music production
[To find out what the various types of equipment used in music production are and what they all do, click here.]

The purpose of this page is to help anyone who may be interested in electronic music production, but doesn't know much about it and to encourage the production of new music. It features common production methods and genuine sound engineer's tips and covers vocals, mixing, processing, using effects, digital audio and studio acoustics. To find out about the various topics, click the relevant blue highlighted titles above, or on the section link panel to the right.

vocals

vocals  •  recording

There are only a few important things to know about recording high quality vocals. Firstly, it is always better to get a good recording at the start of a project than to try to fix it later. There is no point trying to record in a location where there is a lot of background noise. Assuming you are somewhat free of background noise, the next important step is microphone and singer placement.

[To find out about microphones, click here, or for acoustic treatment for recording vocals, click here.]

If you have a pop shield, it should be placed about two to three inches from the microphone with the singer between three to nine inches from that, depending on the strength of their voice and performance. It is recommended that a pop shield is used, to remove the worst of the plosive sound in words with 'p's or 'b's in them. If you don't have one, you can easily make one using material from tights stretched across a metal clothes hanger that has been roughly bent to the shape of a circle. The best way to ensure good microphone placement is to listen to your voice going through the microphone using headphones, as you slowly change distance from it. Also, it is worth doing a few test recordings and comparing them over speakers if you're still not sure that you've got the right spot.

...even slight movements of the singer's head can alter the voice's recorded frequency range and level.

You should be able to notice that even slight movements of the singer's head can alter the voice's recorded frequency range and/or level. While this can be detrimental for a singer who wildly swings their head around during a performance, it can also be used to add emotion to a recording by an artist in the know... Microphones with directional polar patterns (cardioids and bi-polars) will exhibit a 'proximity effect' when the singer is close to the microphone, which can also be used for good effect by professionals. This effect causes a build up of lower frequencies in the microphone's frequency response, which can give a more intimate or personal touch to the recording.

When setting the microphone gain level, it is important to watch the input level as the singer sings at both the quietest and loudest parts of the song. It is also good working practice to leave between 6 dB and 10 dB headroom when recording, ie. the loudest part should register between -6 dB to -10 dB. Recording to 24 bit digital audio affords enough dynamic range to leave a bit more headroom to ensure that extra loud transients in the voice don't make the recording distort, which is very important when using digital equipment.

The only real reasons to compress the vocal as you record are if the singer or performance has very high dynamic range, or if you want a super squashed vocal sound. Even then, it would probably still be better in this day and age to record it clean and effect it later in case you ever change your mind about the sound. However, using a limiter, or a compressor with a high threshold and ratio and no gain, to ensure that none of the signal will go over the 0 dB maximum for digital audio can be beneficial.

Equalisers and effects should not generally be used while recording vocals. If you have a mixer with the ability to record only certain channels, ie. the vocal channel, it can sometimes be helpful for the singer if they can hear a little reverb on their vocals, as long as the reverb output is sent to a channel that is not being recorded.

top

vocals  •  processing

Vocals almost always benefit from compression. Exactly how much is needed depends on the vocal performance...

Different types of music genres generally use different styles of processing, so processing vocals is a very individual thing. Below are some popular methods for achieving different effects that can be useful for most genres.

Vocals almost always benefit from compression. Exactly how much is needed depends on the vocal performance in the recording and the type of music genre being produced. It is common to set the threshold so that you compress up to around 6 dB, using ratios up to 4:1 and fairly fast attack and release times of up to 0.25 seconds. If the attack time is too fast, it will affect the initial transients of each syllable and this can sound un-natural and should be avoided.

[Click here to find out about compressors and click here to find out about using them.]

Sibilance refers to the resonant, hissing part of the sound that the human voice makes when pronouncing letters such as 's' and 't'. Its frequency range lies between around 4 kHz to 8 kHz and is slightly different from person to person. When too much sibilance has been recorded, it can be corrected with either a very narrow bandwidth equaliser, or a de-esser. The problem with using an equaliser to cut the sibilant frequencies is that it is permanently cutting those frequencies even when the sibilance is not present.

