-up will be made.

I need to start off by saying that when it comes down to quality of sound, I tend to approach the wonders of digital audio with a healthy degree of caution. I love my CD player and I personally own and enjoy lots of other digital audio toys. Nevertheless there are still many problems with the digital representation of audio that have yet to be solved. All of these problems relate to digital audio’s sonic accuracy and transparency.
At best, well executed word clock generation, synchronization and distribution will make your studio sound audibly better and allow you to become a better artist, engineer or producer. At worst it will simply eliminate countless technical gremlins that will make you inefficient, ruin your mixes and drive you crazy.

Timing
Despite the many similarities, compared to analog audio, digital audio has a couple of unique and major differences. The first is that an analog audio signal is a continuously varying signal, which is represented digitally by a limited number of discrete numerical values. The second is that these numerical values represent the analog signal only at specific points in time or sampling instants rather than continuously at every moment in time. Sampling instants are determined by various devices and processes, the most critical being Analog to Digital and Digital to Analog conversion. Converters are responsible for transforming an analog signal into a digital representation and back again, this is where it all starts and this is where it all ends.

A sample clock determines when these sample instants occur. All digital audio devices have some form of a sample clock to control their internal sample rate or sampling frequency. In a studio where we integrate many different pieces of equipment that all depend on their own clocks to function we will invariably have sample instants taking place at different times unless we synchronize all the clocks in each piece of equipment and tie the timing of these events together.

This synchronous timing is required because unlike analog audio, digital has a discrete time structure consisting of individual samples. Successful communication between different digital audio devices or the mixing of different digital audio signals together will fail if each device is not producing their bits of data in precise co-ordination with each other.

Poor quality timing between multiple digital audio devices or the improper distribution of those timing signals will result in non-synchronous operation and the result will be the creation of random and highly audible artefacts, often described as clicks, pops or glitches.

There are two standard timing signals used to synchronize the internal sample clocks of digital audio equipment. The first, commonly used in large post-production and broadcast facilities is the AES3 Digital Audio Reference Signal or DARS for short. This bi-phase signal’s carrier is exactly the same as a balanced AES3 signal but no audio data (digital zeros) in the data stream. An XLR sync input can usually be found on pieces of high-end audio equipment and workstations. DARS is distributed the same way as AES3, via balanced 110-ohm digital audio cable and AES3 XLR type connectors.

There are two things that make the DARS or Audio Black signal particularly attractive. The first is its high frequency of operation, which is between 2 and 3 Mbits per second. The second is the fact that the professional AES3 interface is balanced and ground isolated making it relatively immune to induced noise, which can be a major source of jitter.

The most common clock distribution method however is Word Clock. The Word Clock waveform is a simple unbalanced square wave. Word Clock is designed to be distributed on 75-ohm, unbalanced coax cable terminated with BNC connectors. In order for synchronous operation to take place all digital audio devices must be fed one of these timing signals from a master reference Word Clock time base. These are typically referred to as Word Clock generators and sometimes as synchronizers depending on the functionality they provide. Great care must be taken to carefully distribute these Word Clock signals to each piece of digital equipment in the studio or the timing signals will be degraded and audio quality will suffer.

Word Clock Generation
The most basic requirement for a Word Clock generator is that it must be able to produce high quality stable square waveforms. The square waveform will need to be at one of two frequencies of either 44.1 or 48 kHz also referred to base frequency or Fs. The generator must also be able to produce industry standard multiples of either of these two base frequencies yielding Fsx1 thru Fsx4. This x1 thru x4 multiplication generates the common sample frequency sets we are all familiar with in digital audio and generates Word Clock frequencies in the range of 44.1 to 192 kHz.

A special case of declining interest and usefulness is Digidesign’s proprietary SuperClock at FsX256 or a frequency between 11.289 and 12.288 mHz. Because different manufacturers design their products to accept different multiples of Word Clock base frequencies, all multiples from one to four must be supported, and the master word clock generator must be able to produce different Word Clock frequencies from different outputs simultaneously.

It is important to note that not all BNC word clock outputs are created equal. When distributing Word Clock signals, isolation and proper source impedance from each output BNC is important. If the outputs are simply fed from one low impedance source then it will be impossible to correctly terminate a line and a single bad cable or poor connection will reflect back and compromise the performance of every other line. This is one condition that would also create jitter.

