How to Master your finished audio recordings The following describes a rather extensive process for mastering songs to CD. The mastering process is often a long one, and musicians spend lots of hours and dollars to achieve that 'commercial studio' sound. With a whole lot of practice, a good set of ears and the following step-by-step tips/guide, you are preparing yourself to enter the realm of professional music making. Where applicable, some details are listed regarding suggested frequencies or 'settings', but please note that each song is different, so use your judgment and above all, use your ears! 1. Sample Rate/Bit Rate:: If you've already recorded at 16-bit and you've remained in the digital domain, there isn't really any reason to bump up to 32-bit unless you have extensive noise reduction to do. Working in 32-bits to simply add some minor EQ and/or Limiting isn't really going to benefit most recordings (apart from some extended dynamic range) particularly if you are unfamiliar with the noise shaping curves and dithering selections, but it can't hurt, and if you have the HD space...why not?. Certainly, if you begin at 24+ bits, stay there until the final stage. 2. Pre-Dynamic EQ:: This relates to whether or not you are going to use some type of limiter (as per the Hard Limiter) or mild compression on your final mix to give it that much adored PUNCH! If you plan on doing so, you may consider tweaking EVER SO SLIGHTLY some key frequencies that may be accentuated in the compression process. Key frequencies are: 31k (deep bass) 32k (boom) 60hz-80hz (bass, 45hz and below for sub-bottom) 125hz (kick-drum, thump) 250hz (muddiness/low mid) 1.25k-2.5k (vocal fundamental, clap, snare) 4k (sibilance) - 1. A consonant characterized by a hissing sound (like s or sh) 2. Produced by forcing air through a constricted passage (as 'f'. 's', 'z', or 'th' (in both 'thin' and 'then')) 5.8k/6.3k (sibilance) 8k-14k (presence and shimmer) By boosting or cutting (at very small increments) you can really tailor your sound like a pro, using a variety of narrow and broadband Q. Don't boost/cut more than 3dB in any one band, except perhaps in the 250hz range if your sound is very muddy. 3. Limiting/Ceiling:: Make it loud. That's what everybody wants. With the hard limiter, you can set the "Limit Max Amplitude" to -.5dB. Use the 'Boost Input By' to really drive your signal. Be careful not to drive it too hard, as you will start to hear audible artifacts, though they are sometimes pretty cool effects. You can also use a simple compressor from within the Dynamics Processing menu. A compressor preset with a ratio of 2:1 is pretty standard for giving a mix 'punch.' Typically, you would not exceed a 3:1 ratio. 4. Stereo Expansion:: Totally optional. Lots of mix engineers use expansion already in the mixdown process; ie, your background vocals may be mixed with some stereo reverb, the drums may have some type of presence/ambience effect, electric guitars bouncing left to right. All of this can sometimes create a wash of stereo leaving you with no center at all, so use with caution. If you are simply trying to WIDEN a rather narrow stereo field, try using the Pan/Expander to give you a bit more depth and dimension. You can even move the center image around, but be careful, as that can often 'unbalance' your already perfect mix. 5. Post-Dynamic EQ:: See Pre-Dynamic EQ. You may not need any. If you didn't use any before the limiting stage, use lightly to do the final tweaking. Very minimal boost and cutting; just shaping. 6. Normalizing:: Now, if you've hard limited with a max amplitude of -.1, you don't have much room to work with. However, if you've added/subtracted EQ, you may want to boost the signal to peak amplitude. Some say normalize 98% just to be safe. There is really no harm in normalizing to 100%, so long as you are using a good CD burner and adequate software. The levels, as far as Cool Edit, should be painstakingly accurate. 7. Convert back to 16-bit, 44100, Stereo. The preset dither depths and noise shaping curves (pdf=triangular, Noise Shape A) are pretty good, but not necessarily for 32-bit 96k audio. You may want to experiment on shorter files before you find one that best suits your music. There is an explanation in the help file that better defines what curves work best with various sample rates. 8. Saving as Windows PCM Wave and doing it right:: When preparing your song, you want to make sure that when you've made your wave files that you have about 100-150ms of space BEFORE the beginning of the song. This will ensure that all CD players will capture the 'startID index' and play the beginning of the song properly; this value has actually gone down now, as newer CD players need very little (if any) space before the beginning of the track, but it's always a good idea; and what's 1/10 of a second anyway? Of course, if it's a live recording and there is in effect 'no stopping', this is not so much of an issue. But you may want to experiment. 9. Using the CD Burning Application:: The first track on your CD must have a 2-second pause. This is related to the CD burning software itself, not something you do within Cool Edit or to the actual wave file. This 2-second gap is written right after the table of contents on your disc to ensure proper indexing and playback. After the first track, you don't need to have any pausing at all. And, if your software/CDR supports DAO (disc-at-once) it will set up this default for you. 10. Burn baby, Burn (speed):: Always burn at 2x or slower (1x, realtime burning) if it is for duplication. 4x is fine for personal use. The reason being, you will inevitably have fewer CRC/burst errors, and most plants can detect what speed your disc was burned at and won't even accept a CD premastered-cut at 4x or above...too many errors, too many chances for error during duplication. You also want to use a higher-grade CDR. There are many arguments as to what 'dyes' work better, etc. Most plants prefer the green or gold dyes over the blue...but you'll want to check that out first. The other, most obvious things to consider when mastering are that in this final 'make it or break it' stage, you are solely relying on one thing and that is your ears. Try (if you can) a different set of speakers, use speakers and headphones (ones that you know the sound of pretty well) to A-B and listen for boominess or tinniness. Make a 'reference' CD of a few tracks and audition it on a car CD player, boombox, home theatre, whatever you have available. And be sure to give yourself a break! Ear fatigue can set in fairly rapidly during this process, so don't try and 'beat the clock'...make sure you have ample work time but also adequate break time. Mastering is truly an art form, and to 'master' the artform itself takes lots of work, persistence and time. Rock On!! Dynamics Processing Term Meaning Attack Time The time it takes to respond to increases in the input signal level Compressor An audio effect that reduces the dynamic range of an audio signal dB (decibel) a unit of measurement that expresses the relative levels of two electrical voltages Dynamic Range The difference from the highest and lowest levels of a varying audio signal Expander An effect that increases the dynamic range of an audio signal, by reducing the signal level any time it drops below a specific threshold Gain The amount of amplification applied to an audio signal Gate A device that increases the dynamic range of an audio signal by cutting off a signal when its level falls below a specific threshold Hard Knee An input/output curve that changes slope suddenly at thresholds Harmonic Distortion The addition of harmonics that were not present in the original signal, due to changes in the shape of the waveform Limiter An effect that prevents an audio signal from exceeding a specific level Noise Gate See Gate Ratio The factor by which gain exceeding the threshold is reduced or expanded Release Time The time it takes to respond to decreases in the