[Click here to find out a good method for finding the correct frequency with an equaliser.]

A de-esser is a bit like a compressor crossed with an equaliser... it only compresses the input signal within a user adjustable frequency range and only when the level in those frequencies passes a certain threshold. They usually have a listen function so that you can hear and set the frequency range that will be compressed. If you sweep across the frequency range, the correct (problem) frequency will be at the point that most compression takes place, ie. the loudest point. When you have found the right frequency, you can set the threshold by ear. It should be fairly obvious when enough has been cut and the vocal sounds correct... over de-essing results in a 'lisping' vocal.

top

vocals  •  effects

...different musical genres dictate the use of different parameter settings on effects.

When you have a good vocal sound, it's easy to improve it further... using effects. Virtually all vocals will benefit from a little of either reverb, delay, or both. Again, different musical genres dictate the use of different parameter settings on these effects. For instance, the reverb decay time on vocals used in Trance is often much longer than those used in House. It is important not to swamp the vocal in these effects though - it must be allowed to 'breathe'. Make sure that the reverb pre-delay is set to more than 20 ms or so, so that the effect doesn't obscure the all important initial transients in the vocal. As always, experiment and see what sounds best.

There are also several popular ways to 'fatten' or 'thicken' a vocal sound. The main one is known as 'double tracking' and is simply recording the singer performing multiple takes and overlaying a few/several of them together, with one main central vocal track and other, secondary vocal tracks played quieter and panned left and right to varying degrees. This is most commonly used in song choruses, but can be used continuously with good effect if the levels of secondary takes are low enough.

If it is not possible to record extra passes, the double tracking effect can be roughly achieved in other ways from just one vocal recording. For the main method, you will need to have at least one spare mixer channel (either hardware, or virtual in a sequencer) and be able to send the vocal signal to it/them. To emulate the tiny imperfections in time and pitch of the secondary vocals, you need to add chorus, tremolo and/or delay effects. If your mixer can add a delay to the secondary vocal channels then you will not need to use delay effects... if not and you do use delay effects, turn the feedback off and the wet/dry mix to all wet.

The delay time of these channels (secondary vocals only) should be different to each other and less than around 40 ms. According to the Haas theory, delay times of more than this will be perceived by the human brain as two separate sounds, so if that it your intention, use times of over 40 ms. Note that when delaying the secondary vocal channels, the stereo image will appear to move. You will need to use the mixer pan controls to compensate for this and ensure that the overall vocal sound comes from the centre of the stereo field.

[Click here to find out about the different types of effects used in music production.]

The chorus wet/dry mix should also be set to all wet and the feedback, if available, should be set to 0. Again, for best results, set the other parameters by ear and make them different for each secondary vocal channel. If using tremolo, the frequency and its wet/dry mix should be set at points where its effects are minimal.

[Click here to find out other ways to thicken a sound, using chorus or flangers.]
top

mixing

mixing  •  levels

For a high quality mix, levels are important.

For a high quality mix, levels are important. If you have a hardware mixer, channel gain/trim controls should be set so that optimal input levels are achieved. This can be accomplished in a few simple steps. First raise your channel and master faders to unity (0 dB). Play the sound through the channel, while slowly increasing the gain, until the individual channel meter or master meter is showing peaks just about reaching the 0 dB mark if analogue and between -10 dB to -6 dB if digital. When you have set the gain to the optimum level, you can reduce the channel fader to a more appropriate level.

[Click here to find out more about levels in digital audio and here to find out about mixers.]

If you are using a virtual mixer in a computer sequencer, it is important to ensure that the levels feeding your audio interface are sufficiently high... increasing the gain inside the computer is too late, as the crucial point to have a good level is just before the analogue to digital converter in your audio interface.

During the mixdown, it is natural that faders will be raised, but be careful not to fall into the trap of slowly raising all of the faders over time, as this achieves nothing. It is easily done, but you must remember that not all of the sounds can fit at 'the front of the stage'. Try to use sounds that fit into different frequency ranges, that will complement each other when played together, instead of fighting to be heard. If you do have sounds that have frequency ranges that overlap, you can pan them to different sides of the stereo field and give them space that way instead, or use an equaliser to cut some of the clashing frequencies.