Join us next issue, where we continue are look at synchronization in the digital world, tackling the topic of jitter.

Bob Snelgrove is the President of GerrAudio Distribution and the Canadian Product Specialist for Audio Precision test instrumentation.

The Musicality of Mastering by Marisa T. Déry

December 18th, 2002

In this article I will be writing about “The Musicality of Mastering”. Although I will touch on some technical issues, I’d like to focus on the creative process of mastering. The mastering engineer’s role seems to be changing a bit. Whereas before a person would walk into the room and I would EQ it in the best way that I could (adjusting levels etc.) now I’m actually putting more and more special effects in the mix – record noise, backwards snare, flange on a section of a song (à la Britney) – people are asking for my input.

First, I would like to touch on a much talked about subject amongst Mastering Engineers: L-O-U-D-N-E-S-S

Play a CD that is five years old, then play a new release, and you will hear that the difference is staggering.

Ex. Marvin Gaye’s “I Want You” (Marvin Gaye’s Greatest Hits, Motown) then Linkin Park’s “One Step Closer” (Hybrid Theory, Warner Bros.) then Marvin Gaye again.

What is happening now is that music is getting louder and louder at the expense of dynamic range. In the early ’90s, the reference level was -12dB on most DAT players, which is why many old players had a line at -12. Then came the finalizer and people began setting their levels to 0. The problem was that every DAT player manufacturer had a different reference level for 0. Makers of consumer DATs would set the meters hot so that inexperienced users wouldn’t distort their recordings. 0 wasn’t “0” anymore. The finalizer made things worse because you could set the mix with an OUT ceiling of -0.3dB (which is the recommended maximum for CDs), yet still make your program louder and louder (while still remaining at -0.3dB).

The question is, “When is loud TOO LOUD?” All that I can say is that you need to leave room for the music to breathe. People are handing me mixes at 0dB, because the engineer cranked up the finalizer or the limiter conveniently located in the studio. Engineers are concerned that their clients won’t be impressed with their skills, so they give them a “finalized” mix where there is absolutely no room for me to do anything. 0dB is also dangerous because many CD-burner towers assume that if the program is peaking at 0.0dB it must mean that it is overloading, and promptly rejects all the CDs being duplicated (it’s quite impressive to see all those CDs popping out with flashing lights by their side).

A Good Mix
A mastering engineer relies on getting a good recording and mix to do his/her job properly. Too often musicians run out of money after the mix, and are never really satisfied with their mix. Your job is for them not to remember that they didn’t like their mix.

Analog vs. Digital EQ
There is importance in combining noise to the chain going to the DAW. Generally, I will extract a mix from a CD with the Adaptec Toast Extractor into the computer. It will be imported into Pro Tools where I have an endless amount of Audiosuite plug-ins (bells and whistles). This faster technique is used because the client wants everything to be done as quickly as possible. “Extracting” unfortunately (fortunately?) is faster than real-time loading. What I prefer is a DAT or CD master, where I will patch it into an analog EQ then go to Pro Tools. Those mastered mixes, to me, sound human.

There is a breath in those mixes that I cannot replicate with digital processing; there is a noise, a life to those mixes. One must never forget that what you are mastering is music. An artist puts time, energy, emotion and passion into those songs. Out of respect to the artist and the music, you have to make that mix breathe and come alive. You can’t process it to such an extreme that there is no dynamic range, no peaks and valleys, no life. It’s just a block of noise, a block that you can beautifully see in Pro Tools or any other program (ex. the L1 set at 12dB threshold).

Audio Restoration
Another thing that I would briefly like to touch upon is Audio Restoration. Whether you are dealing with old reels or 78 RPM records try to make them sound as natural as possible. There are many outboard EQs and software plug-ins for that purpose. The Waves restoration package is one that I use a lot. Yet even there, you must listen with musician’s ears.

Resist the temptation to get rid of the entire hiss, especially with orchestral music! It’s not only about the sonic quality; it’s also about the music. Be creative when you are working on these programs. I have a little Casio keyboard at work and when I can’t figure out what frequency is humming at full volume (I’m stuck and/or tired), I’ll grab the Casio and find the note on the keyboard. I have a chart that associates the notes of a piano keyboard with frequencies, so if the note (or hum) is Middle C, I’ll look at the chart, and find that I need to notch out 261.63 Hz – it’s a start.