input signal level Soft Knee An input/output curve that changes slope gradually around thresholds Threshold The level at which the compressor or expander starts to operate Transient A momentary high peak level Waveform A graphical representation of an electrical signal Dynamics processors are a class of audio effects that modify the dynamic range of an audio signal. The dynamic range of an audio signal is a measure of the level of variation between the loudest and softest parts of the signal. The volume, or level, of a digital audio signal is normally measured in units of decibels, or dB. When working with 16-bit digital audio, the range of values for signal level ranges from approximately -96dB (silent) to 0dB (maximum volume). The decibel scale is logarithmic -- each increase of 6dB (say, from -96dB to -90dB, or from -36dB to -30dB) represents a doubling of the signal level. There are several different types of dynamics processors. A compressor takes the loudest parts of an input signal and reduces their volume. A compressor only affects those portions of the signal that are louder than a certain level, known as the threshold. Portions of the signal that exceed the threshold are reduced toward the threshold by a set factor, known as the ratio. Most compressors also raise the level of the entire audio signal by a certain amount. This amount is known as the master gain. Both the threshold and the master gain are measured in decibels. The compression of the loudest parts of the audio signal helps to make sure that the average signal level can be boosted without exceeding the maximum allowable peak level. In general, the master gain should not exceed the reduction in dynamic range that results from the compression. If it is larger, clipping may result. Compressors are usually applied with compression ratios ranging from 1.5:1 to 8:1. An as example, suppose that a compressor is set with a threshold of -18dB and a ratio of 3:1. In input signal of -12dB (6dB above the threshold level) will produce an output signal of -16dB (only 2dB above the threshold). The amount by which the signal exceeds the threshold is reduced by a factor of 3. When the ratio is much higher (around 10:1 or more), a compressor effectively prevents the audio signal from exceeding the threshold value. In this case, the effect is referred to as a limiter. An expander takes the softest parts of an input signal and reduces their volume. An expander only affects those portions of the signal that are softer than a certain level, once again known as the threshold. Portions of the signal that fall below the threshold are reduced below the threshold by a set factor, known as the ratio. When the ratio is high, an expander effectively eliminates all portions of the audio signal that fall below the threshold. In the extreme case, this is known as a noise gate, or simply a gate. When a signal exceeds the threshold level of a compressor, or falls below the threshold level of an expander, the signal is modified. The attack time is the length of time it takes the dynamics processor to respond to increases in signal level. The release time is the length of time it takes the dynamics processor to respond to decreases in signal level. In general, attack times are set to be relatively short, while release times are much longer. This means that compressors and limiters generally activate quickly and release gradually, while noise gates close slowly, and then re-open quickly. By adjusting the attack and release times, you can produce smoother changes in signal levels that might otherwise change abruptly. If the attack time is too short, you may experience unnatural or sudden changes in volume. If the attack time is too long, the dynamics processor may not have the effect you want. If the release time is too long, compression may continue through low-volume sections of audio, making them inaudible. All dynamics processors monitor the input signal to estimate the current signal level. The estimate is made using one of several common level detection methods: Level detection method When it is useful Peak Most useful for limiting Average Most useful for instrumental solos RMS Most useful for vocals and speech Dynamics processor The stereo dynamics processor plug-in combines the functions of all the other dynamics processors in a single unit. If you use the graph to control the settings, note that a maximum of four nodes are allowed in the graph. Setting Range of values Compressor Threshold From the expander threshold to 0dB Compressor Ratio From 0:1 to infinity:1 Expander Threshold From -108dB to the compressor threshold Expander Ratio From 1:0 to 1:infinity Attack Time From 0.1ms to 200ms Release Time From 20ms to 1000ms Master Gain From -24dB to +24dB Level Detection Method Peak, average, or RMS Stereo Handling Maximum or Side Chain Envelope Curve Soft knee or hard knee Compressor/Gate A compressor takes the loudest parts of an input signal and reduces their volume. This effect includes the following parameters. Setting Range of values Compressor Threshold From the gate threshold to 0dB Compressor Ratio From 0:1 to infinity:1 Gate Threshold From -108dB to the compressor threshold Gate Ratio From 0:1 to infinity:1 Attack Time From 0.1ms to 20ms Release Time From 50ms to 3000ms Master Gain From -24dB to +24dB Level Detection Method Peak, average, or RMS Stereo Handling Maximum or Side Chain Envelope Curve Soft knee or hard knee When to Use a Compressor/Gate There are a variety of reasons why you might want to use a compressor: · Compressors enable you to increase the overall signal level of a mix, without causing the distortion that would result from exceeding the maximum allowable peak level. Increasing the average signal level makes a recording sound louder, with an apparent increase in “punch.” · If you plan to reproduce your audio using a medium with limited dynamic range, such as magnetic tape, the increase in the average signal level that you achieve with a compressor helps prevent important audio material from being lost in the background noise. · Compression can even out unwanted volume changes to achieve a smoother and fuller sound. For example, if a vocalist moves closer in and further away from a microphone while making a recording, the recorded volume levels may be uneven. A compressor can smooth out theses variations. · During mixdown, you can use a compressor to adjust the dynamic range of individual tracks. This is one way to balance the tracks, and can even reduce the need to use a lot of equalization. · Compressors can be used to increase an instrument’s sustain, by using a release time longer than the instruments decay. · With extreme compressor settings, you can produce new interesting sounds from familiar sounds. When you use a compressor, the master gain adjustment can result in increased low-level system noise. As a result, it is common to use a gate in combination with a compressor. A compressor/Gate combines a compressor and a gate into a single effect, so you can increase the overall dynamic level of an audio signal, without distorting the loud parts, and without excessively boosting low-level noise. The gate threshold is normally set low enough to remove low-level system noise, but not high enough to remove important program material. Since other audio effects can produce background noise, you normally place the noise gate near the end of the effects chain. However, if you’re using delay and reverb effects, you may want to place them after the noise gate, so that the sound trails off naturally rather than being truncated suddenly by the gate. Here are several common problems you can run into when using a compressor: Problem Explanation Breathing..........If the audio material has a wide dynamic range, the output gain of a compressor can vary greatly. The greatest change occurs immediately after the input signal crosses the threshold. If the input signal fluctuates around the threshold, then the output level will increase and