High-Resolution Audio Origins
When did high-resolution audio begin? There’s a great deal of confusion regarding the start of high-resolution audio recording production and distribution. As I sat next to Ryan Ulyate on the recent CES HRA panel on the people creating high-resolution audio, I was surprised at his answer to the first question. When the panel was asked about when and where we got involved with high-resolution audio, Ryan answered, “I’ve been making high-resolution audio records since 1979.” Knowing that a 2″ analog tape machine has limited dynamic range (60 dB without noise reduction) and that subsequent mixing and mastering transfers only diminish that further…sometimes much further…I think it’s wishful thinking to believe that everything recorded during the analog tape era (which persists for many) qualifies as high-resolution. I certainly agree that some absolutely amazing recordings were produced on analog equipment, but there are very real technology constraints (and a magic “sound”) that limit these machines as compared to modern high-resolution PCM digital hardware.
And then the guys on the retail panel starting talking about how they’ve been in the high-resolution audio business since the 70s. They assume that selling the best, high-end equipment of the time qualifies as high-resolution. It would also mean that we’ve made no progress in the production and delivery of great sounding recordings. Of course, it’s debatable whether the technology has been used to produce better sounding records, but we do have the ability to make recording with higher fidelity than we had previously. As great as you think your vinyl LPs sound, they fall far short of “sound that meets or exceeds the capabilities of human hearing…my definition for HRA”. In the end, we may have to face the facts that high-resolution audio doesn’t matter. I certainly think that’s going to be the case with the masses. The can’t tell the difference between lossy compressed files and aren’t going to spent tons of money to move the needle just a little bit.
The first time I heard the term high-resolution audio or my preferred high-definition audio, was in 1995 when progress on a new high capacity optical format was brewing inside of both Sony/Philips and WB/Toshiba. Their competing formats, MultiMedia CD and SuperDisc respectively, finally merged as the DVD-Video format in the fall of 1995. The new movie disc would be able to handle lossy surround sound encoded in Dolby Digital but was also capable of 2-channel stereo PCM up to 192 kHz/24-bits! Whoa! For the first time, engineers and labels were able to release recordings that had greater fidelity than a CD. In fact, Michael Hobson at Classic Records made a bunch of analog to high-res PCM stereo transfers and sold them as DAD (“Digital Audio Discs”).
Nobody was making new recordings at 192 or even 96 kHz and converters had just started extending past 16-bits in those days but the potential was finally there to record and release “high-resolution audio”. I would accept that date as the starting point for HRA. Anyone citing an earlier date is dreaming…and even 1995 is a stretch. The actual date for the arrival of high-resolution production systems and consumer hardware didn’t happen until 1999. That’s when Sony and Phillips rolled out their first SACD player and about 10 discs, all of which were transfers of older analog tapes.
DVD-Audio was introduced in the middle of 2000 with more transfers but most with 5.1 surround mixes.
So we’ve been talking about high-resolution audio for about 15 years. There will be those claiming, “of course, analog recordings are high-resolution…after all they have infinite resolution, right?” Sorry wrong.
The history of high-resolution follows a long and tortured path…one that is still struggling for its day in the sun. With competing definitions, confusing logos, and research studies that prove the opposite of what we want…it’s going to be a long time before it’s done right. If it happens at all.
My take on it is: there is no “one date” when HRA was born:
On one hand, there is the potential for HRA and there is the actual HRA music being marketed. These are two different things.
On the other hand: it is different for every stages of commercial music:
1. The recording stage: HRA probably started with the advent of digital recording at rates better than 44 kHz /16-bits;
2. The commercial support: so before DVD-A, it was not possible to sell HRA music;
3. The domestic playback systems: when all the components (source-preamp-amp-speakers and other signal processors) are able to reproduce more than 22KHz and 90 db. of dynamic range.
But again, as long as no one agrees on what HRA is, anybody can be both right and wrong.
When I was in high school in the mid-nineteen-sixties, I fantasized recording all the radio and tv signals available at a particular time by feeding an antenna’s amplified output into a reel-to-reel tape deck. On playback, the output would be sent to a small transmitter so a radio or tv could tune in any desired broadcast.
Obviously, this wouldn’t work – not because I had deliberately excluded any signals, but because the tape deck simply wasn’t capable of recording radio frequencies.
I was reminded of this long-ago example of “magical thinking” in my youth when I read the suggestion above that some might argue that analog recordings are of infinite resolution – after all, they don’t deliberately filter out high frequencies the way those awful digital guys do!
There seems to be a willful refusal on the part of the romantic thinkers (to use the typology used in Zen and The Art of Motorcycle Maintenance) to accept that a technology they’re comfortable with has limits to what it can do.
Comments like “…analogue recordings are high-resolution…” must come from those who know that Count Basie said: “If it sounds good. It is good.”
Mark, at what bit depth did the early studio digital tape machines record at? I was under the impression they started showing up in studios in the late 1970’s (1979) and the early 1980’s. Ry Cooder and Steve Wonder both made albums in that 1979 that were recorded digital on 3M machines.
Sony and Mitsubishi also made some of the first machines. Frank Zappa was a early adopter of digital recording with a Sony deck. But I don’t know if they were even 16bit/ 44.1 at that early stage.
