The titling of this post is a direct tribute to my tag team buddies Dan and Carmine from the ol’ DJ days. We were (and still are) and bunch of vinyl junkies. We admittedly had a few CDs of some digitized purchases that we toted around and spun, but primarily we threw down vinyl – for a variety of reasons. It put on a nice show visually to the patrons of our wine bar as fantastic eye candy, vinyl sounded much better compared to MP3s through our dbx Driverack PA and Crown 2 x 2000 watt amp system, and many of our top selections were vinyl-only releases which never could be, and still have not been, sold on CD or digital formats (some of these are limited-run foreign bootleg presses at that, so they likely could not have been sold by any other means, and in other cases the artist or label chose to do a vinyl-only release for their tracks). Vinyl is where I personally learned a great deal about the fundamentals of pitch and time manipulation through the tactile experience of “playing” with the music as I would go about cuing up everything during the course of the evening. Unlike digital, vinyl is tactile. It has form, it has shape, it has weight, it has a physically-cut representation of sound (grooves), it even has a strong but strangely pleasant odor right out of the package akin to the ‘new car’ smell phenomenon. So when manipulating pitch and time, not only can you hear what is happening, you get to see what’s making that happen.And there’s a wonderful beauty to learning the function behind the process by way of an analog media, unlike plugins where you press a button but don’t know what physically is being done, you only see the input and the output.
Unfortunately as vinyl fades away in the eyes of my generation and those generations following in our footsteps which may not even know what vinyl or a cassette tape media are, this input/output-only depth of knowledge concerns me. It’s limiting in a rather unfortunate way. If one can’t randomly click their way to a solution from Point A to Point B, then that is the end of the road – if they manage to get to Point B, it’s a happy accident to a fair degree, only replicate-able when the same circumstances are present and a ‘preset’ has been saved. By understanding the underlying process, it allows one to be presented with Point A and Point B, and intelligently problem solve a way to Point B – or even, discover through deduction and reasoning of the process that Point B may not be obtainable, so, how can they instead find Point C (a reasonable alternative), whether that be using a different tool, changing your method, or re-considering the process altogether.
So, back to vinyl and what it has to do with the title and subject of this post. While vinyl has seemed like the immortally immaculate storage and playback format for sound, it is highly susceptible to dust, dirt, grit, and wear-and-tear which cling to the surface – sometimes it’s not even easy to wipe off with a cloth. In the digital domain, all sound information is stored as ones and zeros; binary. It does not matter how high or low of quality the readability is of each one and zero digit is, but as long as each one is legible in some rudimentary way, the data recall is lossless. It does not matter how much noise interference becomes entangled with a digital signal of ones and zeros (inside a cable run, for instance). If the receiving end can still read each digit amidst the noise, everything is perfectly identical and the noise has zero bearing on the output – other than it made the system maybe work harder to recognize the digits easily. With vinyl on the other hand, it is stored within the analog domain as a physically etched groove of the voltage levels of a waveform across a duration of time; stared at close enough, the tiny details of a waveform’s intricacies will visibly emerge.
