Sound recording

The sound recording is the operation which consists in transforming a its into a signal in order to be able to diffuse it or to keep a trace of it in a durable way on a support in order to be able to repeat it. The technologies appeared with the turning of the 20th century made the use of sound recording increasingly current. Thus the Telephone, the Radiophony and the Cinéma have recourru with many processes different of recording. At the same time appeared a divergence between the final goal and a complexification of the methods.

Currently the methods of durable recording concentrate on the support Analogique like the Magnetic band or the Disque vinyl, and the support Numérique like the Compact disk, which makes it possible to diffuse with nearest to identical and possibly to modify it (to treat it).

The operation is carried out thanks to a chain of transducers: a device of capture of the sound waves (Microphonous), a device of amplification (preamplifier) and a device of writing on the support. One also qualifies sound recording the support recorded.

History of the processes

August 1st

Since the origins of sound recording, only three techniques were employed: the mechanical recording, the electromagnetic recording and the numerical recording.

the mechanical recording was the first form of the sound recording and also simplest, invented by Thomas Edison and bearing the name of Phonographe: a mica membrane interdependent of a stylet is driven by the sound waves and serious directly these vibrations on a cylinder moving rotary, leaving a furrow being rolled up on the circumference of the cylinder. It is before the letter the typical example of the recording Analogique. The characteristic of the gramophone is that it is reversible: the same mechanism is at the same time used to record and reproduce the sounds. A new evolution, always based on a mechanical direct recording will lead to the disc and the Gramophone.

The arrival of the electricity and the control of the electromagnetic fields will initially allow the electromagnetic recording on wire then on band, implying the invention of the Microphone and the possibility which results from this to transform the sound into electrical signal and of being able to transport it in a wire. The door is from now on open for the multitrack recordings (simultaneous signals several microphones recorded on several parallel tracks of the same magnetic band), the Stéréophonie, the mixing and the Montage. The recording direct any more but is not differed. The final support of diffusion can be the Magnetic band, the disc Microsillon or the cassette. The broad diffusion of this technology will mark as from the years 1950 the commercial rise of the music recorded with an enormous planetary production of 33 turns, 45 turns and K7.

Years 1970 will be the years of glory of the K7 which will see the quality of its magnetic medium evolving/moving, since ferromagnetic oxide while passing by chromium dioxide, ferro-chromium, cobalt, to pure metal and of innumerable patents (Super Avilyn, Epitaxial, Phase Precision…), outcome with a support of great musical quality. To note that the success of the cassette is partly due also to the considerable contribution of the circuits of reduction of noise Dolby which have, as of its invention, equipped all the tape recorders and readers with K7. The most powerful magnétocassettes of the last generation had the Dolby thus has, Dolby B, Dolby C, Dolby S, the Dolby HX-Pro as of additional circuits of specific treatment of acute (of which the restitution always constituted the weak point of the K7).

It is in the middle of the years 1970 that popularizes little by little the numerical recording . This one will be initially used only by the professional studios and the final support will continue during a time to be the microgroove, simultaneously with the progressive appearance of the CD (invented in 1979, produced starting from 1982). The end of the year 1990 will mark the explosion of the Home studio, allowing whoever, with a moderated investment, to carry out recordings of quality: engravers of CD, recording DAT, recorders of numerical Diskette, recording direct-to-disc (direct on hard drive) with a personal computer.

The sound recording

In the actual position of technology a sound recording professional at least requires a couple of microphones (for the stereophonic sound recording ) connected to a table of mixing or a preamplifier connected in its turn to a recorder which can be either a numerical tape recorder DAT, ADAT or MANDELEVIUM, or a computer with the function direct-to-disc (direct on hard drive).

Until the invention of the Electronic music, there existed only one manner of capturing sounds, it was by using a Microphone . This method of sound recording remains absolutely unchanged since its origins, even if the supports of recordings multiplied and evolved/moved. It is still today the only possible method to record the voices, the noises, the sounds natural and the musical instruments acoustics, in short any sound wave conveyed in the air.

