There are many "tools of the trade" available these days for cleaning up, or spicing up, a recorded performance. Some of these tools/tricks/techniques have been around since the early days of analog tape recording and some are specific to the newer world of digital recording. This FAQ section is really designed to be a sort of overview or primer for people with little or no experience or understanding about such things. If nothing else it'll give you an idea of what's possible these days. I'll start with a quick overview of some of the more popular tools/tricks/techniques and what they do, and then I'll end with a bit of philosophical discussion about the use of such studio trickery.

1. Auto-Tune and vocal pitch correction.

In case you didn't know, it's possible to take a vocal recording and improve it, or change it, by fixing notes which are either sharp or flat in pitch. The most popular tool designed to do this type of work is a software program called Auto-tune (AT). AT was one of the first such products to market (somewhere around 1997 I think) and immediately became an industry standard for vocal pitch correction. There are a number of products on the market now including vocal hardware processors which are designed to work in real time and are used everyday for live performances.

The AT program has two primary modes of operation: 1) automatic and 2) graphical. In automatic mode you basically set it and forget it. It does it's thing in real time as the music is playing. For the most part, automatic mode works ok, but it isn't perfect. Occasionally, AT will make the wrong decision when correcting pitch. For example, if a note is just a little too sharp or too flat, it might try to pull the note in the wrong direction (making it more sharp or flat). This tends to be worse than no pitch correction at all. 

For the most natural sounding results, it's usually better to work in graphical mode. Graphical mode is more of a "hands on" surgical approach which allows you to make pinpoint corrections to the recorded audio waveforms. It takes a lot longer to make adjustments in graphical mode (especially if you're dealing with many vocal tracks) so the advantages and disadvantages of both methods need to be weighed when doing this sort of work. When working in graphical mode, you have to first "encode" or "record" the audio that you want to work on. This is done in real time which means that I have to first play the section of audio that I want to work on. If the entire vocal track needs work than I need to encode the entire track. Multiply that by the number of tracks and... well... you get the idea. Once the audio has been encoded or recorded into AT, I can then set about the task of correcting the performance. This process can be rather tedious and painstaking depending on quality of the original performance. 

AT only works on monophonic source material - meaning music containing one note played or sung at a time. Which makes it a perfect tool for solo vocal tracks and lead guitar or bass guitar notes. AT won't work on polyphonic audio (many notes played or sung at the same time). For that I would need Melodyne (which I don't currently have). 

Samplitude also has some nice pitch-shifting/pitch correcting tools so I have a few different solutions on-hand. If one product doesn't produce the results I want, there is usually another alternative available.

I find it's usually better to do this type of work at the very end of the project because there is often a lot of re-recording or editing going on with the vocals.

2. Vocal doubling or vocal harmonization.

There are plenty of products on the market these days designed to do all sorts of interesting and strange things to a vocal performance. Many of these products are designed for stage use and work in real time during the course of a live show. Vocal doubling, pitch correction and automatic harmonization are just a few of the things that you can do. 

Let's talk about vocal doubling first. People ask me about doubling vocal tracks all the time. This is a very old studio trick. Vocal doubling is pretty much what the name suggests. A singer will sing a lead vocal line and then go back and sing it again. The second track is mixed in with the first track (usually lower in volume and perhaps with some additional effects processing) resulting in a thicker, richer sounding lead vocal. Mixed and edited properly, often times you can't even tell that there is another part. Yet the vocal sounds somehow fuller. So the concept of doubling a vocal track is just a way to add fullness or depth to a voice. The best way to do this (as far as I'm concerned) is to simply sing the same part over and over again. Then you can take those alternate "takes" and blend them under the main lead vocal line. You can do this with with more than just one or two tracks. I've had some singers come in and sing the same part many times and then we mix them all together. You have to be pretty good to do this because in order for this to work, your timing and phrasing has to be right on. 