Which brings us to the creative side of mastering …

Recently, I did a project where the artist came up to me and said, “I have three songs that are mixed, and one that is unfinished. I have an appointment with the A&R rep at DreamWorks on Friday … help me.” So I listened to the said “unfinished” song and began throwing suggestions.

– Why not throw a Janet Jackson type drum loop at the head?
– During the Tag section at the end “chorus” it.
– When it comes back “flange” it.
– In the beginning, listen to the lyric. Play with it.
– Pan to the left when you say, “left”.
– Pan to the right when you say “right”.
– Make it move.

I am now listed as “Remix Engineer” …

There are a lot of plug-ins that you should play with – too many to cover them all, but the L1 Ultra-maximizer (now L2 Ultra-maximizer) must be mentioned, as it is now a must in all productions. Before it was recommended that one should set the attenuation meter/setting between -3dB and -6dB; now anything goes, so all you can really do is match the levels of a new release.

But be aware! It has been said (AES Conference, New York, 2001) that 9 out of 10 songs on Billboard’s Top 10 are distorted, and that songs from the 1970s sound technically better and have more dynamic range than songs released in 2001-2002. We have become a generation of “distorted” listeners (it’s no wonder that teenagers today will be partially deaf by the time they reach 30). Hopefully 5.1 technology might help ease the loudness wars.

In Summary
I’m writing this article because of my concern with where we are going with the loudness wars. I am an engineer by trade, and a musician by birth. I have pursued a career in engineering because of my unquestionable passion for music. I respect Creators and Performers that bare their soul to tape. They rely on the Basics Engineer, the Overdubs Engineer, the Assistant Engineers, the Mixing Engineer and the Mastering Engineer to preserve the integrity of their music to tape. It is our duty to understand their music, their art.

It’s not just about putting a mic in front of an instrument and pushing the record button, or adding highs and lows in the mastering process. It’s about understanding what you are recording, mixing or mastering. Using your instincts to make it sound right. As a technician, your job is to make the music sound as sonically perfect as possible; as a human being, your job is to make the music sound as human as possible (with or without noise).

Marisa T. Déry, a native of Ottawa, Canada, is Chief Mastering Engineer at the Tape Complex in Boston, MA. Her clients include the Mighty Mighty Bosstones, Tugboat Annie, Scientific, Chapter In Verse and RUSHYA.

Good Amps and Power Efficiency

December 18th, 2002

PA amplifiers need to combine the delicacy of a good hi-fi amp with the robustness and reliability of a farm tractor, blending (increasingly) with the low weight and compactness of aeronautical gear.

Good-sounding power amps (ones which add minimal colouration or distortion to the signal, purely making it louder) require great sophistication to enlarge and deliver the signal very precisely over a wide ‘canvas’ of levels and frequencies, while also delivering high currents and voltages.

And these quantities are not delivered into docile power-absorbing elements, but instead into speakers, which are quite complex and ‘reactive’ in the way they interact with the amplifier.

No power amplifiers are 100 per cent efficient – even the best manage only about 80 per cent in reality. The best speakers, meanwhile, only approach 25 per cent efficiency. Best overall efficiency is consequently about (0.8 x 0.25) = 20 per cent.

The average overall efficiency figure is more often between five and ten per cent. Taking ten per cent as an approximate figure, this means to get a certain amount of acoustic power – in other words music at a suitable sound level – in the room, we have to provide about 10 times that power from the electricity supply. And so this is the amount that an audio power amplifier has to handle and ‘process’.

We’ll also want to have some power capability in reserve – since inadequate power results in amplifier overload and bad sound. In general, erring on the side of over-rating is better than under-rating.

And remember that the relationship between watts and loudness isn’t proportional in the way you might imagine. As a reminder, a rule of thumb is that you need to increase the power delivery into any particular speakers by at least tenfold (x10) to attain about twice (x2) the audible level. This appears on a sound level meter as a 10dB higher SPL (sound pressure level) – so, for example, if 100 W gives 90dB SPL, 1,000 W will be required to increase the level (where nothing else is altered) up to 100dB SPL.

In short, much, much more power is needed than you might expect.

This article is reprinted with permission from The Live Sound Manual, published by Backbeat Books, www.backbeatbooks.com. All information is copyrighted and cannot be reprinted without the permission of the publisher.

Recording The Lead Vocal

December 18th, 2002

How Many Tracks Is Too Many?