It is to say how much quality the microphones revêt a paramount importance. In the equipment of professional studio of sound recording, it is one of the element most important and it is often one of those which are most expensive. The sound engineers know it: a perfect signal upon the departure guarantees a recording of quality.

In the sound recording professional, the quality of the Microphone S is not enough. The taker of sound must also find the good placement of the microphones, and certain conditions of recording are often sources of true headaches. If the sound recording Monophonique can seem rather easy to implement, it is not the same sound recording stereophonic known as " naturelle" ; i.e. that which uses a stereophonic base (2 microphones laid out in couple) principal and several auxiliary microphones of which the goal is to specify certain sources. One of the main difficulty resides then in the optimal placement of the principal couple, because any error could not be any more corrected at the time of the phase of post-production. This difficulty is multiplied by ten when one multiplies the stereophonic bases with an aim of simulating effects of plans sound. In addition, since it is a question of multiplying the primary sources (sound recording of proximity), other phenomena intervene. Indeed, more one increases the number of microphones, and more pose problems of rotation of phase at the time of their mixture. In addition the fact of bringing closer a source to a sensor modifies its response curve in frequency: it is what is called the effect of proximity which tends to amplify the low frequencies; effect all the more marked since the microphone is directing. When it is a question for example of recording a Orchestre, the placement of the microphones and then proportionings on the level of the Table of mixing can require hours of preparation and tests which one cannot make the economy even at the time of a multitrack Enregistrement allowing to record each instrument individually and to carry out the mixing later on. One speaks then about post-mixing of Montage or post-production.

The sound recording at the era of electronics

Since the invention of the Electronic music, sound recording entered one new era because the generators electroacoustics, that it is about Synthétiseur S, of sampler S or of charts its of computer directly produce the sound in the form of electrical signal, which can directly be recorded while being freed from a stage, that of the Microphone. It is however necessary to stress that the sounds produced electronically are not natural sounds and the field of application is especially directed towards the production of current music, Bruitage S, sound effects and to a lesser extent of Film music.

At the era of electronics, sound recording can be carried out directly starting from the electrical signal. However certain sources have a signal with quite particular characteristics. Like the Radio operator FM, in a sound recording of Radio operator FM, there are very often a very important weakening starting from 15 Khz (limiting frequency) and a peak to 19 Khz (stereo). Therefore it is recommended that sampling is of 32   000 Hz (2×16   000 Hz).

Supports of the recording

Analogical supports

Numerical supports

  • Tape recording
    • DIGITAL Audio Types (DAT)
    • DIGITAL Compact Cassette (DCC)
  • mechanical Enregistrement
    • Prototype numerical disc of optical Thomson
  • Enregistrement
    • Compact disk (CD)
    • Minidisc (MD)
    • Compact disk super audio (SACD)
    • changeable DIGITAL disc (DVD)
    • general-purpose numerical Disque of high density (HD-DVD)
    • Disque Blu-ray (Blu-ray)
(list to be supplemented)

Features of the recording

In the case of an analogical recording as in that of a numerical recording, to transform a sound wave in variations of tension or computer data, introduced limitations due to the means of recording.

Characteristics of the analogical and numerical systems of recording

Dynamic

“Dynamics” or “dynamic range” of a system of recording is the term used to indicate the difference which exists between the amplitudes minimum and maximum that each link of the system can reproduce or record. That generally starts with the microphone, which converts the sound into electrical signal, to go until the support of the recording, disc, magnetic band computer… You point out that the decibels express a report/ratio. In the case of dynamics, the value given corresponds to the possible maximum level without distortion compared to a reference which varies according to the type of apparatus. In the case of the microphone, this reference corresponds to the level of the standard atmospheric pressure whereas for the electricals appliance it is often the basic noise level which is taken as reference. In the case of the numerical one, dynamics corresponds to a theoretical value given by the formula 20 X Log (2^n) where N corresponds to the quantification of the signal. In practice this value is undervalued by various phenomena (noise of quantification, precision on very low levels… etc). Here some examples:

