Digital audio has a remarkably short history, but apart from a few enthusiasts of legacy technology, the takeover has been total.
Digital audio history probably began when Pulse Code Modulation was invented by Alec Reeves in the 1940s. As is often the case, the idea was good but the technology to implement it was some way off.
Convertors capable of acceptable sound quality were not easy to make and a further problem was that the data rate that came out of an audio convertor was beyond the capabilities of the recording technology of the day.
Today when we take for granted the ability of IT systems to handle huge quantities of data, we forget that there was a time when digital audio recording formats such as CD and DAT out-performed computer recording by a large factor and computers adapted audio recording formats for data storage purposes.
Audio convertors connected to minicomputers became feasible in the 1960s and were used for research purposes. Analyzing airplane and vehicle noise, for example. The data from the convertors would be stored on hard disk drives but the economics were such that it wasn’t feasible for music production.
One early digital audio device that became popular was the delay system needed by vinyl disk cutters that allowed the groove pitch to get coarser ahead of a loud passage. Previously, special tape decks with dual playback heads were needed to provide the advanced sound for pitch control, but a digital delay allowed an ordinary tape deck to be used.
One solution to the high bit rate of digital audio, not far short of a million bits per second for one channel, was to adapt video recorders that were just starting to become available. The video input was fed from a gadget called a PCM adaptor that output two level signals, effectively black and white for 0 and 1 and converted the TV picture into a checkerboard so that data could be recorded.
The first of these was built experimentally by NHK and demonstrated in mono form in 1967. This used a quadruplex tape deck running 14-inch reels of 2-inch wide tape weighing 10Kg. The stereo version was demonstrated in 1969. Denon produced a system based on the NHK machine for vinyl disk mastering in 1972. The machine weighed 400Kg.
Reduced Error Rates
The dynamic range of a video recorder wasn’t really necessary for recording binary, but it did mean that the error rate was quite low, and a simple error correction system would do the job. In due course as VTRs became smaller and industrial and consumer VTRs became available, PCM adaptors became popular and for many people may have provided the first experience of digital audio recording.
For simplicity, a whole number of audio samples was recorded on each line and as the line rate was fixed by the video format, the sampling rate was influenced by the video signal. In a number of PCM adaptors, there were three samples per active line, 245 active lines per field, and a field rate of 60 Hz. That led to a sampling rate of 44.1KHz being standardized for PCM adaptors.
Sony announced the PCM-100 in 1978, and in 1982 launched the PCM-1610, which was intended for Compact Disc mastering. These machines worked with U-matic cassette recorders. The Compact Disc would adopt the sampling rate used in the PCM adaptors.
In the UK, Decca Records went their own way and instead of using a PCM adaptor, they built their own digital audio recorders, which omitted the video stage and recorded data directly on to the slant tracks of a rotary head deck. The Decca recorder contained all of the elements of the later DAT format except miniaturization.
Standardized PCM Adaptors
The format of a PCM adaptors intended for use with Betamax and VHS cassettes was standardized by EIAJ in 1978 and probably the most well-known unit was the Sony PCM-F1 which allowed digital stereo recording of extremely high quality for an extraordinarily low purchase and running cost. These machines became very popular and served to demonstrate the quality advantages of digital audio over earlier media.
Open reel tape was also used for digital audio. Tom Stockham built a digital audio recorder using an instrumentation recorder transport. The BBC built an experimental 8-channel recorder. 3M sold a 32-track recorder running at 45 inches per second. Editing required selective data transfer and two machines were needed. Swedish band Abba used 3M digital recorders starting in 1981.
Probably the most successful open reel format was DASH (digital audio stationary head) developed by Toshi Doi at Sony. The most popular versions were 24 tracks using ferrite heads and 48 tracks using thin film heads, both on ½-inch tape. TEAC and Studer also supplied DASH machines. Remarkably, the DASH format supported razor blade tape editing, using a huge interleave to accommodate the data damage at the cut.
Mitsubishi had a competing 32 track format called ProDigi which was also supported by Telefunken and Otari. Sony got into a bit of trouble when they published an advertisement showing a DASH machine being delivered in the back of a Mitsubishi pick-up truck.