Even the best singers in the world can't sing every single word and syllable in perfect synchronization from take to take so there's usually some editing needed on the secondary tracks to get all the words and syllables to line up. It's fairly important that each word or phrase start and end at the same time. It gets worse and worse with the more tracks you add underneath the main track. It can get muddy and sloppy real quick. So I will go in there and comb through each track - word for word, syllable by syllable, and surgically edit each little bit to make sure the parts line up with the main track. Sounds like a lot of work right? It is! It can take hours to comb through a song's worth of vocal tracks and tighten them up. But in my experience this is the best way to get the job done. It always seems to produce the most natural sounding results to my ears. Once the individual takes have been time-aligned and tightened up, I usually EQ them differently and add some additional effects to those tracks. Basically I don't want a build-up of frequencies which can produce harshness or muddiness. 

The various digital solutions produce good results but to my ears, they all have a sort of artificial quality about them that may or may not work with the style of music you're recording. I have some plug-ins which are designed to create a secondary doubled effect. Dialed in and blended in correctly, you might not be able to tell it was a plug-in. It takes a bit of work to make the software plug-ins sound natural.

Simply cloning the lead vocal track (creating an exact replica of the lead vocal and copying it to other tracks) does nothing but make the original lead vocal track louder. In order to create the illusion of a secondary track (which adds the fullness or depth we talked about), you need to offset the cloned track by a few milliseconds and maybe add some chorus or delay effects to it. There are other tricks such as pitch shifting the alternate tracks up or down a few cents. Again, these techniques tend to produce an artificial sound that might not work with the style of music you're recording. If you have some old Beatles albums go and listen to them and listen to how this effect worked on Lennon's voice. Lennon disliked having to sing the same parts over and over again so the boys at Abby Road came up with a trick using two tape decks. 

Digital vocal harmonizers have been around in many forms for a number of years now. Vocal harmonization has the same inherent problem as vocal doubling - no matter how hard you try, the digital solutions are always going to have an artificial sound quality. They've gotten real good with the tools to do this in recent years but it's not 100% there yet in terms of believability. The human voice just has too many nuances to effectively duplicate with artificial means. It's possible to create 2, 3 or 4 (or more) part harmonies with some of these tools. The fact they work at all is pretty amazing. I'm sure most animated films use these tools to create unique voices for the characters.

3) Drum triggering or sample replacement

There is a program is called Drumagog and it's used to change the sound of a recorded drum set by replacing the sound of each individual drum hit with a different sounding drum hit. Works on all the drums (snare, toms, kicks, even cymbals). Any sound with a short, sharp transient peak can be replaced with something else. I can make a snare drum sound like a kick drum if I want (don't know why I'd want to but it's possible :-) And of course, there are many different electronic products such as drum triggers or electronic drum pads.

Drum triggers are used all the time in live performances to embellish the sound of a drum set or add special effects (hand claps or virtually any percussion sound). A trigger can be either a small device which sits on the head of the drum or an actual rubber pad which looks like a drum head and is attached via MIDI cables to a synthesizer/sampler unit. A lot of the nu-metal guys are using them these days. It's a lot easier to get a killer drum track this way. The alternative is a lot of work and a lot of gear that may, or may not, produce the same effect. The downside to drum triggering is the ability of the triggering mechanism to accurately track each and every drum hit (Drumagog has the same problem). Drummers play with a lot of nuance (believe it or not :-). Small ghost notes are difficult to track properly which makes drum triggering/resampling difficult or impossible for certain styles of music. Your playing technique has be adjusted to fit the use of drum triggers. Sharp, well-defined notes are imperative. The faster you play, the more likely the triggers aren't going to track each and every hit because you're not hitting the drums as hard and it takes a little time for the triggering mechanism to rebound to a ready state.

With Drumagog, I can deal with ghost notes and a lot of other nuances but all this stuff takes time to implement properly. The bottom line here is, if you want a natural sound using samples, it takes time.