  • Disc vinyl: 65 dB

  • Magnetic band: 55 dB
  • Acquisition 16 bits (CD): 96 dB (theoretical)
  • Acquisition 8 bits: 48 dB (theoretical)

A symphony orchestra can exploit of the sounds a range going up to 110 dB. For this reason, one uses systems which compress dynamics, the Compresseur audio in order to prevent that the strong signals are not chopped and the weak inaudible ones. This limitation/compression can be carried out with the sound recording, during the mixing, of the mastering or of each one of these 3 stages. On a tape recorder professional 1/2 inch turning in 76 cm/s and equipped with Dolby SR, one can obtain a report/ratio S/B exceeding 100 dB!

Background noise

In addition to being less powerful than the human ear, the systems of recording have also the defect to emit their clean Bruit. That can be the friction of the play-back head on the disc vinyl or the whirr of the amplifier. This noise, generally rather weak limit the dynamics of the apparatus downwards. It is perceived most of the time like a whirr or a breath which gets along well with a helmet of good quality. This breath resembles a noise of water fall because it has a spectrum which covers all the audio band. I.e. it contains all the frequencies perceptible by the human ear (of 20 Hz with 20000 Hz)

Distortion

All the apparatuses and systems which contribute to record a signal comprise like filter S or generators and in fact introduce distortions signal. Of course, the chain of reproduction does not derogate from this rule and brings it also its batch of distortions and noises, since the play-back heads, the préamplis, the amplifiers until the Haut-parleur S while passing by the cables and connectors.

These filters have an important effect on the phase and frequency response of the signal of which they modify the spectrum. These deteriorations depend on the Fréquence. But the apparatuses of reproduction can also behave in genuine generators and add frequencies which do not exist in the signal of origin; it is what occurs for example in the case from the harmonic distortion where the apparatus superimposes on the signal of origin of under multiples of this one (harmonics).

In audio, the principal types of distortions are the harmonic distortion, the distortion of crossing, the transitory distortion of Intermodulation, and the small last: the thermal distortion.

Characteristics specific to the systems of numerical recording

Noise of quantification

The Quantification necessary to the introduced numerical recording of the noise. It consists in transforming each instantaneous value of the signal sampled into a number included/understood in a scale of values spaced with regular intervals. Most of the time, it will be about a whole scale, and a value will correspond to a level of Amplitude, these levels being linearly distributed. Any representation of a physical size in the form of an integer will be always only one approximation of this one, and the difference between the quantified value and the actual value represents the added noise.

The quantity of added background noise is closely related to the depth of representation chosen, or the quantity of information which one chose to use to represent a sample. Calculation is done according to a signal-to-noise report/ratio, expressed in dB, when the signal uses all the scale of amplitude it has. The depths of numerical recording are often expressed in bits, which corresponds to the number of binary digits that one will use to represent a value. Most current are:

  • 8 bits (vocal recording, little used nowadays, was the prerogative of the first charts its)
  • 16 bits (most universal: Audio CD)
  • 24 bits (used by the professionals and Audiophile S, because it authorizes more sound flexibility processing by limiting the loss of quality, while offering large a Dynamique)
  • 32 bits (seldom used)

It is also current to record in a format in Floating decimal point, which allots more precision to the low values and facilitates calculations.