decrease repeatedly. The resulting rise and fall of low-level background noise can result in an annoying “breathing” effect. To prevent this effect, use a lower compression ratio, or increase the release time. Pumping...........Pumping can occur when you compress a mix where one instrument is louder than all others. Whenever the loudest instrument is silent, even for a brief moment, the overall level of the other instruments will increase dramatically. As the dominant instrument comes in and out of the mix, the compressor will increase and then reduce the level of the other instruments. This results in an undesirable “pumping” effect. To prevent this effect, decrease the level and/or dynamic range of the dominant instrument in the mix, increase the threshold, or use a lower compression ratio. Noise...............Set the gate so that the noise floor is not boosted significantly during quiet passages. Harmonic Distortion.....If the release time is too short when compressing or limiting lowfrequency signals (such as a bass guitar), the compressor or limiter can respond to the actual individual cycles of the waveform, causing harmonic distortion. To prevent this effect, increase the release time. Limiter The purpose of the limiter is to prevent the audio signal from exceeding a threshold value. A limiter may be seen as an extreme case of compressing Its parameters include Setting Range of values Compressor Threshold.............From -40dB to 0dB Master Gain.............................From -24dB to +24dB Stereo Handling.......................Maximum or Side Chain The most common problem caused by the use of a limiter is excessive distortion. This indicates that the limiter is active too much of the time, or that the master gain is set too high. To correct the distortion, lower the master gain or increase the threshold. Expander/Gate The stereo expander expands the dynamic range of the portion of an audio signal which falls below a given threshold. If you use the graph to control the settings, note that the graph shows only two visible nodes, allowing only the initial slope representing the expander to be moved. Its parameters are listed below. Setting Range of values Expander Threshold ................From -108dB to 0dB Expander Ratio ........................From 1:0 to 1:infinity Attack Time.............................From 0.1ms to 200ms Release Time ...........................From 20ms to 1000ms Master Gain ............................From -24dB to +24dB Level Detection Method ..........Peak, average, or RMS Stereo Handling ......................Maximum or Side Chain Envelope Curve .......................Soft knee or hard knee Here are several common problems you can run into when using an expander: Problem What to do The beginning of a sound is cut off ........Decrease the attack time The end of a sound is cut off ................Increase the release time Disturbing sudden change in background noise as the expander cuts in and out............Lower the expansion ratio or increase the release time. The Interactive Graph Three of the four dynamics processors (excluding the limiter) include a graph that lets you draw the envelope curve that defines the threshold levels and ratios. You adjust the envelope by moving the nodes that define the curve. Here’s how: To do this Do this Move a node................Drag it to a new location with the mouse. Dragging with the left button down will cause the node to follow the cursor precisely. When the right button is used, a large movement by the mouse results in a small movement by the node, allowing more accurate control of the node’s location. Zoom in and out of the top right corner of the graph......Click on the text ‘Zoom-in’ and ‘Zoom-out’ Add a node..................Click a short distance away from the existing nodes Erase a node................Press and hold the Shift key, and then click on the node Note that you cannot delete the starting or ending envelope points. Also, keep in mind that the graph does not reflect the effect of master gain on the output levels. The “knee” refers to the bend in the envelope curve. A soft knee transition converts a hard corner into a smooth curve that kicks in more smoothly and gradually on signals surpassing the threshold. Soft knee compression curves are suited for vocals, guitars and other nonpercussion instruments, as well as entire mixes. Hard Knee compression curves are better suited for drums and other percussion instruments. Stereo Handling The dynamics processors offer two choices for processing of stereo signals: Option How it works Maximum.....................The left and right channels are adjusted equally according to the maximum of the left and right channel signal levels. Side Chain....................The left channel volume alone is used to drive the effect, which is applied only to the right channel. This can be used, for example, to automatically lower the level of background music Glossary Term Meaning Attack Time.................The time it takes to respond to increases in the input signal level Compressor.................An audio effect that reduces the dynamic range of an audio signal dB (decibel).................A unit of measurement that expresses the relative levels of two electrical voltages Dynamic Range............The difference from the highest and lowest levels of a varying audio signal Expander.....................An effect that increases the dynamic range of an audio signal, by reducing the signal level any time it drops below a specific threshold Gain............................The amount of amplification applied to an audio signal Gate............................A device that increases the dynamic range of an audio signal by cutting off a signal when its level falls below a specific threshold Hard Knee..................An input/output curve that changes slope suddenly at thresholds Harmonic Distortion.....The addition of harmonics that were not present in the original signal, due to changes in the shape of the waveform Limiter.........................An effect that prevents an audio signal from exceeding a specific level Noise Gate...................See Gate Ratio............................The factor by which gain exceeding the threshold is reduced or expanded Release Time................The time it takes to respond to decreases in the input signal level Soft Knee.....................An input/output curve that changes slope gradually around thresholds Threshold.....................The level at which the compressor or expander starts to operate Transient.......................A momentary high peak level Waveform.....................A graphical representation of an electrical signal ========================================================== ============ ======================================= Using the Studio After you've built a studio, you need to use it to record music, right? Recording guitars Recording drums Recording vocals General recording advice Bouncing tracks Noise reduction Eliminating sources of hiss Cleaning and demagnetizing Recording Vocals For recording vocals, position the mic so you're facing up to it. Use a good mic and make sure you're in a place you're really comfortable working in. Record the vocals last and make sure they mix in well with the music instead of floating on the top (this is actually okay with some songs, but I dislike it in general). I also record vocals with effects so I know exactly how they will sound in the final mix. Also, when you EQ the vocal track, cut the lows (definitely) and the highs (perhaps). This gives clarity to the vocals. Jonathan Epstein again: For vocals I also send through my Concert just to add a bit of reverb and bit of bottom end (weird, I know, huh?). I usually stand about 8-12" from the mic for singing. You lose the popping p's and b's and s's this way and also get a bit of natural reverb. Frank E. Fullerton writes: I usually record vocals last. Quite by mistake, I once left the monitors on when I recorded the vocals. The