[Click here to find out equalising techniques.]
top

mixing  •  spectral balance/curve

...a certain spectral balance must be achieved for a mix to sound right.

As the human ear perceives frequencies logarithmically, a certain spectral balance must be achieved for a mix to sound right. For dance music, this basically means that the bass frequencies must be loudest (around 100 Hz), with a gentle curve lessening the level of higher frequencies gradually by around 12 dB at 8 kHz, then slightly more rapidly above around 12 kHz. This can be slightly different depending on the musical genre, but it is generally around the same. If you have a spectrum analyser, as well as viewing your mixes' spectral balances, it is a good idea to also play well produced tracks that others have made through it and try to approximately match their spectral curves.

top

mixing  •  compressing

There are many situations when compression is beneficial, or can even be necessary. For instance, it can be used to raise the quieter, lower frequencies of a bass drum tail for a fuller sound, to raise the average, or 'perceived' level of a whole mix, or to keep quieter vocals audible and louder ones from standing out too far. In fact, taming unwieldy vocals almost always warrants the use of a compressor.

[Click here to find out about compressors and here to find out out compressing vocals.]
...always carry out comparisons between the compressed and uncompressed sounds, as sometimes compression won't help...

When using a compressor, your ears should be able to hear whether the sound has benefited from the processing, or not. One important tip is to always carry out comparisons between the compressed and uncompressed sounds, as sometimes compression won't help and can occasionally even have a detrimental effect on the sound source. When performing comparisons, it is crucial that you adjust the make up gain to a position where the levels of the input and output signals are equal. This way, you can hear the difference in sound and not just the difference in level... your ears will usually tell you that the louder one sounds better although it may not actually be better! The make up gain can be set to an optimal level when you are confident of an improvement.

The attack and release controls should be set according to the input signal. In general, but not as a hard and fast rule, sounds with fast amplitude attack times, such as drums and percussive instruments, benefit from fast compression attack and release times, whereas longer compression attack and release times better suit sounds with slower amplitude attacks, like pads and strings. It is usually a good idea to use the automatic release time circuitry, if your compressor has it, as it should give a more natural sound because it will self adjust according to the input signal.

The ratio control should be set according to how much the signal needs to be compressed. To approximate this amount, take level readings at the quietest and loudest points when the sound is playing and take away the upper number from the lower number. Often, using a high ratio will determine that a high threshold should be used and a lower ratio will need a lower threshold in order to achieve the same amount of compression. Lower ratios often produce more 'transparent' processing and are used to smooth and raise the average level of a signal, whereas higher ratios are suited to more severe dynamic reduction and can produce more apparent effects, such as 'pumping', when you can hear the compressor working.

top

mixing  •  noise gating

Traditionally, gates were used for clearing unwanted signals, such as noise and hum. When used for this purpose, the attack should be set to 0 so that when the gate opens, it doesn't remove the input signal's initial transients. The threshold should be set at a level that is as low as, or lower than, the quietest part of the wanted signal, but higher than the loudest level of the unwanted signal and the release control should be set at an appropriate level to ensure that the tail of the wanted signal is not cut short.

[Click here to find out about noise gates.]
As most gates can also be triggered from a 'side chain', or key input, they can also be used to produce rhythmic effects.

As most gates can also be triggered from a 'side chain', or key input, they can also be used to produce rhythmic effects. For instance, if you process a pad or vocal sound with a noise gate that is being triggered by its key input, you can have the sound turning on and off in any rhythm and any style that you like. Ideally for this, the signal being fed to the key input should have a fast attack so that the gate will open when it is supposed to. A basic synth sound with no filter, attack or release is good to use because you can then specify exact lengths of time for the gate to remain open. Experiment changing the sonic envelope using the gate's attack and release controls.

If your gate has a range control as well as a key input, then you can create even more interesting results. For example, if you send your bass drum signal to the key input and process a long evolving sound with the gate, you can create a pulsing effect in time with the music, which will be dependant on your settings for the range, attack and release controls.