4) Quantization

Quantizing is the act of time-aligning a performance to match a steady beat or tempo. Most musicians, even when playing to a click track, aren't going to get the notes to be in the perfect place all the time. It's impossible for a human being to play with digital word-clock accuracy. You might be on the beat some times and slightly in front of, or behind the beat at other times. A really good musician with really good timing can play "close enough" on the beat where your ears probably aren't going to detect an ever-so-slight change in tempo. Musicians call this playing "in the pocket". You really can't tell unless there is something else out of time with that performance like a hi hat and a snare drum not locked together. Or a bass player not locked in with the kick drum. Or a guitar player not in time with the bass player. If no one in the band is playing in perfect synchronization with each other, the song tends to sound loose and sloppy (which is a good thing sometimes). But for certain styles of music, it's important that everything lines up on the beat. Quantization to the rescue. Most modern recording programs these days have tools to do this type of work automatically (or at least make the job a lot easier than traditional cut and splice).

4) Compression, EQ, reverb, digital effects

These effects have been around almost since the beginning of time. I'm not sure I would call compression and EQ an "effect" but they can certainly be used in a way that creates the illusion of an effect of some sort. I tend to refer to compression and EQ as "sound-sculpting" tools that are used to get the different tracks to blend well with each other, or to correct for certain audio problems. Compression is a tool that is used to make soft sounds louder and loud sounds softer. It basically removes the wide swings in volume on a track and helps the track sit in the mix better. A singer, for example might have sung some words very softly and those words can get drowned out in the mix. If you add a bit of compression to the track, those softer words will be louder and better heard. EQ is used to add or remove certain frequencies from a recording. Most people have a basic understanding of what EQ is (most modern stereos have a bass, mid and treble control) so it's easy to adjust, listen and understand what the controls do. Reverb adds the impression of "space" to a recording. Without it, the recording will sound very dry and unflattering. And there are tons of digital effects on the market for adding all sorts of wild sounds to a recording. The sky is the limit with this stuff. Unless we're going for some sort of psychedelic sound, I tend to use digital effects sparingly. A little chorus can add fullness to a vocal track for example. Too much can make it sound like, well... something from outer space :-)

5) Copy and paste

No mystery here. We copy a good section of the song and paste it over a not so good section of the song. I do this type of editing all the time. This process tends to work best when the song has been played to a click track and has a stable tempo from start to finish. If your tempo is fluctuating all over the place, then copying and pasting sections of the song might not work because they don't fit the tempo.

6) Brick limiting - loudness maximization

Modern rock and pop music has been mastered in a way that removes all the dynamics from the original recorded performance. Dynamics are the natural swings in volume that occur in most music. One minute you might be listening to a flute part which is playing a gentle and soft passage and the next section might be a screaming guitar solo with crashing drums and thundering bass. The current state of the art is to create a "loud master" which basically removes all the dynamics of a recorded performance and replaces it with... loudness. You can read my opinions on this below but if you want a good example of what this means, pull out one of your CD's from the late 80's or early 90's and compare it to anything released today. You might notice that the volume level of the older CD's is about half what they are today. People equate louder as being better - even if it really means that the music is being damaged in the process.

7) Pitch shifting - time stretching

Closely related to Auto-tune but different. Instead of pinpoint surgery we're talking about "overall" change. For example, sometimes someone will ask me to change the pitch of some music because the key of the music is out of their singing range. Karaoke singers have this problem a lot. Most karaoke decks have a built in adjustment to shift the pitch of the music up or down while keeping the tempo the same. Yes I can do this easily but there is a caveat which I'll explain in a second. I can also change the tempo without changing the pitch. Maybe the song was recorded a few beats too slow or too fast. Easy to change it, but again, there is a caveat.

The caveats are this: this type of work has an adverse effect on the quality of the audio. The amount of change is the issue here. A few degrees (or cents) of change probably won't cause much noticeable damage but the farther you go in any direction, the more noticeable the degradation. The damage will usually manifest itself as a sort of unpleasant digital "warbling". This isn't tape we're talking about where it's fairly easy to speed up and slow down without much damage. This is digital. There's a lot of math involved. So the quick answer to the question is yes I can. The harder answer is, it depends.

The tools I have available to do this type of audio manipulation work pretty well, but they aren't as good as say, Melodyne (which I don't have). I don't get enough call to do this type of work to justify the cost of Melodyne. I've somehow managed to survive this long without it so that gives you an idea of how often this issue comes up. Reasonably "slight" changes are easy. Anything beyond "reasonably slight" is risky. The damage to the music might render the music useless for the intended purpose.