existing tracks were picked up on the vocal track. It was magic. I never realized before how "separate" the vocal track sounded from the rest of the music. Now I always let the monitors leak into the vocal mike. Try it and see if your vocals don't become one with the music. As I say above, making your vocals become one with the rest of the instruments is important. I accomplish this by recording in stereo with a slight delay, adding a bit of a flange that "slides" in with the rest of the instruments (this might not be the best description, but it is how I view it). Frank's idea of using a distant echo of the vocal track might be something worth experimenting with. Depending on the nature of the track, you might either want to make the vocals stand out or be buried in the mix. I really think this is a track-by-track choice to make. Frits van Mourik writes: cut your EQ at 1 kHz about 6-12 dB. Bandwidth (Q-factor) should be a 0.5 octave. This way you take out the "harshness" of the vocals, leaving more room for other instruments. You can crank up the voice in the mix without becoming to loud. Works great on my U87, AT4033, and even SM 58! From Rob Kirbos: When recording vocals, DO NOT use a foam wind screen that fits over the top of the mic. these wind screens eat up the top end as you print the track, then you have to boost the highs on the EQ in the mix down, which adds to the ever popular (and unwanted) hiss. For a wind screen, do the following: Get a wire hanger. Put a loop in it about 6-8 inches in diameter. Stretch some panty hose or stockings over the loop. Get an old pair from your sister or girlfriend [how sexist! ] so you don't get any funny looks in the store as you buy them. A clean pair is recommended. The "Pro" windscreens that you buy in a music store will run about $15-$20 depending on the brand, store, etc. the only difference is that these have a nice little clip and goose neck so you can mount it right to the mic stand. for the coat hanger model, duct tape works just fine. Recording guitars I have a guitar tone I like (be cautioned, I do like warped music) on the 4-track. I do NOT mic the guitar amp. At first, the guitar did sound dumpy. But I found that unless you have a very good amp the sound quality of your recording didn't get any better. I initially got around this by turning the mid-range eqs up when I mixed down, but I found that if you can record using stereo effects (i.e., have an effects processor that gives stereo output (the RP-10 that does), then the sound improves greatly. This means using up two tracks, but it's worth it, IMO. The other thing you can try is to go direct from your amp out to recording device instead of micing the amp. This I've found produces quite a good sound that is comparable to one produced by micing the amp, but with lower hiss. If you want to be more experimental, you can do both: mic the amp and and record direct from the amp. This leads to some interesting sounds. Jonathan Epstein says: I've used a few different methods depending on the sound I want to achieve. The last piece I did was a light sorta-love song, I used an Ibanez with 2 humbuckers going with the tones turned way down onthem pacthed through my Fender Concert tube amp out through Line recording jack. The second guitar track was a Tele sent thorugh the same amp in the same way. I mixed the first mid-Left and the tele mid-right. My most common practice for recording my guitar is this: I use my Tele (with 13s!) and go into my amp. Then I mic the amp up close, and also go out through the line recording. I love the tone of guitar/amp and not micing the amp, I get to crisp a tone and miss a lot of the bottom end. This way I get both, the nice deep tone and the crispness with line out. I throw one hard left the other hard right. I am very happy with my guitar sound. For mics I use an Audio Technica ATM41a and (get this) a nice set of digital headphones. It acts like a PZM, and captures the sound very well considering. For bass, I usually send it through my Concert and use the line out, instead of micing. I like to add a little reverb too. Gives the bass a nice, almost fretless sound. I am also very proud of the bass sounds Ive been able to get. My one problem is that, it strats to sound kind of muddy when I have a lot of busy tracks. I need compression, but dont have the cash right now. Chris IX writes: I use the Hughes & Kettner Tubeman. It is a real tube preamp/cabinet simulator which sounds remarkable. You can plug it direct into your mixer. I makes sounds from clean jazz to shred. I like to record w/o effects because it makes punch in/outs cleaner. Otherwise reverb/echo tails tend to get truncated. Also adding f/x after the fact will help to smooth over and give tracks with lots of punch in/out more continuity. Lorr Safratowich has this to add about recording guitar direct: I've found that taking a signal from my amp's direct out and running in thru a 1/3 octive eq set to emulate a speaker's frequency response (I use the tech info from a EV 12L) and then into my board has given me consistently good guitar sounds. I think the kicker is to cut just about everything above 7000hz on your guitar signal. One could adjust other freqs to taste for body. Also, to keep the noise levels down, like to zero (other people don't seem to have the passion for the lick I need 20 takes to get!), I've taken a 10 ohm resistor rated at 50 watts, put it in a project box and wired a 1/4 inch jack to it. I then plug my speaker wire from the amp to it and have an effective load box. I wouldn't advocate running a 100 watt Marshall head wide open into it, but for lower level things, it works and is a lot cheaper than the marketed items. I did check this out with an amp tech and its ok as long as one is prudent with the volume knob. Recording Drums Jonathan Epstein again: For drums, I've used a couple of techniques. Two Radio Shack PZMs (my best friends!! I wish I still had these, someone swiped them from me a year ago) one over head and one in the kick. Lately Ive just used the AudioTech mic or headphones hanging overhead in my reverb-rich attic. I actually like the sound better from the headphones. Ive been recording thorugh headphones for 6 years now and think it adds a nice true effect. I played stuff I did for my brother and he said he didnt even notice the difference. Matt Macchiarolo has this to contribute: First, I place the drums in a suitable room, usually my living room, which is the biggest in the house. (I have to wait for my wife to leave, first.) I'm a big fan of close-miking, so I mike each piece separately; bass drum, AKG DA20 (THE mike for bass drum); 2 toms and hat, Shure SM57's, snare, a 30-year old Shure SW55, overheads, a pair of AKG C-408's. The advantage to close-miking is that you can pan each piece for a true stereo effect. The key is proper mic placement; I don't touch the EQ until I get the best possible sound from each mike, both by itself and in the mix, and the only way to do that is to place them properly. I mike the bass drum by first putting a pillow in the drum to kill unwanted overtones and speed up the decay. Then I put the mike inside the drum, about 2-3 inches from where the beater strikes the drum, then play around with it to get the best sound. After that, I still might have to EQ some highs into it. For toms, I place the 57's in such a way so that leakage from the cymbals is minimized; this usually means placment so that the "rear" of the mike (where the cord is connected) is pionting directly at the cymbals. For hihat, I place the mike at the perpendicular to the very edge of the hat, to make it "cut" more...placing the mike toward the center makes a "brassier", heavier tone. I mike the snare's batter head at about a 45 degree angle. Overheads are pointed over the cymbals, but they still pick up the ambiant drum and room sounds; these condenser mikes were originally designed for close-miked percussion or horns, but they work fine as OH's too. Once the kit is miked, I