Computer Recording

Current revision

A digital small studio

Introduction

Computer recording is a standard today among both studios and home users, even though there is some state-of-the-art studios that still use analog tapes. The evolution has developed very fast since it first was introduced round the 70/80's. In this article I will cover some of the history and the basic concepts and terms of digital recording. (there will however be some more in depth articles on the subject in the future. ed.note)

History

The art of digital recording was pioneered by Tom Stockham at MIT as early as 1961, when he was making the first digital tape recordings with the assistance of a large TX-0 computer and a A/D-D/A converter from Bernie Gordon at EPSCO. He later founded the company Soundstream that recorded the first 16-bit digital recording in 1976 at the Santa Fe Opera. However, digital technology was present even before in instruments like the Hammond organ and the Mellotron and in some effect-processors.

In 1975 Sydney Alonso/Jon Appleton/Cameron Jones developed the Synclavier digital synthesizer at Dartmouth College. They formed the New England Digital Corporation the next year (later became Demas) and sold their product mostly to the recording industry. In 1979 was the Fairlight synthesizer introduced, which actually was a workstation capable of making full recordings stored on floppy-disks.

The beginning of the digital era came when Philips/Sony made a collaboration about the standards of the compact disc in 1980-81, even though there was some attempts to make a standard before. (Betamax,Pioneer Laserdisc,PCM) The specifications were 16-bit encoding and a sampling frequency of 44,100 per second, which still is the specification of a compact disc. (CD) The first truly digital recordings were made in 1982 by Tom Jung at the jazz-label DMP, who recorded directly down to to disc.

In 1981 was the first "white papers" of the MIDI-specification released by Dave Smith and Chet Wood from Sequential Circuits (Prophet-5) at the AES fair. This was truly revolutionary and two years later, in 1983, the first synthesizers from Roland, Yamaha (DX7) and Sequential Circuits implementing the MIDI protocol. (see terminology)

The use of MIDI in home/smaller studios wasn't established until 1986 when Steinberg made their first release of Cubase for the ATARI-platform, even though there was some programs even earlier at the AMIGA-platform. LOGIC was released about the same time, and for years a battle raged between the two opponents. However, large studios was using Macintosh computers and Pro Tools already back then.

The DAT-format (Digital Audio Tape) introduced in 1986, was setting a new standard in digital mastering, but it took several years until the prices of consumer machines dropped to reasonable levels.

The next landmark came in 1991 when ALESIS released their first multitrack for digital recording based on Video tapes. 3M has made multitracks for digital recording since the beginning of the 80's, but due to it's price, there was only studios like Abbey Road, Polar Studios and others that really could afford the digital technique. The 8-channel A-DAT was a technical revolution and allowed up to 16 machines to be used in conjunction. Originally priced at 4000$, it made it possible for smaller studios to adopt the digital technique.

In 1998 there was an agreement upon the DVD-Audio Format 1.0 specifications. The capacity would be the same 4.7/8.5/9.4/17 GB as DVD-Video, but the sampling rate of 88.2 kHz and 176.4 kHz will be higher for a frequency response of 0-96 kHz rather than 0-48 kHz for DVD-Video at 44.1 kHz and 5-20 kHz for the audio CD. The Maximum Transfer Rate for Audio will be 9.6 Mbps rather than 6.1 Mbps for DVD-Video or 1.4 Mbps for audio CD.

Hard disk recording system were introduced in the beginning of the 90's when Fairlight ESP Pty Ltd developed the MFX2, the first 24 track disk recorder. It was followed in 1993 by the RADAR system which was a replacement for OTARI's analog 24-track systems. The prizes on hard-disks as well as the storage volume has continued to drop since then, and is by the time this is written not really an issue anymore.

Digital Technique

A 4-bit sampled signal

AD/DA Converters

In order to make a digital recording, the first point to mention is the converters for converting an analog sound to a digital. These are called AD/DA converters. (see terminology) In the recording process, the DA-converters are used most for actually getting the sound out to the monitors. The AD-converters is more critical since the more precise a converter is, the more will it cost. AD/DA-converters exists in all soundcards to computers, and is one of the most important issues when doing a good digital recording.

About sampling of the analog signal

In order for an AD-converter to sample an analog signal, some basic theory applies. Due to the Nyquist–Shannon sampling theorem, a perfect reconstruction of a signal is possible when the sampling frequency is greater than twice the maximum frequency of the signal being sampled. Eg. if a tone of 440 Hz is going to be sampled, the frequency of the sampler must be at least 880 Hz. Hz stand for Hertz, and is the frequency of samples taken/second or the movement of air of a note/second. Based on these theories its pretty simple to understand that the greater sampling frequency - the greater will the reproduction of the analog signal be.