On a side note to vinyl, the groove is actually etched with a lathe using a (usually) sapphire stylus onto a soft lacquer disc, which subsequently goes through a silver electroplating process to create a negative Master (the grooves are raised). The lacquer is usually destroyed in the process, so this negative Master is used to create a metal positive Mother (playable like a normal record), and from there it is electroplated similarly to create what is known as a negative Stamper (each yielding a life of about 1,000 presses, the Mother is re-electroplated often to generate more than one Stamper as needed), and is subsequently stamped under more than 1,000 psi (as I recall) onto super-heated vinyl and subsequently blast cooled with water into a solid state, which are then cut and trimmed, bonded with a label, inspected, packed and shipped. So technically the vinyl you own is a physical stamping of an etching created in the exact invert mechanical process you utilize to play that record back. And speaking of which, when we play a record, the diamond tip of your needle’s stylus traces this groove, and the minute vibrations of this tracing process are translated back into their mV Phono-level sound signal, which is then pre-amplified with an outboard unit (although some come with exceptional onboard processors like the Numark TTX-series) into a Line level signal and applies the RIAA curve to invert and cancel out the equalization that was required to store the sound onto the vinyl in the first place (boosting the lows back and suppressing the highs). The resulting sound which feeds out from the pre-amplifier is the sound we all have come to know and love as what our vinyl record sounds like. But herein lies a problem. with the very first step – the stylus traces the groove. But what happens now when dust or dirt get stuck in the groove? The stylus traces that, resulting in crackles heard on top of the sound, because these granules cause the stylus to report back voltages of random nature, which have an exceptionally short period, and contain no harmonic information. Another definition of that is noise. What if somebody unfortunately dragged a pen across your record and you played it, or the stamper did a bad job? The stylus traces that too, resulting in loud snaps and pops on top of the sound, because what has happened here is the groove in these places have been effectively destroyed. With the angular momentum and torque of the platter, the stylus has to keep moving forward, and in doing so these hollowed-out or non-existent grooves cause the stylus to physically jump out of the groove and back into the next one very hard, and in some cases where there is some groove still left but it has been blurred into noise by the damage the same way that a gardener trims a hedge to a clean edge, the stylus reads a bunch of random noise in that area, resulting in spurts of fuzzy crackles. Depending upon how uniform the “streak of damage” is it’s orientation to the platter surface, these snaps and pops will continue with a predictable interval of occurrence on each rotation. Even worse, vinyl records are often pressed with recycled vinyl rather than virgin vinyl, the former of which contains impurities which subsequently are blended into a newly-pressed record and result in all sort of surface noise of crackle and pops – a Kathy Brown vs. Hott 22 white label bootleg I have is almost certainly 100% recycled vinyl, it sounds so awful above and beyond anything else I own. Unfortunately with vinyl as well, especially played a lot at a venue in which alcohol and cigarette particles accumulate readily within the air, these particulate matter forms adhered to our vinyl in addition to storage dust. Compound that with the light wear-and-tear damage accumulated through repeated plays (the stylus does slowly degrade a groove), occasional mishandling, and so forth, it’s quite easy to see that we had some records within our bag which had a lot of cosmetic damage, which led to them display snaps, crackles, and pops when we played them. With all of the vinyl we played, these little bits of tasty, cherished analog-ness found their way into almost every corner of our mix sets; snaps, crackles, and pops subtly woven in all over the place. And henceforth, The Rice Crispies Mix was the name loving-endowed by my tag team partners Dan and Carmine.
What does that have to do at all with what’s actually on my mind for this post? Well, I’m sorry to say but The Rice Crispies Mix is not limited to vinyl! Yes, indeed, we find it in film/tv sound too, and it’s here to stay. These sounds I speak of are known as ticks, clicks, pops, and lip smacks. It cannot escape being found in sound dailies, in the voiceover booth, or in the ADR session. It is up and down and left and right and front and back. It rears it’s ugly head everywhere in post production sound, and its utter eradication is one of the darling tasks of our venerable Dialogue Editors and ADR Editors, one such ‘dirty job’ task which tops of the list of thanklessness because of how subtle of an art form it is, it’s sheer tediousness, how completely invisible of a process it is to the overall sound, and how much of a lasting effect it has on the final product. In fact, a book by John Purcell is even titled Dialogue Editing for Motion Pictures: A Guide To The Invisible Art, Chapter 14 of David Yewdall, M.P.S.E’s book about post production sound is titled Dialogue Editors: Hollywood’s Unsung Heroes, and I’ve heard many times within the community that the saying goes: “good dialogue editors are worth their weight in gold”
A picture is said to be worth a thousand words, so here’s a smattering of visuals of these little critters collected while in the process of handling a dialogue edit ‘deep-cleaning’ pass, the critters of which have been highlighted.
Legend= vertical pale-grey grid lines define a measurement of:
1/25th of 1 frame (30fps) ≈ 0.008 seconds (30 fps)
1/25th of 1 frame (24fps) ≈ 0.010 seconds (24 fps)
Meaning, the total duration of nearly each visually-provided blue bar ≈ 0.016 seconds (30 fps) or 0.020 seconds (24 fps)
Damn these are small!! And their diminutive size is a understatement to both their persistent presence and ability to be under-estimated. It’s an important task to handle these, and not for the fainthearted and non-patient individuals alike, but it is a very essential process. The intent of this post is to both share a little bit about what these bits of digital gunk are and their origins, and a little bit of what Dialogue Editors and ADR editors cope with on a daily basis which is all but invisible to the outside word. So this post a little glimpse through the window into the world of Dialogue & ADR Editors.