assign each of the 7 channels used by the mikes on the Mackie 24-8 console to two submasters, then pan the hat, toms, OH's to approximately match the placement of the drums from the drummer's point of view (Snare and B.D are panned center). The submasters are patched to the inputs of the TSR-8 tape machine...essentially I'm mixing the drums to 2 tracks when I record it. Usually I record everything dry, then add a bit of reverb to the overall drum mix at mixdown to give it a "live" sound. If I want to add an effect to an individual piece (like a gated delay on the snare) when doing the initial tracking, I usually record the effect on a separate track and mix it all together at submixdown (bouncing all the other tracks to 2 track) or final mixdown. This seems really involved (and it is), but it's worth it when your drums tracks sound as big as life. I know many people don't have seven mikes at their disposal, and that's OK, I've heard a lot of good sounding mixes using two mikes. I think a recording miked well with two good mikes will usually sound better than one miked poorly with 10 crappy mikes. When purchasing, don't skimp---get the best mikes you can afford, because the sound hits the mikes first. I got a chance to use the Radio Shack mikes, and to tell the truth I was pleasantly surprised. The sax player at last week's gig who up till now used a Shure SM57 said the new mike was a little bit clearer and didn't require as much high-end boost as the 57. (Of course, if he would wear earplugs at the gigs, he wouldn't need to boost the highs toward the end of the show.) Last night I non-scientifically compared the RS mike to my Audio Technica 4050 studio mike and the AKG C408 mini condenser. With the pre-gain on the mixer cranked, I found that the RS had only slightly more self-noise than the AT4050 and not nearly as much as the AKG. (the AKG was designed for very close-proximity drum miking so you would never have to crank the gain up that much in real life). For a mike powered by an AA battery, I was impressed! It did sound a bit different than the AT, especially in the low-mid range, but it still sounded good to my ears. And it seemed to be about as sensitive as the AT; it picked up the clock ticking across the room! Granted, my tests were far from scientific, as I don't have any high-end test equipment except for my ears. It's no Neumann, but for 70 bucks, I think it's a good buy! I think it would be a very good vocal and instrument mike; I plan to use them for drum overheads and/or snare/hihat mike. Again, it's Radio Shack catalog #33-3007 Unidirectional Condenser Microphone. It's been discontinued, so you might have to hunt for them. Frits van Mourik has this to add: When recording/mixing toms, remove the area around 800 Hz. This will take out the "cardboard box"-sound. If you are losing the toms in a mix dominated by fierce guitars, then add some narrow-band eq around 3.5 kHz. The attack will pierce through the mix and psychoacoustics (whatever that may be) will do the rest. Can't find that "smack"-sound in the bassdrum? Tune it as low as the drummer feels comfortable with. The beater now smacks into the drumhead instead of bouncing off and there you are! Ken Wronkiewicz has this bit of pragmatic advice: Live drums are a pain to record. If you use a synth's drum set (which has come a long way from the old 808s) you'll get a much better sound for most purposes than trying to record a drum set and getting everything just right. General recording tips I almost always do two vocal tracks. Since I'm into warped music, I tend to process one track extensively and leave the other one alone. In general, I find that there's a fuller sound with recording things just slightly differently in two tracks. The trade off between this and the decline (though I don't notice it) in sound quality when you bounce tracks is a judgement call you have to make. For every song I've recorded, I usually do both: record everything in mono/one-track (using up only 4 tracks) and then record it in stereo/two-track (using up to 8 tracks) and see how it sounds. People, in general, react better to the stereo sound, even though the mono performance might be superior. My sample is based on 4 songs, but there is some "magic" to doing everything in stereo. In general, I have found it better to use up 2 tracks to record the full stereo sound of a guitar effects unit or keyboard and do bouncing (see below for how I do my bouncing) than to use only one track. Be sure to play your tape on other players after you do your final mix so you can see how it is going to sound on different machines. Important! Make sure you write down everything you do! Especially when you mix down, make sure you have all the equaliser controls marked on a sheet of paper. What I do is photocopy the front page of my manual which has a plain figure of my tape deck console and then I blow it up and use a coloured pen to mark my settings. Bouncing tracks Of course, if your compositions require more than 4-tracks (besides the keyboard, which could be a sequencer containing many virtual tracks), you are going to have to bounce tracks. Combining tracks onto a single track on the 4-track is a bad idea, I've found. For whatever reasons, it ends up missing something. A better option is to mix down to a GOOD casette deck (a DAT would be great) and then re-record the mixdown on your 4-track. If you are using a normal tape recorder, then you can take the mixdown tape and use that for further recording, thus saving yourself a generation. Be careful that you're consistent with the noise reduction scheme that you are using. If I require 6 or more tracks, my usual procedure is as follows: record the keyboard on 2 tracks, record the guitar on 2 tracks, mixdown to a 2 track and then you have 2 more tracks free. And so on. If you have only one vocal track and two keyboard tracks, then depending on the sound of the guitar you want, you might be better off just using one track for the guitar. If there's a guitar solo you want to add between vocals, and don't want to use up an entire track, use the punch in facility for the vocal track(s)! This is where auto punch in has been the greatest help for me. At any point, you can always mixdown with some of the existing tracks, but this is somewhat irreversible. For example. I sometimes use up 4 tracks on sequences and guitars, and then add the vocals on mixdown. This means I need to get the vocals sounding right (levels and performance) in one shot. Using this technique in an extreme sense might mean that you don't even need a 4-track, particularly if you do digital overdubbing using a PC. Noise reduction (NR) Use high speed recording. This goes without saying, but I tried out it out both (high and normal) ways, and it does result in a better signal/noise ratio. I happen to have dbx NR in my 464, so I always have this on when I record and when I playback. I do not however record with dolby B or C turned on on the mixdown deck when I mixdown. I find doing this results in a "dull" sound. I have pretty much obtained similar results when I don't record using dolby B or C turned on and turn it on when I play the sound back. This might violate everything you've heard about dolby noise reduction, but it's true in my case at least. Try it out. Michael R. Kesti has this to say about noise reduction: To understand how to use Dolby noise reduction, it helps to first understand how it works. Dolby noise reduction's goal is the elimination of hiss, which is present in any magnetic medium recording process, but is especially a problem with the Phillips cassette format, due to low tape speed and narrow tape width. It's approach is to pre-emphasize the high frequencies during the record process, and to apply a matching de-emphasis during the playback process. Because the hiss is not present in the pre-emphasized signal, but is present in the signal that to which the de-emphasis is applied, the result is a reduction in hiss level (an effect increase