Another popular trick using the key input is to process the output signal of a reverb processor to create a gated reverb effect. This is when the tail of a reverb effect is abruptly stopped and way before its natural end time. To create this effect, use a long reverb decay time... the longer the time, the fuller the initial reverb reflections and the more pronounced the effect will be. If you use this reverb on, say a clap sound, write a MIDI part in your sequencer that will play a basic synth sound into the gate's key input whenever the clap sounds and then extend the length of the notes to the desired length of the gated reverb effect. Rhythmic gating can be used on just about any source and with any rhythm, so possibilities are endless.

top

mixing  •  equalising

An equaliser can make a huge difference to a sound. In general, boosting mid to high frequencies (around 4 to 6 kHz) will bring a sound forward on the sound stage and make it stand out in a mix. Too much high frequency boost on a sound will end up making it seem thin, so care must be taken. Equal care must be taken with deep bass (less than 80 Hz), as too much of that can make a sound or mix lose clarity. It is important not to just boost most of the frequencies, as that just makes a sound louder and doesn't benefit it tonally. It is rare that you will need to boost a frequency range more than six or seven dB's or so... if you find that you do, it may be better to improve the sound at source.

[Click here to find out about equalisers.]

The equaliser is also an excellent tool for correcting any unbalanced sounds with resonant frequencies, such as vocals with excessive sibilance. Having said that, it should be stated again that it is always best practice to get your sounds sounding as good as possible at their source, which may entail re-recording or going back to program your synth further.

One useful technique for finding problem, or 'sweet' frequencies, is to use a parametric equaliser with a very narrow bandwidth (a high Q value).

One useful technique for finding problem, or 'sweet' frequencies, is to use a parametric equaliser with a very narrow bandwidth (a high Q value). When the frequency control of this equaliser is swept across the frequency range, the resultant signal will become excessively loud at the troubling frequency, ie. it is easier to locate a problem frequency when it is amplified. When the frequency has been located, cut the gain significantly and slowly increase it and possibly widen the bandwidth, until it sounds right. When boosting using this method, it is usual to increase the level until you hear the frequency range stand out and then reduce it by a few dB's, so that the sound becomes enhanced, rather than overwhelmed.

When cutting frequencies, a narrow bandwidth is usually required, whereas boosting generally benefits from wider bandwidths. It is important to realise that the human ear often confuses something that sounds louder for something that sounds better, ie. a sound that has had an equaliser boost applied will (usually) appear to sound better than the original sound. For this reason, it is good practice to compare your equalised and original signals at equal levels. Many equalisers will have some kind of gain control for this purpose... remember that the point of equalising is to correct or improve the frequency balance of a sound, not to increase it's level.

Mixes can easily become 'muddy' when too much, often unneeded, bass clouds activity in the lower frequencies.

Mixes can easily become 'muddy' when too much, often unneeded, bass clouds activity in the lower frequencies. It is common practice to remove the excess bass using high pass filters or low shelf equalisers. If you can't hear the lower frequencies of a sound in the mix, you won't miss them when cut, but as bass uses up a lot of energy, its disappearance will allow overall levels to be raised, ie. generally, a sound with less low bass will appear louder than the sound with its full bass at the same dB level. It is also common to cut 'clashing' frequencies in a sound, so that it will fit in better with other sounds in the mix.

One last important fact to mention about equalisation is that apparent frequency boosting results can be achieved by actually cutting frequencies elsewhere on the frequency spectrum. If you want to hear more treble in a signal, you can cut the mid, or high-mid range with a narrow bandwidth equaliser. Likewise, you can raise the apparent bass level by cutting in the mid to low-mid range. Experiment with cutting different frequencies to see which frequencies it enhances. Also, try sweeping a high cut, very narrow bandwidth equaliser across the frequency spectrum and listen to the effect... it sounds almost like a phaser effect and can be used with good results if it can be automated in a sequencer.

top

mixing  •  effects

Effects play a big roll in electronic music. The kinds of sounds that we hear on records and lust after are quite often fairly basic sounds that have several choice effects added to produce a more pleasing and original sound. The trick to making a sound sound right with multiple effects is to not over do it... keep your eye on the wet/dry mix - modulation-based effects, such as a flanger, never really need more than a 50/50 split, often nearer 30/70 and time-based effects, such as reverb, usually sound right set at less than 30/70. In particular, using too much reverb or delay can have the effect of reducing the apparent size or closeness of a sound.