Do you have any insight on these early machines?
There were very early digital audio recorders in the 1970s. IN fact, I met one of the original engineers at the recent CES show. The first digital audio system was designed and built by Soundstream, which was based in Salt Lake City and founded in 1975. The prototype machine ran at 37.5 kHz and had word lengths of 16-bits. Later machines moved the sampling rate to 42.5 kHz.
The 3M machines ran at 50 kHz and 16-bits.
For the purposes of your own website, you are free to adopt any definition of HRA you wish – and it’s good your consistent –, but I haven’t seen the same definition adopted by anyone else.
Referring to the JAS requirements, I’m sure there were electronic components in the 70s that could pass a 40 kHz signal, that’s not really very high. I’m not sure about loudspeakers though. I guess it might depend on in what sort of +/- response window you have to reach 40 kHz.
I’ve been advocating within the AES, CEA, and NARAS to establish a meaningful definition of high-resolution. The one I have proposed makes sense. It’s not full of numbers and formats. The JAS definition is the closest that I’ve seen to matching my own definition. If the existing definition of “better than CD quality” remains with the arbitrary source “categories”, then high-resolution means nothing.
A.S. What could you say about Musepack ~225 kbps ?
Analog tape consists of finite number of particles, thus it is finite quality. However, it still features much better impulse response than 384 kHz right out of the box, but nothing more.
So, what to do?
First, sound is being captured by a 768 kHz / {?} decits non-anti-alias-filtering Oversampling ADC.
Since I just have found the way to substantially improve upon the performance level of current CPUs, the sound is then being further oversampled highly to make its impulse response far better than that of even professional analog tape. {Note: in a 44.1/16 recording, impulse response is not clipped, it is only time-smeared, so it can hypothetically be recovered to its initial state; it is responsible for revealing some ‘aura’ around instruments called ‘airiness’, aside from natural 3D image.}
The result is very obvious: live sound .
Mark,
I would like to clarify some of the chronology you quoted in your article.
The first machine that could record and play back at 96kHz, albeit only with a 16-bit wordlength, was the Pioneer D-07 DAT machine, which was released in 1992. The internal converters were not very good and the standard machine had no digital interface to allow the use of external converters.
In either 1992 or 1993, ADT in London modified the D-07 and produced an interface that enabled the D-07 to be connected to dCS 902 A/D converter and allowed 24-bit 96kHz recordings. dCS’ background was A/D and D/A converter design for the aerospace and military industry – either in IC or discrete form – and the converter design in the 902 was entirely their own. It was a discrete implementation using a topology developed from a project for an airborne radar. It used what they called their Ring DAC, which had exceptional linearity. Measurements made by dCS on a dCS 900 A/D converter, which used the same Ring DAC, showed better than 24-bit linearity. These were made using a measurement system they had developed for in house use, as the nature of their work meant they were developing converters that had levels of performance that exceeded the capabilities of the then commercially available measurement systems. This measurement was detailed in a technical paper. Unfortunately I no longer have a copy.
NHK, the national television company in Japan, used the Pioneer/dCS system to record a 24-bit 96kHz sound track for a demonstration high definition TV laser disc they produced.
In 1995 dCS released the dCS 952 24-bit 96 kHz D/A converter and used together with the dCS 902 A/D it became possible to record and play back 24-bit 96kHz recordings using the four tracks of the Nagra D digital tape recorder in 24-bit 48kHz mode.
In 1996 the Dutch recording engineer Bert van der Wolf working for the recording company Kompas, started making recordings at 24-bit 96kHz using dCS converters and initially the Nagra-D and later a work station made by the Dutch company Augan, which allowed editing. These recordings were down sampled for release in the CD format using the dCS 972 sample rate/format converter. Many of these appeared on the Turtle Records label.
Also in 1996 MTX Digital Mastering of Switzerland made 4-channel 24-bit 96kHz recordings at the Montreux Jazz Festival using dCS converters and a Pyramix work station.
The dCS Elgar consumer D/A converter was launched at the Stereophile show at the Waldorf Hotel New York June 1996. This could play back 24-bit 96kHz recordings from the Nagra-D.
In 1997 the dCS 904 24-bit 192kHz A/D and dCS 954 24-bit 192kHz D/A converters were released. In June of that year Bert van der Wolf made the first 24-bit 192kHz recording. This was of a jazz group called Toys and was made in church in Deventer, Holland. The recorder was a Genex GX-8000 machine that provided 8 tracks of 24-bit 48kHz capacity: each 24-bit 192kHz channel being spread over four 24-bit 48kHz tracks. The recording medium was a magneto-optical disc. The following day he recorded the Orpheus String Quartet in a monastry chapel (I forget the name) in Tubingen, Germany. The day after that he recorded numerous classic car engines at the Mahy Motor Museum in Belgium and then the following month the Gonville and Cauis choir in the Queen’s College chapel, Cambridge. This was all part of a project commissioned by Samsung in Korea, who wanted 24-bit 192kHz recordings (and converters) so that they could lobby the Working Group 4 to have this format included on the DVD-A specification. Pioneer in Japan also made 24-bit 192kHz recordings with dCS converters at this time for the same purpose.