in signal to noise ratio), and a flat system frequency response. It is important to understand this because it shows that the Dolby noise reduction system is a two sided system, that is, its process is applied during both record and playback. This means that if a track was recorded with Dolby enabled, it should be played back with Dolby enabled, regardless of whether that track is being sent to a final mixdown or being bounced. I will add this technical detail regording NR from my Technics manual: dolby B reduces noise about one-third. Dolby C reduces noise about one-tenth, and HX-Pro is something that allows recordings without drop of the level of the sound source's high-frequency range. Andy DeFaria: the difference between B and C seem to be that C boosts more treble than B. As a result I tend to use C during record if available and never put any Dolby on during playback. This is probably wrong but sounds best to me (usually). Bruce McGee: I looked into the dbx issue and I think that db B and db C should not be used with this in recording. The dbx in some circumstances has a very controlled algorithm for current bias, which may in fact result in total cancellation of recorded frequencies when used in conjunction with db C especially. It would be interesting to hear the opinion from the experts. Steve Shumake: If you happen to be using floor type stomp pedals for effects, I have found that using the batteries is much quieter than using the AC adapters. The adapters cause a lot of hum, but just watch your batteries so that use always have fresh ones. I used to notice that even way before the batteries were out of juice that distortion was audible and increased as the batteries drained. I fixed this by going with a rack digital multi-processor! Ken Wronkiewicz: if you are doing digital hard disk editing, you can get rid of hum and noise by taking it out by hand. So I'd select when I wasn't singing and then I'd have the sample editor fill it in with zeroes. Advice from Jason Olson of New Spectrum Sound Note that using a computer with a soundcard can introduce a lot of noise. Sometimes moving the soundcard around will help (it did in my case---it was sitting right next to the video board and I placed it in another slot and it greatly reduced the amount of noise). Plus if you're recording using a mic, you could have various humming noises (from the monitor and the fan, for example) picked up. The way I solve this problem is to keep my sequencing separate from recording guitars and vocals. I dump a mix of all the sequenced instruments onto my 4-track and I go from there. Eliminating sources of hiss I've not mastered this yet, but it can be done if you're clever about how you channel your inputs. I've found the least hiss is when I turn the volume of my instrument up instead of turning up the trim control (though this is by no means absolute) on the 4-track. In the case of the keyboard, I found this very noticable. I keep the trim completely down, and turn the volume of the keyboard up until I get the recording levels I want (the "ideal" recording levels are described in my 464 manual). The same goes for the output of my RP-10. For vocals, unless you are recording from an amp, you probably need to turn the trim up. This is my major source of hiss. George T. Talbot contributes with: one additional tip about hiss & microphones is that the mixers that manufacturers put into 4-track tape recorders are cheap. I had a lot of trouble getting decent sounds on tape with my Tascam Porta-2. I went out and bought the 12-channel Mackie mixer (~$400). It made a WORLD of difference, especially with microphones. Before I had the mixer I had to use one of those transformers that you plug into an XLR microphone cable to plug the microphone into the unbalanced input of the 4-track. The Mackie (as do the other low-cost mixers) has balanced XLR inputs. The dynamic range I get from the microphones now is better than the tape deck, and the only hiss I get is from the tape and the guitar amp. Having the mixer also makes recording stuff a lot easier. I like to record vocals dry and process them on mix-down, but singers like to hear the effects (reverb, etc.) as they are singing. With a mixer, it's pretty easy to do this. James Eibisch writes: a way to cut down background hiss is to record with high levels, being careful to balance it against distortion from overloaded circuits, so putting the signal/noise ratio in your favour on mixdown. Trey Canipe writes: One discovery that I have made is a tape on the market for four trackers: the Radio Shack LN30. It has the high band of the spectrum dampened to hiss. It gives you really quiet recordings and doesn't kill your highs. I used high bias tapes before and i had clarity but still picked up some hiss when I bounced down. I would give up any (unknown to my ear) tiny bit of sensitivity in return for the amount of hiss that is done away with. I use these tapes for four tracking and for copies. I think that analog can be used for demo purposes and still make impressionable tapes. Some people don't "hit the tape" hard enough on their original track and that makes for hissey recordings after the level is brought up. This is my technique: I record drum tracks first. I hit the tape hard by hitting kick and snare and trim up until it gets the rattle in the monitor from overfeed. I back down just barely below overfeed while pounding the kick/snare. You can use Dolby but I don't. I find that with the low noise tape and hitting the track so hard, very very little hiss is on that drum track. The reason I say drum track first is because now you can track another instrument at any time you feel like (retake over and over) vocals any time, etc., because now every thing will be in perfect time (guitar player can come in one, singer the next). Also, if the drummer needs the guitar to hear in order to play the song, put the guitar on headphones and and play with him. Now, you need only a real basic "bass and snare" type drum track and you can dub in all kinds of fills on other tracks. The main idea is perfect timing and let everyone else do their thing. I have seen guitars tracke out first with a click track and then drums later but it is hard for me to do that. Cleaning and Demagnetising I clean at the beginning of every recording session. I use the Radio Shack professional head cleaner. For demagnetising, I use the Radio Shack battery operated tape demagnetiser and I do this every other recording session on all my tape decks. ========================================================== ============ ====================================== Mixing, Mastering & Mixtering Mixing, Mastering? While mixing and mastering were quite separate processes in the past nowadays the line gets thinner and thinner. Why is that? And why does or doesn't this matter? Because of the many new high quality plug-ins, sample editors, digital effect processors and computers, nixing and mastering has gotten easier and many people can afford it nowadays. Just over a year ago editing on a computer with a sample editor took about 2.5 times the time it took on an expensive Pentium 200 compared to a cheap Celeron 400 System now. Outboard effect processors are equipped with 20 or 24 bit A/D converters and are packed with high quality digital effects, and getting cheaper. New plug-ins seem to pop-up every week with high quality effects and even cheaper. The only thing that hasn't gotten cheaper is the expensive analog gear some people use for mastering. But should or do all these things matter to the way music is produced, mixed and mastered? Mixing Mixing used to be the process, and still is, of mixing multiple or many tracks back to a stereo track. The most important factors while mixing, of course, are getting all tracks at the right level in the mix. Usually this involves setting the equalizers for each independent track and if needed some compression on a track. The basic idea behind the mixing is getting the stereo track