[Click here to find out about different types of effects with audio examples.]
With the right parameter settings, flangers and chorus effects can thicken an input signal...

With the right parameter settings, flangers and chorus effects can thicken an input signal and/or emphasize particular frequencies in it. The LFO frequency controls the apparent thickness, with faster rates equalling a thicker sound and the range or depth affects which frequencies are enhanced, with smaller depths enhancing higher frequencies. Feedback is usually not required for this thickening effect.

There is a nice thing that you can do with reverb that works extremely well with vocals, but also sounds good on just about any sound. It's when you hear the reverb build up to the start of a sound instead of, or as well as afterwards. First you need to record to audio the start (first few seconds) of the sound that you want to have the effect on and reverse it. Then you put a suitably long reverb effect on its mixer channel and record just the effect. When you reverse the reverb's audio file and place it in front of the sound, the job is done.

Another popular trick is to control various effect parameters, either with MIDI, if using hardware effects units, or with automation if using computer based effects. This way, you can guarantee that your flange will reach its peak at the right time, or you could raise the feedback levels on your delay at different points throughout your song. Possibilities are endless: have different effects coming in and out at different places in a bar, increase the wet/dry mix of effects during the song's chorus, increase a flanger's feedback in an instrument's solo section, etc.

top

digital audio

digital audio  •  levels

There are several things that everyone interested in music production should know about digital audio. The first and most important point to note is that unlike old analogue equipment, where the volume went from say, 0 to 10 on an amplifier, or from -infinity to +10 dB or more on a mixer, comparative digital equipment goes from - x dBFS (decibels full scale) depending on the quality of the unit up to a maximum level of 0 dBFS. Digital mixers may show the same level scale as their analogue counterparts on their interface, but their output can never pass 0 dBFS.

...when levels in a digital system are pushed into the red, serious digital clipping occurs...

Also, levels on old analogue mixers and tape machines were often pushed into the red to get pleasing saturation/distortion effects that were brought about by various components in the machines. (Most) digital equipment does not have these components and furthermore, when levels in a digital system are pushed into the red, serious digital clipping occurs, where the top of the audio waveform is 'cut off' and errors are introduced into the sound. This can sound extremely unpleasant and is the reason why you should keep an eye on your levels. This should not be confused with analogue (or emulations of analogue) clipping, which can actually improve a sound... click here to find out more.

top

digital audio  •  bits

Digital audio is stored as a continuous series of binary numbers and a bit is one digit of one of these binary numbers. The higher the number of bits (called bit depth/bit resolution), the longer the binary word and therefore, the higher the maximum value of the equivalent decimal number able to be produced. In an audio system, this maximum decimal value refers directly to the dynamic range able to be achieved by the system.

A 16 bit audio system will have 65,536 possible values that relate to levels of amplitude.

A 16 bit audio system will have 65,536 possible values that relate to levels of amplitude. The process of assigning a sample with the value that is closest to that of the input signal at that moment is known as quantisation. As audio signals have both positive and negative components, these values are converted to give a range of -32768 to 32767 amplitude steps. In a perfect system (none exist), this equates to a maximum reproducible dynamic range of 96 dB. A perfect 24 bit system (none exist) should be able to reproduce a 144 dB dynamic range, although around 110 dB is more usual in real products today.

The human ear is capable of differentiating between sounds across a huge dynamic range of somewhere around 120 dB, with between 120 dB and 130 dB being our threshold of pain and enough to damage and/or perforate the ear drum. That equates roughly to the difference between the sound of a pin dropping some distance away in a large silent room to the roar of a nearby jet plane, or an orchestra playing at their peak. Looking at the figures above, it is clear that current audio bit depths can't actually reproduce the full dynamic range that we can hear.