Ticks & Clicks
They don’t carry Lyme disease, nor are they responsible for malicious drive-by downloads, yet they are both mutually undesirable much in the way these two situations are to us humans as a species. There once was a time when these were an absolute pain to deal with. And then ProTools came with 16-bit waveforms! What’s so special about that? Well, with 8-bit waveforms most ticks and clicks are not visibly detectable (for example, Ticks & Clicks, Samples 1, 2, and 5 would pretty much not be drawn) so it required an excruciating accurate and discriminating ear to find them. The silver lining here is that us dialogue editors who have been blessed with working under the 8-bit waveform paradigm have become very accurate and efficient in catching these by ear alone – and now the 16-bit (double resolution) waveform is just icing on the cake to speed us up in the edit. Ticks & Clicks are fast-attack transient snap noises, almost like the sound you hear when you get that snapping discharge of static electricity, and they sit in a classification gray-area between Lip Smacks and Crackle. These suckers tend to be found at home living around ‘S’ and ‘S’-related sounds, even with ‘F’ sounds. But all is fair game when it comes to these, as sometimes they can be related to the voice itself, and other times they can be environmental sounds such as dolly movement (dollies are infamous for being the bane of a Dialogue Editor’s existence), actor movement, crew movement, or technical issues such as interference or a bad cable. There can be many reasons and sources behind them to list all, but luckily these guys are quite easy to eradicate because their nature of being a quick burst of unnatural sounding noise sound very foreign to the ear and stick out like a sore thumb.
These guys, as with pretty much all of the other artifact types further described, can be brought to the foreground and “stripped naked” by running the dialogue through a compressor (RTAS, not AU), as compressors accentuate these transients of wholesome noise. Personally, I like to play my dialogue through such a chain during the second and third pass to help me weed out these artifacts – than again, I am usually requested to fully premix my dialogue in addition to editing it, so setting up a sidechain at this point is a natural step in my regular workflow. A recommended compressor chain for premixing (or just monitoring your edit to test it out) is about a 4:1-6:1 ratio, -8 threshold, 7ms Attack and 35ms Release. It’s just enough to properly babysit the RMS levels and stonewall the peaks, but still allow your dialogue to retain most of it’s dynamic richness and life – in other words, with this type of compressor you are gently mixing into it, not using it to control everything in your mix; there’s a huge difference here. Depending upon the nature of the dialogue, the threshold may have to be played with, but that tends to be a good starting point. What I like about this as well is that it takes already obvious transient artifacts and really makes them loud and nasty, so when you roll over them they serve as a constant reminder to eradicate them out of sheer annoyance!
It’s all fun and games until your monitors burp a disturbing thump. Pops, or plosives, are the result of air blowing along the surface of a microphone diaphragm from the mouth, wind buffets, mechanically-transferred movement such as boom pole handling noise, or random production noises being picked up such as bumping a box an errant footstep in the background. Pops is a pretty diverse category, but it tends to all fall under a definition of being a very loud, yet very short, burst of sound energy usually defined by the presence of very visible low-frequency (wide) modulations.