More often than not, the lead vocal is the track that contains the most emotional content of the song. With repeated attempts at recording the vocal, you run the risk of losing that emotion and “magic”. So while it’s ideal for the singer to nail the perfect take in one or two tries, a good engineer knows how to respond the other 90 per cent of the time.

The answer is to compile the best elements of a few different takes into a single, composite performance where each line, each phrase and even each syllable is sung just the way you want. This process is called “comping”. It’s done on nearly every record you hear, even the ones you’re convinced are single, complete takes.

Tip: If the singer is hesitant to record this way, claiming “artistic integrity”, remind them that they’re free to sing the song through from top to bottom, without interruption. Meanwhile, just switch tracks while you’re winding back to the top after each take. (Make sure you’re only sending the current take to the headphone mix – it can be very disconcerting for a singer to begin a song and hear two voices coming out of his mouth.)

In this digital age of virtually unlimited available tracks, it’s tempting to record 5 or even 10 different takes before comping the vocal. But using that many can really overwhelm you and confuse the process. Try utilizing two or three tracks instead. Starting with your first take, tell the singer it’s only a practice take for the purpose of further level adjustment (when in fact you’ve already adjusted everything and are ready to go). This is useful for anxious singers, taking the “pressure” off them.

After two or three takes, stop if you have terrific performances overall. If not, go back to the track with the least inspired take and record over it. Hopefully, you have gained the singer’s trust by now and don’t need to inform them of these details. Continue with this process until you feel that, within those two or three tracks, you have the makings of a great performance.

When you’re ready to start comping, draw lines on the lyric sheet so you can make little notes (check marks, yes, no, good, bad, maybe) on each line of each take. Involve the singer in this process only if they insist – the more they analyze their own performance, the less they’re likely to respond with an inspired, heartfelt one. Once you have usable takes for each line, bounce the winners onto a fresh track (you can also bounce certain lines from “alternate” takes into one take that just needs a few fixes).

Tip: After you have a comped vocal, get away from it for a while (dinner break, TV break, whatever). Then listen to it with fresh ears, and with the singer, to see if you still need to fix something.

This article has been reprinted from the Studio Buddy software. Written by acclaimed producers/engineers Michael Laskow and Alex Reed, Studio Buddy gives hints and tricks on various recording techniques. To download a free copy, go to www.studiobuddy.com.

Crossovers: How Many ‘Ways’?

December 18th, 2002

Whether the crossover is set up for two, three, four or more bands will be largely pre-determined by the PA’s speaker system. Here’s an outline of what the different numbers of ‘ways’ generally accomplish, and how they’re placed.

Two-way
Two-way crossover systems require a minimum of two amps, so they’re referred to as bi-amped. The split signal feeds just low frequency (LF) and high frequency (HF) drivers. Such a simple set-up is typically restricted to monitors or mini PAs, because for most kinds of music it’s not possible to cover the audio range fully enough (particularly with low-enough bass) with only two types of drive units.

Three-way
Three-way systems (sometimes called ‘tri-amped’) feed low-, mid- and high-frequency drive units, and are the most widely used configuration for ordinary FOH PAs and more elaborate stage monitor cabs.

If you require a three-way PA system on a tight budget, you can use set-ups such as ‘bi-amped + passive split’ or ‘two-way active with passive split.’ All that’s required is a single two-way stereo crossover, a minimum of one stereo power amp per side and three-way cabs wired for this approach. The system is still bi-amped, but is also three-way. It’s achieved by ‘splitting’ the HF band’s signal (really mid + high) using an additional passive crossover in each mid/high speaker cab. Hence ‘passive split’. The active crossover splits the bass from everything else. The passive then splits the HF from the midrange – like a sequence of two forks in a road network.

Despite the limitations of passive crossovers, passive splitting of HF (only) can work quite well.

One restriction of the ‘passive split’ configuration is that the top-end’s level can’t be limited with any discrimination – any protective limiting is ‘lumped in’ with the midrange (we’ll cover limiting in more detail shortly). But this only affects the ultimate sound level capability. The excess levels in one or the other will turn down both. Still, the limit has to be set low enough to protect the tweeter, rather than the mid, which would handle more power.

If higher sound levels are really needed, more active ‘ways’ are simply going to be required. But you can still add a passive split to these.

