Audicom

Audicom stands for: Audio en Computadora (Spanish). Released in 1989, it was the world's first PC-based broadcast automation system to use audio data compression technology based on psychoacoustics.

The audio card was designed with standard CMOS logic ICs and used hardware to perform an audio data compression algorithm named ECAM, based on the critical bands masking principle, a property of the human ear. Today the same principle is used in all the lossy audio bit compression systems, MP3 included. The audio card was designed for the old ISA slots of the PC and works using direct memory access to the host memory. The driver and the client application were developed by Gustavo Pesci [http://www.gustavopesci.com Gustavo Pesci Bio] ] , while the hardware was designed by Ricardo Sidoti and Elio Demaria. [http://www.lanacion.com.ar/Archivo/Nota.asp?nota_id=187775 La historia del Audicom] ]

The radio automation system was launched in 1989 under the name Audicom. After 20 years, almost all the radio stations in the world are using some similar technology, now manufactured by a lot of companies around the world. [http://books.google.com/books?id=zD9WAAAAMAAJ&q=solidyne+audicom&dq=solidyne+audicom&pgis=1 Audicom introduced at LAB] ]

History of Audicom's software versions

Audicom (mid 1980's): first version, included a hardware compression audio card: ADX903

Audicom II (1989): most popular version for DOS systems, exported to more than 30 countries. Introduced the concept of Playlist for computer systems, and created the concept of Live Assist. When working in automatic mode, the music was selected from several directories with the /RANDOM command allowing the rotate styles along the day.

Audicom III & 4: not released, but developed internally. Audicom 4 was the first version intended to be a Windows version (for 16 bit Windows), but due to problems in the Win16 platform it was canceled in favor of a Win32 system (Audicom 5). The change (including staff change) caused a delay to the project of two years.

Audicom 5 (Introduced at CAPER 96, commercially available at 1997): first Windows system available. There was two versions (both Win32 applications): ECAM and multimedia. The multimedia version included the option of an 'audio co-processor card' (a compressor-expanser for greater dynamic range with conventional sound cards). New ECAM cards allowd to perform the mixing on the same card. Multimedia version used two cards for mixing and 3 cards for mix and cue (which was pretty hard to install on a same PC). The Audicom 5 introduced a novel concept of Artificial intelligence for selecting music called AutoDJ, which allowed for '1 second precision' to meet the spots schedule on real time (the operator can add or remove titles and the autoDJ will meet the schedule without cut any music). Other novelty was the virtual locutor, which announces weather condition and time (based on a text to speech algorithm). A weather station was available as optional (SMT-100).

Audicom 6 (October 2000): Only a multimedia version. It introduced a new play engine based on DirectX, allowing mixing on the same audio card. First broadcasting audio system to support MP3, allowing cross-fade within two different MP3 files and CUE concurrently (all software decoding). The product was partnered with the PCI audio card SX46.

Audicom 7 (August 2002): Included multiuser support, network ready, included autoEdit (automatic editing of MP3's files without recompressing), full MP3 mixing. Also was partnered with an advanced PCI audio card: DSP48.

How the ECAM technology works

When we start developing the perceptual coding technology for Audicom (the ECAM system), the PCM coding was well known for its telephonic applications. But our decision was to use, instead of PCM linear coding, a more complex coding named Sigma Delta Adaptive Modulation. Since we needed to develop a real system that can be manufactured with the late 1980 decade technology, after considerable research, we decided to use a Sigma Delta codec designed for speech coding in telephone applications. At that moment audio quality converters were not manufactured because it did not have commercial application... We change the clock rate and all external components to work in wide bandwidth: 20 - 20.000 Hz. We tested it at 96 Kbits/second and of course it does not sound well (this was not a surprise because was predicted by the Sigma Delta theory).

At this stage we start to use the band masking concept in a way different that the proposed for speech coding, because we were looking for audio quality at the full audio band. We use a crossover-like bank of filters with four bands that follows the Ehmer curves ["Masking by Tones Vs Noise Bands",Richard Ehmer, Journal of Acoustical Society of America, September 1959 ] We compress each band in a 2:1 ratio when record the sound. (2:1 means that each 20 dB the signal changes, the output only changes 10 dB). When reproduce, we did a complementary expansion.