Due to their errant nature, they are not bound to ‘rules’ as some of these other artifacts are such as those closely related to the human voice. Pops can occur anywhere, at any time, yet luckily serve as more of nuisance than an actual headache. They are visually very noticeable, looking almost like a seismograph reading of a tremor, with frequencies usually extending so low that the period of the wave is exceedingly noticeable. Pops, Sample 1 clearly shows this – but this one is actually a little trickier to spot visual became it overlaps desired content of which does have a low-frequency sound, as seen by the next unaffected blip to the right. What is noticeable is the pop’s low-frequency flutter modulating the desired content, greatly affecting its amplitude in a rather erratic fashion. When heard, pops are usually very loud snaps which have a mid-frequency thickness to them, but sometimes they can extend exclusively into the upper frequency range. What differs a pop from a tick is both their amplitude/intensity difference (ticks are quieter and subtler), and that pops have a more roundish tone (ticks sound almost like that annoying click of static electricity discharge when we touch something). More often than not pops can make you jump out of your seat when listening to dialogue at reference level. I know I’ve been started a few too many times. Depending on where pops lie, over a line of dialogue or over backfill (quiet background noise), they can be handled differently, yet usually they require a bit more finesse when landing on top of lines because the more severely disrupt the harmonic modulations over a greater duration of time than do ticks/click and lipsmacks. Sometimes reversing works, other times creative stitching of like harmonic structures heals the wound they create. But in the end, these nasties usually require some extra grunt work for a Dialogue Editor than some of the other types when they don’t land in a free-and-clear piece of backfill – worse when it’s a breath or wind plosive rather than a short loud ‘pop’ sound, some of which can taint multiple seconds in duration rather than milliseconds. When this happens, the pop causes the Dialogue Editor to seek out alternate takes to slip in a replacement for the offending section, or failing that, must find a phonetic fragment from another word elsewhere in the edit to steal it to heal the problem. Personally I’ve done this many times, stealing bits and pieces of other words elsewhere to fix problems when a simple sticking up didn’t work and/or there was not good alternate take. This process requires extreme care and precision and respect for the human voice’s intricacies of infection, cadence, pacing, context, and emotion, so it’s not for the faint of heart nor is it for everyone. Although for those who have the gift of making such edits work as one of their regular tools and the patience and endurance to roll up their sleeves, the world becomes their oyster – just about anything in their edit becomes your secret-weapon tool set!
How pops be avoided in the first place? Well, it’s all about context. If we’re talking about wind plosives, using a windjammer kit (a suspension blimp, or zeppelin, coupled with a furry cover, known offhand as windjammer, sock, dead cat, and many other names) such as those made by Rycote and Rode can usually do the trick, as their suspension systems are build to stop 30+ MPH wind forces dead in their tracks from hitting the microphone diaphragm in the first place. The other type of plosive, handing movement, can be addressed in a few ways. For one, maintain a relaxed stance with your arms not locked, as contracting these muscles can result in tremors over a short period of time, which make your hands tire quick, and begin shaking, which of course leads to shaking the mic and transduction of low-frequency bumps, or plosives. With breathing, it’s a matter of not aiming the mic directly at the mouth. The sound energy of the human voice does not sit at the opening of the mouth, or even with the oral cavity. The energy and essence of the human voice sits in a sphere about 8-10 inches in front of their mouth, so this is where we want to mic – and do so at a creative angle in which the mic does not terminate at the mouth. In booming practices this means aiming downward at the actor, at an angle, so the mic terminates at their chest whist it’s line-of-sight is passing directly through this 8-10 inch sweet spot. In a studio situation, similar holds true, and luckily in controlled environments such as studios we can employ a pop filter. This is an attachment which connect to the mic stand and sits between the mouth and microphone. It’s either a mesh stocking material or perforated metal mesh/surface. The purpose of it is to fiter pops, creatively enough – it allows the sound energy to pass through, yet it’s surface area acts as a diffuser for the air energy and reflects it away from the mic for the most part. It’s not perfect, but for actors who breath heavily and/or have trouble controlling their ‘P’ and ‘B’ sounds (some have real trouble with this, others are immaculate), it can severely mitigate the issue. In production, this usually is not a problem because the mic is much further away, yet it is something to watch out for as a source of pops in studio recordings such as Narration, Voice Over, and ADR. There are many other rarefied circumstances that cause pops, including bad cables and connections (see the section Crackle), yet the above are the primary culprits of Pops and solutions to said culprits.