8) Vocal removal or vocal elimination

I think I make reference to this in the Karaoke FAQ section too but I'll mention it again here. Sometimes, under very limited and precise conditions, it is possible to remove a lead vocal part from previously recorded and mixed commercial music. Karaoke singers ask me to do this when they can't find the song they want to sing released in a normal karaoke version. The problem here is that the techniques used to make this happen usually destroy the music so badly that the resulting audio is useless for anything other than casual practice. I've seen the process work amazingly well but that is very rare indeed. The original recorded song has to have been mixed and mastered in a way that facilitates this process, and that hardly ever happens. The tool/technique works under the assumption that the lead vocal has been mixed dead center in the mix. If you take the stereo signal, invert one channel 180 degrees, supposedly anything in the center will be canceled out. The problem is, most vocals occupy the same frequency range as the guitars and various other instruments in the mix. Generally, you can't remove just the vocals without removing something else in the process. Given enough time and tweaking, it might be possible to zone in on "just the vocals" but again, this is very song-specific and most likely, there will still be enough damage to the music to make it worthless for anything serious. I've tried it often enough to say this - don't ask. I don't even want to bother trying. It's a waste of time.

Having said all that - here is a link to a website with examples of a product that supposedly works. As you can hear from the examples, the process worked pretty well. My guess is that they chose the sample material wisely. The target material contained a single mono voice with little or no effects processing and little or no backing vocals or other processing. Under these sort of narrow parameters, it's possible to achieve good results. If you're inclined to go out and buy one of these devices, be my guest. Here's another device. And another. And another. And more information about the process here.

Are these things tools of the devil?

Not hardly but if you do a little searching around the various chat forums devoted to audio engineering and production you'll likely find some heated discussions and debates related to the use (or overuse) of some of these tools.

Let's talk about Auto-tune to start with. If there's any one particular product that has gotten slammed the most it's probably Auto-tune. And yet almost every song you hear on the radio these days has been treated with Auto-tune. Especially in hip hop and urban R&B. I can hear it a mile away. The problem, in a nut shell, is that no one sings like that normally. To get a life-like and transparent result with Auto-tune you need to really be careful how you adjust the product. I think there is a tendency to overuse the product in places where there doesn't really need to be any correction in the first place. It's one thing to salvage a performance with a little nip and tuck here and there, and quite another to just get slap-happy with the tool and suck all the soul out of the performance. Having said that, Auto-tune is the product that was used to create that funny "Cher effect" on  her hit song "Do you believe". That funny robotic sound on her voice is the result of adjusting Auto-tune in a way that is used as an interesting and artistic effect. That record sold millions. Who's to argue with success?

I will state flat out that I'm not a big fan of teflon-coated music. Too much of it tends to sound like elevator music to me personally. But this may be a sign of the times in fact. I grew up in an era where people had to actually be able to sing, and sing well, in order to sell records. These days, you don't need to be able to sing that well to sell a million copies. That's not to say you can get by with no talent whatsoever. The irony of Auto-tune is that you really do have to be able to sing fairly well in order for the tool to do it's job. You need to be close to the pitch you're trying to hit otherwise you'll end up sounding like Mickey Mouse or Donald Duck. It can take a horrible amount of time to spit-shine a vocal track into perfection so time and money come into play here too. Multiple that by the number of tracks on the album and there ya go. And let me state that Auto-tune is not going to help you one little bit if the performance lacks excitement or passion or whatever you want to call it. A poorly sung vocal track that lacks inspiration might be perfectly in tune but it will still sound like shit. Know what I mean?

Auto-tune tends to be a slippery slope. What I mean is, once you start down that path it's hard to stop. It's rare that a singer will sing all the way through and only need a slight tweak here and there. Usually, there are plenty of rough spots - or spots that need some sort of attention. Once you start tweaking, it brings focus on other weak sections. This is especially true on backing vocals. If you pitch-correct the main track it will make the backing vocals not sound right so now they have to be pitch-corrected too. Pitch-correcting one word in a verse will usually bring attention to other words nearby that aren't right. The slippery slope. Once you start down that path it tends to be all or nothing.