to sound good while all the important instruments are there, the left / right and stereo balance is right and a good direct/depth ratio. Mixing effects While effects were used mainly on the busses of the mixing tracks, these days it's easier to have almost any effect you want as an insert on a separate track. Not to long ago most effects were either recorded directly on to a track with an effect processor between the microphone and the recording device and you would have to live with it, or you did have to do a lot of editing to get all the effects on the dry recorded track. With all these effects running on tracks, mixes can sound very good already so why do you need mastering? At least, that's what a lot of people seem to think when making/recording music and mixing. Mastering Mastering used to be the process of processing the stereo track to get it ready before getting it onto the final product which the consumer puts in his stereo system. The basic idea behind mastering is getting the mixed stereo track to sound good on many different types of stereo's, car systems, walkmans and on the media it is used or played on tape/CD or broadcasted on radio/TV. Mastering effects To get a stereo track to sound good on a media like CD, processors / effects like: compressors, limiters, equalizers and the latest kind of maximizer or loudness effects are used. These effects should not really introduce it's own sound but should be high quality and that's the reason the analog processors are quite expensive. But then there is the digital world and the line seems to disappear. Not only do programmers include very transparent effect but they also included effects that have the warmer sound of analog processors. Both? Now that most people have many effects and also some mastering tools and easy ways of burning CD's what does the mean for mixing and mastering? For instance it opens up a new way of doing both. You can actually hear a little of what the mix will sound like when having it mastered or mastering yourself. Just print all the mastering effects on one of your own mixes with a sample editor like a compressors, equalizers, limiters and loudness/maximizers and listen back how it can sound when it's mastered. You can even put it on a CD and start listening to these semi mastered songs on your car system and other systems. Don't like it? Go back to mixing and see what you can fix. Or even better put the effects or an effect/mastering processor on your outputs or between your computers soundcard or mixing desk and your monitor system and hook up some stereo systems too. Now you don't only hear the mix you would normally do but you can also hear how it would sound after additional mastering processing and maximizing on both your monitor system and even a crappy stereo system, or multiple. So now you can both mix and "master" at once, or at least listen to what the mastered version can sound like. So why doesn't it sound like the songs you listen to on TV, radio or CD? Well you need another plug-in. It's called the experience plug-in. Oops.... That one hasn't been invented yet, has it? The important thing about mastering is that when you already have a good mix you need to listen different than you do with mixing. The changes when made during mastering are much more subtle and with a different purpose. You don't have to concentrate on the individual tracks but you have to listen to the effect of the processing on the mix and have to wonder how it will sound on different systems, speakers, environments. A little different than mixing. Impossible when you are mixing? Not really, but as many things in life it takes time, practice and experience. Back to mixing.. err mastering.. mixtering? ========================================================== ============ ======================================= AUDIO RECORDING WITH COMPUTER SEQUENCERS The recording studio inside the virtual world of the computer is real enough, but sometimes you have to treat it with care to get the best from it. Paul White offers a few tips on the subject. Computers offer us MIDI, audio recording, mixing, virtual effects, virtual synths and CD manufacturing facilities, but it doesn't pay to take them for granted. The following tips will help you get the best out of your system, whether it runs on a Mac or PC, and most assume that you already have a system that's up and running. If you're planning to buy a PC system but aren't sure what to go for, check out the FAQs on our web site and give some serious consideration to buying the system preconfigured from a single vendor rather than assembling it yourself. If you want to buy a Mac system, either buy one of the newly obsoleted grey Macs or wait until the peripherals and software copy protection needed to work with the new candy-coloured Macs are ready. In either case, buy the fastest machine you can afford - even if you can't afford it! 1. Optimise your input signal level at source rather than relying on normalisation to bring the level up: if your signals peak at only half the maximum level, you're effectively halving the signalto-noise ratio of your recordings and wasting half of the theoretical resolution of your system. Digital processing such as EQ or reverb may also introduce far more noticeable rounding and quantising errors in low-level recordings. Use the level metering provided in the software and try to keep your peak levels just a few dBs below clipping. 2. Regardless of whether you have 16-, 20- or 24-bit recording, the real quality of your recording will be defined by the source. For vocals, consider buying a voice channel type of device that combines a good mic amp with EQ and compression. This may also be used when miking other instruments, and many feature an instrument DI input suitable for use with bass and clean electric/electroacoustic guitar. 3. The fact that computers and recording software are such good value for money can lead you into believing you can make do with equally cheap components in the rest of the studio. This simply isn't true. With good capacitor vocal mics now available for under £200, there's no excuse for using your old gigging dynamic microphone. 4. Use quality monitor loudspeakers and set them up so that you're at the apex of a roughly equilateral triangle with the monitors pointing directly at you. You don't need to monitor loudly, but you do need enough volume to overcome the physical noise your computer fans and drives make. 5. Use a separate hard drive for audio if at all possible as this will increase the number of tracks you can play back at the same time. This also allows you to defragment, or even reformat, the drive regularly without disturbing your program files. Most modern drives are suitable for audio use, but if in doubt, get a drive that is badged as being suitable for AV applications. The faster the drive you buy, the more tracks you'll be able to play back, though very fast drives may need a special fast SCSI interface card to make the best of their capabilities. If you really can't afford a separate drive, at the very least create a separate partition on your main drive for audio use. 6. When choosing or upgrading a soundcard, try to get one that can provide at least four outputs - and a digital S/PDIF out if you own a DAT machine or Minidisc recorder. This way you can use one pair of outputs for tracks that use software-based plug-in effects while the other output can carry tracks that you want to effect using external processors. 7. Reverb is the most important effect in the studio, and good reverbs take up a lot of computing capacity. For this reason, it may be worth considering buying a soundcard with its own hardware reverb processing, such as the Lexicon Studio, the Yamaha DSP Factory or the Yamaha SW1000XG. The SW1000G also includes onboard synth sounds that can be patched through the same hardware effects as the audio tracks. 