These little guys, are well, the sound of ones lips being smacked. The saliva present within our oral cavity causes these sounds as we open our mouths to speak or breath. In real-world situations, such as speaking with a friend over lunch, we aren’t aware of these – but a microphone is highly sensitive to it. They can sound like short little click noises, or more elaborate, multi-click squelchy noises sort of like the sound of a growling stomach but pitched up like a chipmunk voice, and in places of sibilance they can be excruciating difficult to detect because of their similar frequency makeup. Usually lip smacks are found hangin’ out just milliseconds before a word is spoken, especially if it starts with a sharp consonant, as is the case in Lip Smacks, Sample 1 – the highlighted bit is a lip smack, but the blip right beyond that is a sharp ‘C’ consonant. The other hiding place for these is at the very tail end of a word or phrase during the last modulations of vocal decay air release just before silence. They are also found mid-word when a new syllable is being sounded out in cases where the mouth is having to contort in weird ways (e.g. not a fluid transition like the word ‘owl’) It all depends on the talent (some actors don’t present the issue as much as others), and what’s being done with their mouth.
The difficult part about these guys is using correct discrimination when addressing them, as they look (and sound, when scrubbed over) startlingly like the very-important consonant articulations like ‘P’, ‘C’, ‘CH’ ‘GUH’, and ‘TH’ amongst others. Get rid of these along with your lip smacks and you’ve turned your dialogue intelligibility to mush, much like a body without the bones. I believe that because of this and the near-constant second guessing required for addressing them, lip smacks are the most difficult and stubborn of these artifact sound types out there. The one upside is that since lip smacks typically occur on and around consonants and between syllables, they can be quite easy to isolate from the sound source itself because there are no strong underlying harmonic modulations compared to those found in vowels, which inevitably means that when removing lip smacks there’s almost no chance of being on the receiving end of a weird vocal flutter or snapping sound occurring to the fixed word/syllable in question. That one little upside makes these lip smacks a lesser evil than it has a reputation for being.
Now, what can be done about it from a preventative? Well, there is no smoking-gun answer really, however having the talent sip room-temperature water before delivering their lines can help a great deal, and avoid eating. Drinking too much water or incorporating various temperature differences has an effect on saliva production and it’s consistency. And well, the ‘avoidance of food’ part pretty self-explanatory considering that saliva is crucial in breaking down food with its secreted enzymes. A bone-dry throat is no good either, because that stimulates further saliva production to compensate for the lack of lubricant in your mouth that a sip of water would otherwise provide in place of saliva. This is why you tend to see a surplus of room-temperature plastic water bottles arranged at ADR stages on a daily basis.
This is one of those interesting “3rd-wheel’ type sound artifacts, as there is no visual for it because quite frankly, it’s not even hardly visible – but it is very audible, very recognizable, and very annoying, so it’s worth mentioning. Cracking can be succinctly defined as being a random grouping of a series of ticks lasting for a defined duration of time. Crackling tends to be a more “fuzzy” sound which is attributed most of the time to either cloth rubbing against a microphone (like lavalier radio mics), or signal interference either by another offending signal source intervening or by loss of signal’s robustness amongst the white noise background (again, like lavalier radio mics). Thus, crackle tends to be one of those special artifacts uniquely tied to that type of microphone recording, and that’s why a boom is also used in production as an alternate sound source. Rarely if never have I had to deal with crackle on a boom recording, but almost without fail I’ve run into it on a lavalier radio mic recording.
Crackle is a uniquely irritating artifact for a few reasons. First, they can almost never be surgically removed by hand because they taint too much of the original sound source at too close of an interval that there’s virtually nothing to salvage. Second, almost always when crackle is run across it means that the take is usually ditched and considered ‘no good’; this is corollary to the first reason. The more broad-band the crackle is, the more non-salvageable it becomes. Thirdly, if you use an automated tool to de-crackle the sound, there will very likely be a loss of fidelity of the higher-frequency content which in effect reduces intelligibility and clarity of dialogue at a tradeoff of less or no crackle. This all being said, there are two tools worth considering because they do a decent job in a pinch and can actually salvage lightly-crackled material quite well without sacrificing very much high frequency content, if any at all – I have experienced this within my own work in the past, although nowadays I tend to first opt for an alternate dialogue take and use de-crackling as a last case scenario. These tools are Waves X-Crackle and iZotope RX Decrackle. Again, they do a pretty decent job. Sometimes it’s a matter of running multiple passes with the tool, wherein you begin with an aggressive pass and then follow it with subsequently gentler passes. But in the end, the only real way to win against crackling is to not play the game to begin with. Seek out alternate takes that are clean, and go from there, it’s almost always worth following that route instead.