The same slippery slope applies to quantization. Quantization can make a song sound unnatural because it wasn't meant to be played like a machine. Speeding up and slowing down is a stylistic effect and adds drama and excitement to music. You probably don't want your songs to be speeding up or slowing down in a dance club though. So you need to be careful where and how much of this wizardry you apply. Trying to quantize a chorus section might bring focus on the verse right after it. Here again with the slippery slope. Once you start it's hard to stop. And if one song is in perfect alignment but the other songs aren't? I don't know. It doesn't sound right as a "big picture" sometimes. We're more of a "singles-based" market right now so who knows or cares what you use and when I guess. I like to mention this to bands who are in the process of recording the tracks: In my mind, it's usually always better to re-record the entire song if it appears like there is going to be a lot of time involved in fixing the timing. I can easily spend hours trying to time-align a recorded performance. Compare that to the 5 or so minutes it took to record the song in the first place. Don't forget. Time is money :-)

The business of loudness maximization is starting to become a critical problem in the industry. Consumers are starting to notice the problem more and more and are fighting back. It's a sort of dilemma because people (who don't know better) tend to equate loudness as being "better" when in fact, it's actually worse. Removing all the dynamics from a recorded performance - I can't even think how or why anyone would think that's a good thing to do except for radio stations who want to make sure that you hear everything in your car, even with the top down. I mean, musicians play music with different dynamics for a reason sometimes. When you see a band play live, they usually play soft during certain sections and loud during other sections. When you listen to people speak, they don't TALK LIKE THIS ALL THE TIME. I think you would probably get annoyed being in the same room with someone like that and eventually leave :-)

Brick limiting is actually a form of audio damage because in order to get the volume of the music up to today's levels, you need to clip the tops of the waveforms which causes digital distortion (clipping). Do a search on the latest Metallica CD and most likely you'll find something in regards to the backlash concerning the sound quality of that CD. I recently pulled out one of my older Van Halen CD's and was surprised at the way it sounded compared to stuff being released today. The volume was about half and the music seemed to lack the stylistic punch of today's modern sound. And yet, to my ears, I preferred the older, "less loud" mix because all the dynamics were intact and there was no digital clipping noise.

As far as all the other trickery, these are just creative tools used to make a statement. I don't pass judgment on the use, or non-use, of any of this stuff because there is a time and place for all of it. They're just tools. Like a paintbrush or camera. If it helps the song along, then it's a good thing. If it's not helping the song, then perhaps, it's not the right tool for the job. You have to be careful with effects sometimes because they can date a song. If you want your music to stand the test of time, sometimes it's better to stick with more natural production methods. I think that's one of the reasons why Aerosmith and Led Zeppelin and bands from the 70's are still successful today. Their music still sounds pretty fresh. But some of that "freshness" was grandfathered in along the way with up-to-date remastering. 

Will anything we release today still sound fresh 20 or 30 or 40 years from now? Who knows. I was in a Borders book store not too long ago and I happened to pick up an Audiophile magazine that had some record reviews in the back. They specifically rate their album reviews on whether the records (ok, CD's) have been brick-limited to death or not. CD's which have retained their natural dynamics get higher points than the latest Metallica CD which sounds like an overly loud distorted mess and I don't mean that in a good way :-)

Personally, I think with any of this stuff you need to take it all in perspective. Like I said. They're just tools. If using Auto-Tune means you don't have to sing the same part 20 times until you get it right then maybe that's a good thing. Use the time you just saved to start working on the next song. If quantizing your songs so they have a steady tempo means the difference between getting that coveted weekend paying gig at your local pub or playing at open mic night on Monday night for free, then do whatever it takes to get your foot in the door. At some point though, you're gonna have to actually get up in front of people and try not to make an ass out of yourself so think about that too. Of course, most people are there to just drink and cruise for chicks and aren't paying attention to the fact that you're flat every other note. If they can dance the night away and the bar is selling lots of beer then you're golden.

Back to LnL Recording FAQ