Typically this is done for occasions when super-tweeters (working at frequencies above about 14-16 kHz) might be used – for example providing a PA in a small venue where high frequencies won’t be largely absorbed before they’ve reached the audience.

Four-way
Four-way crossovers (quad-amped) will either send signals to a sub-bass driver, plus low, mid and high units; or else to low, low-mid, high-mid and high-frequency speakers. It’s mainly used for more up-market FOH PA.

Five-, Six- and Seven-way
These were more common in the past, when PA developers experimented with different schemes – and before accountants worked to prune tour costs to the bone. The frequency ranges became further sub-divided and the speakers used were increasingly specialized. These higher-way systems are more complex, but offer potentially higher sound quality and ‘maximized’ power handling.

But the ‘law of diminishing returns’ sets in fairly steeply once the crossover has split-up the audio range into four fairly equally-sized bands. For one thing, the weight of amps and cabling, as well as the wiring and rigging complexity, is bound to start increasing substantially, but without much worthwhile increase in sound level or quality. There’s also an acoustic trade-off, in that it can be increasingly hard to synchronize larger numbers of separate sound sources.

Despite this, you will occasionally still meet five-, six- and seven-way systems. There are commercial analog (active) crossovers with five and more bands, some of them flexible modular types. Or else they’re bespoke (custom-made), way crossovers. The more upmarket digital crossovers also typically offer up to six bands.

This article is reprinted with permission from The Live Sound Manual, published by Backbeat Books, www.backbeatbooks.com. All information is copyrighted and cannot be reprinted without the permission of the publisher.

Becoming a Producer

December 18th, 2002

Almost as soon as assistant sound engineers learn enough about studio recording to spell “EQ”, their ambitions begin to drift towards the possibility of working as a producer. They see dollars, or power, or the seductive pull of being totally in charge of the artistic direction of a recorded work, and sooner or later, I get the inevitable question: “How do I go about becoming a producer? Is there a course I can take?”

It all depends on how you define “producer”. In its simplest sense it can mean anyone involved in sound production – so if you are a songwriter, musician or sound engineer, you could be regarded as a “producer of sound…”

If that’s what you want (and you have the talent), then experience, skills in recording (i.e. “music production”) can be gained through on-the-job training, sound engineering courses or a combination of both. There are many courses from which to choose.

However the term “producer” in its more accepted sense describes the person responsible for the total sound and feel of the finished track. Responsibilities generally include choosing and arranging songs, selecting and rehearsing the band and any additional musos, working with the engineers during tracking and mixdown, and guiding the mastering engineer during final post-production.

The producer may employ others to help realize their vision, or may take a hands-on approach to some of the duties such as recording or mixdown. In any event, a producer generally leaves the imprint of their own style on the finished product while still allowing the style and talent of the performers to shine through.

Becoming a recognized producer therefore involves developing a professional skill-set through a number of different avenues. Essential areas include music knowledge, theoretical, technical and practical knowledge of sound engineering, mixing and post-production. A very good set of ears and ability to work effectively with people is a must. A knowledge of the capabilities of a wide variety of studios and other audio facilities helps too.

Audio courses may help, but they need to be extremely comprehensive. The Germans have a word for the technical side of a producer’s job: “Tonmeister” – Master of Sound. There are some “Tonmeister” courses around, but check first that they are respected by the industry, because there are some that are “Tonmeister” in name only. One of several courses with a good reputation is the degree course at Surrey University in the UK. Any short, part-time courses claiming to make you a “Master of Sound” may be stretching the truth just a little!

Whether you do a course or not, without a track record in the real world it is unlikely that you will ever get a producer’s gig – unless you are very persuasive.

Many successful producers have not completed any formal training at all, except perhaps for music. They have paid their dues through the experience of being musicians and engineers and show such a fine grasp of the sensibilities of putting together recorded works that others ask them for help with their music.

When that happens, and continues to happen, you are a producer.

Vyt is the Managing Director of Audio Training Consultants, who operate the audioEd Pro Audio Resources Site, www.audioed.com.au. Before ATC, Vyt owned and operated several professional recording studios and an accredited audio school for more years than he cares to remember.

Contact

4056 Dorchester Rd., #202,Niagara Falls, ON
Canada L2E 6M9 Phone: 905-374-8878
FAX: 888-665-1307 mail@nor.com
Web Site Produced by NWC

"