What can be done to prevent crackle in the first place? Since it largely an issue revolving around lavalier radio mics, more care can be taken into properly using them and/or protecting them during production. For the ever-present cloth rubbing issues that is the source of much crackle problems, many prevention method exist utilizing a combination of surgical tape, mic clips, moleskin, and creative placement. Smaller radio mic varieties such as the Countryman EMW or B6 are tiny, and this makes it easier to hide with less surface area being in contact with the mic. Ultimately, the lesser the surface area the mic and its fixtures are coming in contact with to anchor it, the less cloth noise issues there will usually be. I have been told the best places to mic are in the divot just under the sternum or even mid-sternum – the both of which are because when the torso is rotating, this area mains solidarity thanks to the underlying bone structure. Because of that, the chances that cloth will be pushed and pulled over these non-deforming surface mans that the cloth holds its place for the most part; when the cloth is holding its place, it doesn’t make movement, and that translates to a recording free-and-clear of cloth crackling. Another major source of crackling deals with RF interference and/or operating range between a radio mic transmitter and receiver. Some models such as the Sennheiser G2, from the appropriately-named series EW (eww) are notoriously bad in this department. Their range is quite terrible and the signal is very prone to RF problems. I cringe at the thought of working with dialogue shot on these, only because they are consistently poor in build and quality. Many past dialogue edits over the years working with source from this radio mic series has left a sour taste in my mouth about these to say the least. However, Lectrosonics on the other hand capture signals which are usually a delight to work with. They are almost always clear of RF problems, the signal is very strong and crisp (when the signal chain is calibrated properly), and they the have superior range capabilities required for production purposes where the sound mixer is many times nowhere near the actors. Anyone who’s serious about production sound mixing uses it, it’s just that good. A third source of crackling to consider is sometimes XLR cable problems or loose XLR connectors, wherein the wiggling back and forth causes the voltage current to break up, resulting in digit crackling. Properly maintaining your gear and cables can help mitigate this, as the connectors are especially prone to the same type of snap crackle pop issues that vinyl records are susceptible too with regards to dirt and grit. Compressor air can blow this stuff out quite easily, so it’s always good to have a spare can around (and most sound professionals probably do!). On occasion, manufactured connectors are defective in size which result in no tight contact. In this case, it’s better to contact the manufacture’s service department or replace the connector yourself if you’re up to the soldering challenge and who what you are doing. Whatever the choice, these kinds of problems are best dealt with by nipping it at the source rather than using tape or some other means to temporarily “correct” the issue. Not only are they going to usually still further produce the problem you’re trying to mitigate in the first place, but sometimes it can be dangerous. With XLR cables themselves, I recommend either using either of the two tradition “looping-up” methods, the over-under or the half-twist, however absolutely do not fold the cables akin to how AC power cables are folded in packaging. It’s the quickest way to damage the internal wiring by sharply kinking it. It will lead to crackle, clicks/ticks, or even worse, no signal – and just at the perfectly wrong time too! What’s more, these same preventative measures apply to preventing the other mention sound artifacts. So do yourself a favor by doing your dialogue editor the favor of capturing the best possible sound you can with gear that is properly maintained at all times and utilized with sound techniques and practices. Your dialogue editor will thank you!
These sound artifacts suck, they really do suck indeed. They can ruin your day, waste your time, and piss you off sometimes. Although with patience , some elbow grease, and a little flair for problem-solving, they can be eradicated. And it is done so on a daily basis with extreme precision and tactfulness by the many talented Dialogue Editors and ADR Editors within our industry. These editors are very important, as it has been said to me, “they are worth their weight in gold”. And rightly so. Stop to think about it for minute. Where would the likes of Academy Award-winners Brad Pitt, Meryl Streep, George Clooney, Julia Roberts, and all others before and since be if we heard them unattractively smacking their lips throughout a film? Think about it for a moment, and you’ll realize how gravely under-appreciated yet important Dialogue Editors and ADR Editors truly are.
Happy wrangling to all, and to all a good night!