A recording in 16 bits will have a broad range of values of 65  536 units and a constant stage of 1. The signal-to-noise report/ratio can be deduced: 20 \ log_ {10} \ frac {65536} {1} \ simeq 96 dB

It is recommended to reduce the impact of the quantification by artificially adding Bruit to the quantified signal, in order to eliminate the distortions at the price from a light additional background noise, but more " naturel". (see Dithering , Noise shaping )

The noise is particularly made hear on CD in the passages of low level. Indeed, the constant and linear quantification being, the low acoustic levels are cut out in a coarse way, the discrete values for the décrires are fewer. When one has 16 bits to describe a sample, that is worth only on the totality of the dynamic beach which is to the maximum of 96 dB. However on a low level of about 20 dB, only the 3 or the first 4 bits are significant, increasing by as much the error of quantification and consequently the noise. They are the limits known and difficult to correct CD. Only a beach of values of larger quantification (20, 24 or 32 bits) makes it possible to be freed from this defect, particularly audible in the classical music.

Noise of sampling

See sampling for more detailed information on the problems arising.

Sampling consists in cutting out the sound information which arrives in electric form (thus analogical) at regular time intervals. The speed to which the chart its records points is the frequency or gives rhythm sampling. The standard of the Audio CD fixes the sampling rate at 44100 Hz, which means that 44100 values are captured each second. If one does not take a precaution before sampling, it may be that one introduces a strong background noise even a denaturation of the signal, because the frequencies higher than half of the sampling rate will not be restored but, and it is more awkward, will be reflected on the lower frequencies.

It is necessary to make so that all the frequencies higher than half of the frequency of acquisition are eliminated, or else the restitution of the sound will be polluted by parasitic sounds (Bruit and interferences between the sampling rate and the frequency of the sound). In the case of the rate of the Cd-Audio of 44,1 Khz, one does not need sounds of frequency higher than 22 Khz (according to the theorem of Shannon).

To strongly attenuate the awkward frequencies, one uses filter S. “Filter” is a rather broad term which indicates an apparatus able to retain or to filter part of a sound. One uses for example a low-pass Filtre to attenuate the high frequencies, inaudible but awkward for acquisition.

Data compression

More and more, of the algorithms of data compression like MP3 or Ogg Vorbis are used to gain an invaluable place on our supports of recording. These algorithms are known as “destructive” because they destroy part of the aural signal to reduce (as much as the selected flow imposes it) the size of the file its final. The programs of compressions use a model of the human ear in order to eliminate useless information. For example, if two frequencies are close one to the other, weakest could be eliminated if she is regarded as masked by strongest. For this reason, one finds on Internet a certain number of tests and recommendations on the manner of using this software to avoid an audible degradation of the recordings. It comes out from those which the author consulted that many coders MP3 systematically filter the signals higher than 16 Khz and limit themselves to a “bit misses” (translated by “flow”) by 128Kps. These figures prove largely insufficient in many cases if one wishes to preserve a quality of listening close to that obtained with CD. On the other hand the systems of data compression like gzip, bzip2, lha or zip do not deteriorate the data but have lower compression ratios. It is necessary moreover, to decompress all the file before listening to it, which rather badly adapts walkmans and other readers. There exist all the same nondestructive formats of audio compression like the format ape or flac, which can listen to same manner as those with the format MP3 or Ogg, but the compression ratios weaker (are of course divided by 2 instead of 10 on average), and few audio readers support them.

Whereas the memory supports become increasingly accessible and offer almost unlimited storage spaces, one can legitimately wonder whether technologies of sound compression still have future. For the next years, the answer is clearly yes, and for (at least) two essential reasons:

  • sound compression brought new uses with it: rather than to store 200 albums on a support, which enormous was even compressed at the beginnings of this technology, one can now store 2  of it; 000, and it then becomes interesting to refer them in a database to facilitate the search for pieces and the diffusions sets of themes.
  • the data compression does not bring only one profit of storage space, it also makes it possible to reduce the data flows, and thus to contribute to reduce the costs of the components for the reading (in the walkmen for example) and the transfer times whatever they are.

See too

Random links:Crane (constellation) | Gellainville | Yesterday Went Too Soon | Nemrod (resistance) | Adventures of Marco Polo | Nouvelle_plage_de_Smyrna,_la_Floride