It therefore seems obvious that digital audio has one final progression to be made. With high frequencies well over the human frequency range already being achieved today with 96 kHz and higher sampling frequencies, all that is left to attain is being able to digitally capture and reproduce sounds over the whole 120 dB dynamic range or more. Enter 32 bit audio, with a theoretical maximum dynamic range of 192 dB which should mean that even modestly made devices could easily achieve 120 dB ranges... it's only a matter of time.

top

digital audio  •  sampling frequencies/rates

Sampling frequencies are measured in Hertz (Hz), or more accurately, kilohertz (kHz). One Hertz refers to one cycle per second and 1 kHz is the same as 1,000 Hz, ie. 1000 cycles per second. It is important not to confuse sampling frequency with the human's audible frequency range which extends between 20 Hz and around 20 kHz.

The common 44.1kHz CD standard sampling frequency has 44,100 tiny recordings of audio for every second.

The common 44.1 kHz CD standard sampling frequency has 44,100 tiny recordings of audio for every second. This may at first seem like an adequate amount, but it means that signals near the top of our frequency range (11 kHz +) are distorted to a small degree, with only a couple of samples to describe their changes in amplitude and frequency. Often, producers make their music at higher sampling rates these days, with 48 and 96 kHz being the most common.

48 kHz audio will have approximately an extra 1 kHz of clarity, ie. frequencies above 12 kHz will very slightly start to lose definition, although it should be pointed out that this would be hard to notice. 96 kHz sampling frequencies can reproduce all frequencies audible by humans and considerably more. There should never be any need for higher sampling rates although some professional audio equipment already have 192 kHz rates and higher... of course, manufacturers will always tell us we need newer, improved systems.

top

digital audio  •  dither

As digital audio is comprised of a series of numbers, any processing that is done to the audio results in mathematical calculations being performed by the computer, digital mixer, etc... the more accurate the results are, the cleaner the audio signal will be. However, most of these calculations result in additional decimal places and therefore, a need for greater binary word lengths (bit resolutions) to represent them. This is why most digital sequencers, mixers, etc. have higher internal bit depths than the bit depths of their inputs and outputs and often use internal dithering before re-quantising the signal to be sent to their (digital) outputs.

In the most basic terms, dither is simply a form of low level noise... it is actually used to improve the sound quality of digital audio...

In the most basic terms, dither is simply a form of low level noise. Strangely and unlike analogue noise, it is actually used to improve the sound quality of digital audio, whenever it is quantised, or re-quantised (recorded, or had its sampling or bit depth altered)... most good ADC's (analogue to digital converters) add dither to their inputs as standard. As previously explained, digital audio is (currently) not a wholly accurate representation of its analogue comparison. There are always parts of an audio signal that do not exactly fit into the rigid digital steps, known as quantisation. For example, a 16 bit recording will have 32,727 possible amplitude values, but the analogue signal that was recorded had an infinite number. When a sample is taken whose amplitude does not exactly fit into one of the 32,727 steps, quantisation errors, or distortion occurs.

Keeping this simple, when dither noise is added to a signal during the quantisation/sampling process, it allows for a greater depth of clarity to be achieved, by effectively spreading the many quantisation errors over the whole audible frequency range. Furthermore, dither also allows for the capturing of extremely low level parts of the signal that would otherwise not be loud enough to register in the least significant (smallest value) bit. It does raise the noise floor (by approximately 5 dB), but is usually unnoticeable, particularly when certain types are used.

There are a few different dither types, which use noise with different frequency balances and are often referred to as either coloured, or filtered noise, or sometimes as noise shaping. Two of these which are very popular are the POW-R consortium, LLC's POW-R dither (type 3) and the Apogee Electronics' UV22 signal, both of which avoid placing the noise at frequencies that the human ear is most sensitive to (roughly between 1 kHz and 5 kHz) and are available in many software programs.