Digital Audio Workstations (DAW)

A digital Workstation (DAW) is made up of two main components, eg. the part that handles MIDI-data and the part that handles AUDIO-data. To connect a DAW to an external MIDI-device. an interface is needed that allows the units to "speak" with eachother through the MIDI-protocol. When working with broadcast and other recording environments there is also a need for a DAW to be able to synchronize with external devices. Therefore it is common for a DAW to have different functions for sync. via common protocols.

Audio and ASIO

In the Windows environment the OS is not connected directly to the processor of the soundcard, eg. Windows processes audio data through several "layers" (different levels of the OS). In audio applications this is critical since the signal is simply delayed. Steinberg developed a technique known as ASIO, to connect directly to the soundcard on computers running Windows to avoid this problem. The regular soundcard in a computer is most likely NOT to handle ASIO, though there is a little application called ASIO4Free that can be used to simulate ASIO. Therefore a dedicated soundcard is essential when running DAW software.

Computers and audio

As mention before, the real critical point when recording audio is the AD-converters. Once the sound is into the digital domain, the computer does the processing of the sound before it is converted back to the speakers via the DA-converter. Some points of interest is to be considered though, when using a computer for musical purposes.

a. The speed of the hardrives. The faster a drive spins, the better, since it will take less time for the OS to pick up the data needed. A large buffer memory (see terminology) is also preferred, since the performance of the disk will be faster.

b. harddisk is divided in memory areas, when a computer saves data to disk, it is taking the first free spot available, which makes the data to be "spread" over the disk. Fragmenting solves this problem by putting similar data in memory blocks and leaving the free amount of data in one spot. It is obvious that if the data needed by the audio application resides in different spots on the hard-drive - it will take longer time to find it - it will result in heavier processing and in reality noises and scratches can be heard.

c. Dont' bother the OS with other tasks at the same time - eg. screensavers, automatic updates, network functions etc. that the computer does in the background shall be turned OFF. The less processes a computer doing while working with audio, the more processor power can be used for audio.

About buffer size

In order to provide the soundcards I/O ports with a regular stream of data, (eg. audio) the internal memory in the computer can be set by the DAW to pre-store different sizes of "buffers" in order to keep the stream constant. Otherwise, an interuption of the stream will most likely cause pops and noises in the sound. This function let's the user adjust the sound system for minimum latency - eg. the faster computer/disks/bus speed, the lower can the buffer size be set and then the latency will be less. Recommended is to set the buffer size high, and the lower and lower until it starts to crack and pop. The most DAW's have measurement functions for this matter, though.

Synchronization

Synchronisation functions in a DAW is for the normal user not of a greater issue, but for professionals and people in broadcast/movies it is most important. Before the recordings got digital, MIDI was synced to audio via a timecode that was recorded on the tape and then played back to the computer, hence the computer then "listens" to the timecode and plays after that instead of syncronizing internally. The advantages here is that even if the tape-recorder is 100% precise (in speed), the computer will be in sync anyway since it is controlled by the timecode. The same formula applies to syncronising with A-DAT's, movies, external equipments, several computers etc. There are several standards for this matter.

Plugins

In order to let third-party developers sell products for their respective platform, the manufactorers of DAW's has developed plug-in systems that let another software run in their environment. The most known systems is respectiveley RTAS (Pro Tools) and VST-compatible. (Steinberg) A lot of other DAW manufactorers also supplies these techniques, which can be seen as branch standard. Common plugin contains Reverbs, Soft synthesizers, Samplers and other effects. This open architecture makes it easy to get a DAW the way the user wants it, but there is a drawback about compatibility with 64 bit systems vs. 32-bit systems. Eg. plugs for 32-bit systems may not run on 64-bit OS. And plugs for 64-bit will definitely not run on 32-bit OS.

Recommended setup

To prepare a computer for audio recording is fairly simple, since every running program, tool etc should be kept to a minimum. It is also essential to have as much internal memory as possible. The processor shall be a native one - not a celeron/duron -type (for PC), since they operate with parallell half of the amount a native processor does. As mentioned before, standard disks will work fine BUT a large buffer memory will improve the performance. An imortant thing about the disks is, that if one is using a lot of samples etc. they shall reside on it's own disk, since the disk availaible for just pure audio like vocals, guitar etc. doesn't have to bother about the sampling streams and therefore will run more effecient. A good idea is also to keep the OS and the DAW on the same disk and dont mix this disk up with the other's. This will give us a total of three disks, which may seem a little much, but on the other hand the drives is fairly cheap today. But I will at least recommend two.