8. Unless you are using a fairly sophisticated soundcard with onboard DSP processing, you're likely to experience some latency or delay when monitoring the signal you're currently recording through the system (See Martin Walker's article on the subject in SOS April '99). The new ASIO II drivers will minimise this problem for compatible hardware, but it won't cure the problem in all soundcards. An alternative is to use a small mixer and arrange to monitor the computer's input rather than its output when overdubbing - a separate mixer will usually be needed to combine your audio and external synth/sampler signals anyway. Monitoring the input source will avoid latency problems, but will mean you have to monitor without plug-in software effects. However, a simple hardware reverb unit is generally all that's needed to put you in the mood for a good performance, and you can probably make use of this when mixing if your card has more than two outputs. 9. Use Antares' Autotune plug-in not only to clean up vocal pitching, but also to tighten up guitar solos (a low-cost VST 'light' version is due very soon). As long as you set a slow enough tracking time, regular playing will be unaffected, but whenever you sustain a note, it will automatically settle on exactly the right pitch. This can be particularly useful for slow pieces that use a lot of string bends. You can also emulate that Cher 'Believe' vocal-type sound extremely convincingly by just setting the tracking speed to maximum and dialling in the correct key for the song rather than leaving Autotune on its Chromatic setting (although of course, Cher's producers claim Autotune was not used on that recordeing - see SOS February '99). 10. One problem that most guitarists come up against is that the computer's monitor interferes badly with the guitar pickups, resulting in a nasty buzz on the recording. Some humbucking pickups are reasonably good at rejecting this buzz providing you don't sit too close to the monitor while recording, but singlecoil pickups tend to be very badly affected. One way around this problem is to switch off the monitor just before recording and use keyboard commands to start and stop the recording process. If you can't switch the monitor off for some reason, sit as far away from it as possible when recording and rotate your position to find the null point where the buzz is least obtrusive. You might also use a noise gate pedal to keep your guitar quiet between phrases. Flat-screen LCD monitors are becoming cheaper and they both save space and eliminate the electromagnetic interference generated by the scan coils of a typical monitor. If you record a lot of guitar, or are short on space, such a monitor could be a good investment. 11. Physical noise is also a problem when miking instruments or voices in the same room as the computer. If possible, turn off unnecessary external drives, CD-ROM burners and so on, as these often make more noise than the main computer, Set up your mic (ideally a cardioid model) as far from the computer as possible and improvise an acoustic screen between the mic and the computer using a duvet or sleeping bag. Also make sure the surface the mic is pointing at is absorptive rather than reflective. Work as close to the mic as you can without compromising the sound (and always use a pop shield for vocals). 12. Virtually all sequencers capable of recording audio have a waveform edit page (though it isn't always called that) where it's possible to highlight and silence selected portions of audio. If background noise was a problem, you can sometimes improve matters by manually silencing all the gaps between words and phrases. This doesn't take as long as you think and can really improve the quality of a recording, especially where there are multiple audio tracks. It's a good idea to normalise your audio recordings before processing them so as to minimise rounding errors at the processing stage, though don't use this as a substitute for getting the record levels right in the first place. Normalising can generally be done from within the waveform edit page. 13. You can also use the Waveform edit page to clean up guitar solos. Often you may end up with an almost perfect take, but perhaps there's too much squeak or finger noise between notes, or maybe you caught the next string just after bending a note. You can use the silence function to surgically remove these little errors, though you may end up with a more natural sound if you leave them where they are but instead reduce them in level by between 6 and 20dB. 14. Try to record all parts dry - don't add reverb or delay unless you really have to. If you need to hear reverb to create a good performance, fake it at the monitoring stage, but don't record it. This way, you'll be able to edit tracks without cutting holes in the echo or delay effects you've added, then when the editing is done, add the necessary delay or echo, which will help hide your edits, making the recording sound quite natural. 15. Plug-ins always take up a certain amount of your computing power, so if you want to add the same delay or reverb-based effects to several tracks, use a single plug-in configured as an aux send processor rather than using a separate Insert plug-in on every track. You can use the Aux Send controls in the same way as those on a regular mixer to add different amounts of the same effect to any tracks you like, all for the CPU overhead of a single plug-in. Note that under normal circumstances, you can't use the aux send with processes such as EQ, compression or gating these have to be inserts. 16. Often, it's cheaper to buy a hardware reverb unit or signal processor than to buy a decent plug-in that does the same job, and the chances are the hardware unit will still sound better. Don't try to force your software to do everything for you just because it can - very often you'll find you can get a better sound with discrete boxes, and of course they won't load your CPU. Even if you don't have a multi-output soundcard, you can still compress signals as you record them, ideally using a voice channel type of device as described earlier, and the same applies to EQ. Only the best digital EQs sound as natural as even the most basic analogue equalisers. 17. There are lots of tricks you can do using the audio manipulation facilities provided by your sequencer. These vary from model to model; pitch-changing and time-stretching, which are invaluable for massaging audio sample loops, are supported by most machines. You may also find other tools for level maximizing, denoising and so on. Many of these work off-line, so you can use them even on a slower machine - you just have to wait around a while for the results. 18. Consider using CD-R to backup your audio files along with your song files. Though you can't rewrite a CD-R, they're so cheap now that it doesn't really matter. If you create a 600Mb partition on one of your drives and store (or copy) your audio and song files there, you can back up the entire partition in one go. Of course the same CD-R machine can be used to burn audio CDs of your finished songs. 19. Most computer audio systems run best if you get rid of any superfluous software such as screensavers and games - and make sure you have no more drivers than you actually need (Extensions for Mac users). The cleaner your system, the less likely you are to run into problems. Also, check manufacturers web sites to make sure you have the latest drivers as improvements are being made all the time. 20. Do some tests to find out how many tracks and plug-ins your machine can run without falling over, then try to work with no more than half to two-thirds this number. Most sequencers include some kind of CPU activity monitor to help you. The demands on your CPU aren't constant, and sometimes a lot of heavy processing loads can be imposed at the same time, which can cause a machine running close to its capacity to crash. Your disk drive will also slow down as it fragments, so try to allow for this - you can't be expected to defragment it after every track you record.
© Copyright 2018