To summarise, dither should and should only be used when digital quantisation or re-quantisation is performed and usually only when processing files, ie. not during the mix, but on the finished mix file. Most ADC's already have dither systems, so you only really need to use it when processing an audio file with around 20 bits or less, or when converting a larger format file into a smaller format file, ie. 24 bit, 48 kHz to 16 bit, 44.1 kHz for CD - there is no benefit from converting a smaller format file into a larger format file. It should also be noted that cumulative dithering will have a detrimental effect, so it is beneficial to perform your mastering, processing and/or sampling/bit depth conversions in as few steps as possible... the depth conversions should always be the very last step in the process and are often best left to a mastering studio to undertake if the tracks will be released, as their equipment will be of far higher quality.

top

digital audio  •  formats

The 'Red Book' audio CD standard specifies the properties of the CD format. In particular, it uses 2 channel (stereo), 16 bit PCM, with a sampling frequency of 44.1 kHz. It can not deviate from these specifications in any way, ie. to burn to an audio CD, the file must be in a 16 bit, 44.1 kHz format. This gives CDs a bit rate of 2 x 16 bits/sample x 44,100 samples/second, which equals 1411.2 kbit/s.

The Super Audio CD (SACD) format stores data in a completely different way to the CD/DVD's PCM method. It uses Direct Stream Digital (DSD), which is only 1 bit, but with an ultra high sampling rate of around 2.8 MHz... the 1 bit effectively states whether a sample is louder or quieter than the previous one. This is supposed to be able to deliver a dynamic range of over 120 dB, covering the whole human audible frequency range and more. It is definitely a higher quality format than CD, but they don't play on most CD/DVD players and recordable machines are very expensive.

The DVD-A audio format also allows for much higher levels of quality than a CD and even surpass that of the SACD. It can use up to 6 channels (5.1 surround sound), in 16, 20, or 24 bit resolution and at sample rates up to 192 kHz. As computers and DVD-A both use PCM encoding, it is an easier format to burn music to, although good DVD authoring software can be quite expensive too. Different files on a DVD-A disc can also have different properties, ie. you could have a surround mix and a stereo mix on the same disc.

Digital audio on computers comes in an alarmingly wide variety of different, often non-compatible formats.

Digital audio on computers comes in an alarmingly wide variety of different, often non-compatible formats. High quality audio files can be very large, so there are many data compressed formats which trade lower audio quality for smaller file sizes. Compressed audio files are most commonly used when low memory capacity or bandwidth is an issue, such as music on the internet, or .mp3 players.

The main uncompressed computer audio formats are Microsoft's Waveform (.Wav) for Windows users and Apple's Audio Interchange File Format (.AIFF) for Mac users. They are both roughly equivalent in being able to accept multiple sampling rates and bit depths and as neither use data compression, they retain the full quality. Most audio software on both the Windows and Macintosh platforms will read both formats and they are equally industry standard audio formats.

MPEG-1 Audio Layer 3 (.mp3) files are very popular among consumers these days, although as they are (sometimes heavily) data compressed, they are not much use in music production. They use a clever compression algorithm that reduces the size of audio files by removing parts of the signal that, due to phychoacoustic reasoning, it deems that we wouldn't actually hear anyway. The amount of data reduction is scalable using the bit rate, with higher bit rates removing less data and therefore giving better quality. The highest .mp3 bit rate is 320 kbit/s.

... given the choice, always use either .wav or .aiff uncompressed files.

Windows Media Audio (.wma) and Apple Lossless Encoder (.m4a) are Microsoft's and Apple's own data compressed formats, but are not compatible on each other's platforms. Many manufacturers/audio software designers also have their own, non compatible audio formats, but one thing is important to producers... given the choice, always use either .wav or .aiff uncompressed files.

top

studio acoustics

Many of today's producers are working from spare rooms, or bedrooms in houses that were not designed to be studios. There are several problems that arise from this situation and the main ones are caused by sound waves repeatedly reflecting between parallel walls and becoming standing waves, or producing flutter echoes... this why many professional studios have no parallel surfaces, often having a sloping ceiling too. Different frequencies have different sized wavelengths... the wavelength of an audio sound wave can be calculated using the simple formula:

wavelength = frequency / speed of sound (343 m/s)

The wavelengths of the lowest and most prominent standing waves in a room will be two times the length, width and diagonal length from corner to corner of the room respectively. Higher frequencies with wavelengths that are exact multiples of these will also be affected, but to lesser amounts as the multiple is increased. The speed of sound actually changes with temperature, but can be summarised as being around 343 m/s at 20°C + 0.6 m/s per 1°C above that.