Soundcards

There are a lot of different soundcards on the market today, and the prizes are getting lower and lower due to the technical development - at the same time, of course, the high end sound cards are getting better and better. What is most critical about a souncard is the AD/DA -converters as said before. When choosing a soundcard the first thing to consider is what it shall be used for. Eg. shall it be used for just overdubbing/writing music a two channel card is just fine, but in order to record "live" events with more than one musician, a multi-channel card is desireable. In general, it is better to pay for good converters rather than many channels since the final result will be a lot better. Also channels can be sub-mixed via small mixers before going to the DAW. It is also essential when using samplers/software synths to have a good card, since the sound that is rendered from the smpling engine/synth is dependable of the quality of the card rather than the samples/sound itself.

Samplers & Soft Synthesizers

The most recording musicians today uses sampled instruments from the computer rather than having a dedicated sampler. There is also a lot of software synthesizers around that emulates the classic synths as well as new one's. The difference between a sampler and a soft synth is really that a sampler is built on sound files that are played back, and then are processed, while a soft synth is purely computer generated sound from different waveforms.

A sampler will therefore put more stress on the computer and it's devices, but it also depends on the quality/length of the actual samples. Most samplers/soft synths are used as plugins in the DAW and is controlled via an external keyboard through MIDI. Nowadays, there are really cheap keyboards that works well for this purpose and the importance of hardware samplers/synthesizer's has decreased, unless for live purpose.

Other hardware

Since the use of software synthesizers and samplers has been very popular, a lot of the computers resources is beeing used by these. Instead of using additional computers, hardware can be purchased to perform effects task, synthesizing and other thing that are processor critical. When this article are written the are two popular consumer systems, Powercore and UAD. These systems will put a lot less stress on the computer, hence it can concentrate on pure tasks as recording instead.

Terminology

AD/DA -converters
AD stands for Analog-Digital and DA stands for Digital-Analog. These converters is crucial when transferring an anolog signal to the digital domain, and also has a big impact on the sound when playing back a CD for instance, since a CD-player uses DA-converters.

ASIO
Audio Stream Input Output. A computer technoly by Steinberg that let's a Windows application connects directly to the soundcard (if it supports ASIO) by by-passing Windows normal drivers for sound, which giveas a problem known as latency. In other environments this problem doesn't exist.

Buffer memory
The amount of memory stored in the processor (or in a device's processor) that is available for "queing" the data to be processed.

Bus speed
The frequency in Mhz which the computer transfer's the data between it's devices/units.

DAT
Stand for Digital Audio Tape. A technique for storing data on tape also used regular data storage.

I/O ports
In and Out ports. Exists on a soundcard/computer and is used to transport data from and to the device.

Latency
The difference in time before a sound is played back vs. the sound in real-time. Often measured in ms. (milliseconds) Latency should be less than 6-7 ms to avoid problems with synchronisation during recording.

MIDI
MIDI stands for Musical Instrument Digital Interface, which is a technology that describes a user's interaction with a (digital) instrument, rather than performs any actual computer-sound processing.

OS
Operating system

PCM
Pulse Code Modulation.

Popular DAW's

Here is a list of the most popular DAW's and their respective homepages. Some of the extremely high-end system may not be covered, since this article is written for the common user.

Commercial
ACID Pro/Cinescore from Sony
Cubase/Nuendo from Steinberg
Digital Performer from MOTU
energyXT from XT software
FL Studio (Fruity Loops) from Image Line Software
Live from Ableton
Logic Pro/Logic Express/Garageband from Apple
Master Tracks Pro from GVOX
Orion Platinum from Synapse Audio
Pro Tools from Digidesign
REAPER From Cockos
Reason from Propellerhead
Renoise
Samplitude/Sequoia/Music Maker/Music Studio from Magix
SAWStudio from RML Labs
Sinfonia, from Realtime Music Solutions
Sonar/Project5/Home Studio from Cakewalk
Storm from Arturia
Tracktion from Mackie

Free
Ardour - Linux, Mac OS X, FreeBSD
Frinika - Java (cross-platform)
LMMS - GNU/Linux, Windows
MusE - Linux
MuseScore - Linux, Windows, Mac
Musette - Windows
Qtractor - Linux
Rosegarden - Linux