Although famous for traditional tape recorders, Ampex never made a digital audio recorder, but the Ampex tape division supplied digital tape that was preferred by many DASH format users.
Open reel digital audio was very much a temporary phenomenon as the slow access time caused by winding tape reels limited productivity and other media provided faster access.
Digital recording and mastering recorders certainly improved the quality of vinyl disks, but it was inevitable that a consumer digital audio disk would emerge. Philips had an optical disk technology suitable for mass replication by pressing that was used in their Laservision video disks. Using a modulation scheme devised by Kees Immink and called EFM it became possible to record up to 75 minutes of stereo audio data on a 12cm optical disk. Sony provided the error correction system that allowed the disk to survive consumer handling standards.
Without error correction a digital medium would be useless as occasionally a high order bit might be corrupted which would produce a very loud noise, drowning the true sound. Error correction had been in use for some time, but it was not until it could be implemented at consumer prices that the Compact Disc would be viable.
16-bit as Standard
Originally the CD was going to be a 14-bit format, but late in the day it was decided to go to 16-bits. Philips had developed a good 14-bit DAC that was suddenly out of the running. They solved that problem using oversampling.
At the time of its launch, the recording density of the Compact Disc was far ahead of anything the computer industry had. Before long the technology of CD was adapted for data recording to produce the CDROM, a read-only disk that was ideal for software distribution.
CD was a conservative format, technically better than it needed to be. The Super Audio CD was hailed as an improved replacement. No one could hear the improvement and it failed. The MiniDisc was a recordable format using magneto-optics, but it employed a crude compression system that impaired the sound quality and that too failed, as did the Digital Compact Cassette (DCC).
The rotary head DAT format didn’t fail, but it fell foul of record company lawyers who saw the ability of DAT to copy CDs as a threat to their profits. DAT was prevented from becoming a consumer product and the disgruntled developers tried to re-purpose the format for professional purposes by adding proper time code. DAT was also turned into a computer data storage format.
The equipment manufacturers had their revenge. They simply brought forward the recordable CD. As it used the same format as the regular CD there was nothing the record companies could do. Before long there was a recordable CD drive in every laptop computer.
Computer Audio Restoration
Adhering to Moore’s Law, computers got cheaper and faster and so did disk drives, which became attractive for audio editing because of their fast access. About the first use of such a system for audio production was in 1967 when Tom Stockham used a PDP-11 to restore some historic recordings by legendary tenor Enrico Caruso.
Denon built an in-house hard disk audio recorder in 1982. The first commercial disk-based recorders products were the AMS AudioFile and the DAR Soundstation. Since then disk recording has become the norm for audio and tape has practically died out. Most radio stations use automated playout based on file servers.
Many small audio mixers contain convertors that allow them to be plugged in to the USB port of a laptop. Suitable software, some of it free, allows the laptop to become an audio workstation. Every musician who wants one can afford a personal recording studio and the big studios lost business.
Developments in flash memory led to audio recorders with no moving parts that obsoleted the Compact Cassette as a portable sound source. Many of these used audio compression to extend playing time and the sound quality was marginal. Musician Neil Young was rightly disappointed with the way his recordings sounded through these things and founded Pono, which made uncompressed personal players. Unfortunately, they were as over-specified as the compressed players were under-specified and the project ultimately failed.
The march of information technology led to the Internet and it soon became clear that downloading an audio file was a lot easier and cheaper than obtaining a physical recording medium.
Today digital audio is just another form of data that can be stored on practically any data medium and just needs to be reproduced at the correct rate. Audio specific formats are no longer required and there won’t be any new ones.
Why Did You Read This?
You might also like...
Digital audio relies completely on the accuracy of quantization and it is important to see how it works.
The consumer video market is growing fiercely competitive between content creators and content aggregators and OTT live and OTT VOD formats are growing increasingly fragmented. Consumers are benefiting from the tremendous choice of content and sources for that content while…
Once upon a time, the cause of data corruption would be accidental. A dropout on a tape or interference picked up on a cable would damage a few bits. Error correction was designed to deal with that.
Digital audio relies completely on sampling and no treatment of the subject can be complete without looking at how it works.
The requirements for data transmission have changed out of all recognition since the early days of computing where the goal was simply to make something that worked. Today that’s the easy part.