Bert van der Wolf’s sessions were also recorded at 24-bit 96kHz on to a Nagra D and these were edited with other 24-bit 96kHz recordings on a Sadie work station, down-sampled and released as a dCS demonstration CD on the Kompas label.
That year Bert started to record a lot of his commercial projects at 24-bit 192kHz, using the Pyramix work station – he may have used an Augan system at first. These recordings were converted to 16-bit 44.1kHz with a dCS 974 sample rate/format converter.
In 1998 the dCS Elgar was upgraded to accept 24-bit 192kHz data.
So, true 24-bit audio converters and high-resolution recording and playback hardware (professional and consumer) has been around for quite a while.
You mentioned that the first SACDs released were all copies of analogue tapes. I spoke with one of the senior engineers at Sony who worked on the SACD project and he told me that there were some native DSD recordings in the first batch of releases. These included, “All TELARC recordings (except for reissue titles), all DMP recordings, San Francisco Symphony’s Mahler cycles, Yo-Yo Ma recordings, and following titles are native DSD made by using Sony’s Blackline recorder or Philips’ multi-channel recorder.”
Robert Kelly
Thanks Robert. Most SACDs (85%) of them are still transfers of older analog tapes or conversions from PCM. The arrival of high-resolution distribution happened with the launch of the DVD-Video format in 1996-97.
Hello Mark,
Thanks for posting my comment.
All the best,
Robert
Digital Recording Time (copied from http://en.wikipedia.org/wiki/Digital_recording)
In 1938, British scientist Alec Reeves files the first patent describing Pulse-code modulation (PCM).[1] It was first developed as a telephony technology.[2]
In 1943, Bell Telephone Laboratories develops the first PCM-based digital scrambled speech transmission system, SIGSALY,[3] in response to German interception of military telephone traffic during WWII. The twelve transmission points were retired after the war.
In 1957, Max Mathews of Bell develops the process to digitally record sound via computer.
In 1967, the first digital audio magnetic tape recorder is invented[clarification needed] by NHK’s research facilities in Japan. A 12-bit 30 kHz stereo device using a compander (similar to DBX Noise Reduction) to extend the dynamic range.
In 1970, James Russell patents the first digital-to-optical recording and playback system, which would later lead to the Compact Disc.[4]
In 1972, Denon invents the first 8-track reel to reel digital recorder.
In 1975, Thomas Stockham makes the first digital audio recordings using standard computer equipment and develops a digital audio recorder of his own design, the first of its kind to be offered commercially (through Stockham’s Soundstream company).
In 1977, Denon’s music company Denon Records, a division of Nippon Columbia, recorded the first all-digitally-recorded commercial album using their Denon 034 multi-track system. The album was Archie Shepp’s On Green Dolphin Street, released initially on LP record and eventually on Compact Disc.[5]
In mid-1978 Sound 80 Records digitally records “Flim and the BB’s” (S80-DLR-102) as a backup to the direct to disc recording. This digital session is released as an LP record, but because the recorder was a prototype 3M digital recorder, the master tape becomes unplayable when the prototype is disassembled in 1979. No Compact Disc reissue is possible.
In 1979, the first digital Compact Disc prototype was created as a compromise between sound quality and size of the medium.
In 1979, the first digitally recorded album of popular music now with vocals, Bop ‘Til You Drop by guitarist Ry Cooder, was released by Warner Bros. Records. The album was recorded in Los Angeles on a 32-track digital machine built by the 3M corporation. Also, Stevie Wonder digitally recorded his soundtrack album, Journey Through the Secret Life of Plants, three months after Cooder’s album was released, followed by the Grammy-award self-titled debut album of American singer Christopher Cross which was also 3M digitally recorded album.
In 1982, the first digital compact discs are marketed by Sony and Philips,[6] and New England Digital offers the hard disk recorder (Sample-to-Disk) option on the Synclavier, the first commercial hard disk (HDD) recording system.[7] Also that same year, Peter Gabriel releases, Security and The Nightfly released by Donald Fagen, which both were the early full digital recordings.
In 1984 Sony released the Sony PCM-501ES digital audio processor, which for the first time allowed consumers to make their own digital recordings, using a VHS or Betamax video tape recorder as the storage media.
In 1990, digital radio begins in Canada, using the L-Band.[8]
In 1991 Alesis Digital Audio Tape or ADAT is a tape format used for simultaneously recording eight tracks of digital audio at once, onto Super VHS magnetic tape – a format similar to that used by consumer VCRs. The product was announced in January 1991 at the NAMM convention in Anaheim, California. The first ADAT recorders shipped over a year later in February or March 1992.[9]
In 1993 RADAR (audio recorder) Random Access Digital Audio Recorder or RADAR is the first single box device used for simultaneously recording 24 tracks of digital audio at once, onto hard disk drives. The product, manufactured by Creation Technologies (iZ Technology Corporation) was announced in October 1993 at the AES convention in San Francisco, California. The first RADAR recorders shipped in August 1994.
In 1996, optical discs and DVD players begin selling in Japan.