Whichever frequencies have wavelengths that are equal to multiples of the lengths of the room's walls will bounce back and forth exactly, reinforcing or canceling those frequencies...

When music is played in a room, the sound waves will bounce from wall to wall. Whichever frequencies have wavelengths that are equal to multiples of the lengths of the room's walls will bounce back and forth exactly, reinforcing or canceling those frequencies, dependant on whether the whole or just half the wavelength fits in the room... this principle is used in the production of musical instruments to create harmonics depending on the length of the column of air in a wind instrument, or of a string on a guitar.

In large rooms, this is not such a problem, as the wavelengths that fit the size of the room correspond to very low frequencies. The problem is that when you are working in a small room, the wavelengths that resonate will be much higher and well within the audible frequency range. This will create an uneven frequency response in the room and can be responsible for mixes that sound good there but sound wrong when played elsewhere. This effect is most clearly apparent in small shower rooms, where the reinforced frequencies are in the upper mid range.

Furthermore, as standing waves also cause sonic troughs as well as peaks, this means that particular frequencies (dependant on the size of your room) will appear louder in some places in the room and quieter in others... try walking around your room and listening in particular to the bass end. Without going into the science of it all, it should also be pointed out that there will always be an approximate doubling of bass power just in front of each wall and a quadrupling in the corners. For this reason, it is important not to put monitor speakers (except maybe a sub) against the walls and especially not in the corners. If you can not avoid this, some good speakers have 'acoustic space' settings, which you should set to half space when your speakers are against a wall, and quarter space when they are in a corner.

top

studio acoustics  •  acoustic treatment

...how can anyone produce a good mix when they can't accurately hear the music they are producing?

If your music production centre is in a room that is affected by standing waves, it is advisable to get some kind of acoustic treatment to remedy the problem. After all, how can anyone produce a good mix when they can't accurately hear the music they are producing? Bass traps are often quite large due to the long wavelength of bass frequencies, are used to 'soak up' excessive bass in a room and are ideally situated in the room corners, where the effect is most prominent. Mid and high frequencies are dealt with by diffusers and absorbers, which are usually much thinner products than bass traps and easily attached to walls. The most important places for them to be situated are directly in front of, behind and to the sides of the monitoring position, ie. where you sit while mixing.

Diffusers do not reduce the level of affected frequencies, but instead diffuse them evenly throughout the room. This prevents the build up of particular frequencies and ensures that the room has a better frequency response. However, before you rush out and buy some, it's worth pointing out that in an average room, there are other things that can do the same job, such as blinds, shelves, and other wall mounted items. Absorption is the same method that bass traps use, but for higher frequencies and it reduces the level of these affected frequencies. It is only necessary when your room has excessive mid or high frequencies, such as a room with bare plastered walls would have. Thick wallpaper has a similar effect and will stop flutter echoes.

If you record vocals, a vast improvement can be gained by placing a few acoustic absorption panels in the recording environment. They should be placed on the two walls in the corner of a room and the singer should stand with their back to them with the microphone facing towards them. This way, the microphone will only pick up the sound of the singer and far fewer reflections of the room sound, ie. mains and computer noise, etc. It is also worth noting that either mattresses and/or thick duvets will do the job equally well.

Although acoustic treatment is highly recommended, it often does not fit on the 'studio budget'. If you have none, it is important to check your mixes on other audio systems regularly, ie. in the car, on headphones, at your friend's house, etc. In addition to this, it is also important to regularly play other music through your studio monitoring system, particularly tracks that you believe are mixed well. Over time, you should be able to learn what the frequencies that your room reinforces or cuts sound like and be able to make suitable adjustments in those frequency ranges that you know are right, despite them not sounding right in your room.

top

 

 

Copyright © 2015 nitetime productions. All rights reserved.