To improve the audio quality we use a new Psychoacoustic concept developed by Oscar Bonello; the Burst Masking [ Oscar Bonello, Burst Masking, “Enmascaramiento por Rafaga”, EAA Symposium of Architectural Acoustics and Iberian Congress on Acoustics, Madrid, Spain, October 2000, ISBN 84-87985-03-3 ] The burst masking theory deals with the concept of the “integration time” of the human ear. If a a subject listen a burst of N milliseconds of the masking tone plus a second tone or noise, the masking threshold is dependent of the value N. When N is very high we get the Ehmer curves. For low values the masking increases. For example, if a masking curve (static) is applied to a pure tone of 1 kHz at 90dB SPL, it can be found that is possible to recognize a tone of 5 kHz, 65 dB below the tone of 1 kHz. But the masking theory shows that within a burst of 50 ms this value is reduced to 53 dB and for 5 ms to only 36 dB. . [” Multiband audio processing and its influence on the coverage area of FM stereo transmission” Fig 7, Oscar Bonello, AES Journal", New York, March 2007 ]

Modern Systems for Radio Automation

In 1983 Oscar Bonello a radio automation pioneer, and Engineering professor at the University of Buenos Aires and Fellow Member of the AES (USA) start researching about using the recently developed IBM PC computer. The problems that he faced were: 1) Create an audio PC card of good audio quality; 2) Create a bit compression algorithm; 3) Create the Automation software to be run from the PC. [http://www.aes.org/e-lib/browse.cfm?elib=2384 New Improvements in Audio Signal Processing for AM Broadcasting by Bonello, Oscar] ] The audio card was designed with standard CMOS logic ICs and perform by hardware the ECAM bit compression algorithm based on the principle of Critical Bands Masking a property of the ear (see a brief History at MP3 ). Today the same principle as the one used by Bonello is used in all the lossy audio bit compression systems, included MP3. The audio card was designed for the old ISA slots of the PC and works using direct access to PC memory. The driver for this software was developed by Gustavo Pesci, a young software engineer, formerly a pupil of Bonello at University. The card hardware was designed for Ricardo Sidoti and Elio Demaria, both brilliant electronic engineers.

The Radio Automation system was launched in 1987 under the name Audicom The software initially ran on DOS and was further changed by Sebastian Ledesma to Windows, using an Artificial Intelligence concept. Twenty years after the original Bonello invention, almost all the radio stations at the world use the same technology, now manufactured by various companies around the world. Mr Bonello refuses to apply for any type of invention patent.

Modern systems typically run on hard disk, where all of the music, jingles, advertisements, voice tracks and other announcements are stored. These computer files may be either compressed or uncompressed, often with a minimal compression as a compromise.

Scheduling was an important advance of these systems -- allowing for exact timing. Some systems use GPS satellite receivers to obtain exact atomic time, for perfect synchronization with satellite-delivered programming. Automation systems are now more interactive than ever with digital mixing consoles and can even record from a telephone hybrid to play back an edited conversation with a telephone caller.

The simplest arrangements can even be done on software as basic as Winamp.

There is also a trend in radio to use voice-tracking production techniques which make the live presenter in the studio, less critical in the overall broadcast; a great number of Internet radio and variety hits stations use automation software and voice-tracking to completely replace live on-air talent

PC Cards

The Audicom cards are a family of PC audio cards developed by Solidyne. They were intended to the broadcasting market and some models included the Audicom broadcasting system.

In mid 80's Solidyne, a broadcast manufacturer company, started the project of recording audio in a PC to automate the spots emission (at that time done by cartridges). By that time, the hard disk were very expensive and extremely small (a PC hard drive allowed 10MBytes of storage), so the engineers knew that they needed to develop a system to reduce the required bits as much as possible in order to be successful. The result was the ECAM audio system, the first system was showed to the 'Secretaria de Ciencia y Tecnica de la Nacion de la Republica Argentina' in 1988. [http://www.aes.org/e-lib/browse.cfm?elib=6674 PC-Controlled Psychoacoustic Audio Processor by Bonello, Oscar Juan] ]

Models

#ADX 903
#ADX 922
#ADX 925
#ADX 925 Serie X
#SX 48
#SX 46
#DSP 48

References

External links

* [http://www.solidynepro.com Solidyne] Company's web site.