Although we were using solid-state devices, including the very latest versions of Field Effect Transistor amplifier stages, it was entirely analog. Even decoding the tones in the command signals was done by phase-locked loops contained on chips developed for detecting Ma Bell dial signals from the new, at that time, push button phones.
All my training up to that point had been in the analog realm. That’s the area where small changes in voltage and current represent the signal. It was the time of meters with little red needles and dials. Soon I was enrolled in college studying the mathematics and physics behind the small electronic devices. There I encountered the digital domain. The realm of numbers … whole numbers.
By assigning one of two values to a voltage or current, the small variations no longer mattered. These discrete values were assigned a mathematical “one” or “zero,” no in-between. Based on voltages that swung high or low, these were decoded in a very certain way. No small doubt. Soon meters were replaced by digital readouts. Like the difference between the automobile speedometer and the odometer, no interpretation was required. No tiny marks between 50 and 60 to recognize if you were over the speed limit of 55.
Later, when I began my career teaching electronics at the college level, I chose digital courses to become my area of expertise. So I’ve lived in both worlds. A living arrangement I had already experienced as I made the transition from Navy basic training in vacuum tubes to the more common in the late sixties solid state and transistor circuits. So, again, I made a transition from one technology to another.
While the music reproduction industry had seen steady improvements and new technology from the 78 rpm records of the forties to the 33 1/3 rpm Long Playing album of the sixties, from mono to stereo and even beyond with the advent of Quadraphonic, all these technical advances were basically analog. In fact, if you looked at a groove in a record with a microscope, you would see a reduced image of the actual sound waveform as clearly as if you used an oscilloscope to view the voltage waveform. Whether recorded in magnetic transitions on recording tape or undulations on a vinyl groove, this was a real-world, analog picture of the music.
It suffered from the same problems as all analog: noise and interference, scratches and pops, dust and wear. The frequency response was twisted to fit the medium using ingenious RIAA equalization standards to improve the quality and deal with the imperfections.
But music can also be digitized. That’s really what a musical score is — sheet music. Music notation turns the sound into discrete notes that are then reproduced via the skill of the musician … a human being. Old fashioned player pianos with their paper rolls filled with holes are discrete — a word akin to “digital.” With a player piano, no human hand is need to interpret the music.
Naturally, an electronic version of this discrete player was due. The development of lasers and so-called “optical disks” (or discs) gave the medium. Now all that was needed was the method.
Sony and Philips set up a joint task force of engineers to design a new digital audio disc. After a year of experimentation and discussion, the team produced the “Red Book CD-DA” standard. First published in 1980, the specification was formally adopted by the IEC as an international standard in 1987, with various amendments becoming part of the standard in 1996.
Philips coined the term "compact disc" in line with another audio product, the compact cassette, and contributed the general manufacturing process, based on video LaserDisc technology. Philips also contributed eight-to-fourteen modulation (EFM), which offers a certain resilience to defects such as scratches and fingerprints, while Sony contributed the error-correction method, CIRC.
The Compact Disc story, told by a former member of the task force, gives background information on the many technical decisions made, including the choice of the sampling frequency, playing time, and disc diameter.
The logical format of an audio CD is described in that document produced in 1980 by the format's joint creators. The document is known colloquially as the “Red Book” after the color of its cover. The format is a two-channel 16-bit PCM (Pulse Code Modulation) encoding at a 44.1 kHz sampling rate per channel. Four-channel sound was to be an allowable option within the Red Book format, but has never been implemented. Monaural audio has no existing standard on a Red Book CD; mono-source material thus is usually presented as two identical channels in a standard Red Book stereo track (mirrored mono).
Interpreting the engineering specifications, the 44.1 kHz (thousand Hz or cycles-per-second) yields a maximum bandwidth from near zero to 22 kHz. Actually, that isn’t a bad bandwidth and exceeds the bandwidth of all but the highest quality turntables for LPs.
The Nyquist (or Nyquist-Shannon) theorem gives us the expected bandwidth for a given sampling frequency. However there are concerns with aliasing, which in CDs is eliminated by filtering the top frequency on playback. That makes this 22 kHz a hard stop for maximum frequency response. There are also concerns with transient response and phase shifting, but these same concerns apply to LP reproduction since RIAA equalization uses capacitance and inductance which can cause phase shifts too.
My personal opinion, which doesn’t really count for much, is that the frequency response of CDs is adequate and not really the area of concern with regards to high fidelity and faithful reproduction of the music spectrum. Limitations of the human ear … especially such an aged ear as mine … limit enjoyment of high frequency signals more than the CD specifications. Maybe dogs can hear those highs, but most humans can’t
Of much greater concern is the 16-bit encoding. This limits the discrete volume or loudness levels to a range of 216 or 65,536 values, while analog can, at least in theory, record an infinite number of values from zero to maximum. This limitation determine the dynamic range or the level of the softest and loudest sound. In addition, this limited number of values to represent the instantaneous voltage values produces a crudeness in representation akin to rounding errors. It just isn't a very precise representation due to the small encoding word size. Volume range is purposely reduced to preserve resolution.
To maintain a smooth transition from soft to loud, CDs typically raise the soft passages and limit the loud passages. This, and other considerations including musical style have led to amplitude compression of the music dynamic range. This occurs with analog recording too, but it has become almost a sonic trademark of modern music. Just where the art ends and the technology begins is becoming quite confused. (One reason that commercials seem louder than the program sound it a heavy level of amplitude compression. That makes the overall impression of volume seem loud.)
Analog also limits the softest portions due to surface noise. And the dynamic range of human hearing is limited too, and greatly affected by the absolute volume or sound pressure. Still, this is a major deficiency of the CD standard when compared to other recording techniques.
In response, many new and improved digital recording methods, increase the coding length to at least 20 if not 24 bits to vastly improve dynamic range. Rock and roll is not known for dynamic range. It is usually loud and louder. But symphonic music is more likely to have a wide range from the soft to the loud passages, and is quite limited by the basic CD bit length.
It was these technical choices by Sony and Phillips, made based on the capability of the technology available in the late 70s and various engineering concerns such as cost to create discs and cost of players to reproduce and marketing considerations in the 80s that determined the technical choices.
The frequency band and dynamic range of the CD is set in stone. It can not be changed without making all the current players obsolete. A later standard that allows for higher resolution was to come and is mostly associated with video and movie soundtracks. These DVDs could record music digitally with higher resolution, but the vast majority of modern music on optical discs is stuck with 70s CD technology that is now 40 years old. It is the digital equivalent of the wax cylinders Edison first used to invent recording and the home music player in the 1800s.
There are many advantages to digital recording including the near elimination of noise, interference, and wear as well as the ease of reproducing and copying without changing the fidelity or adding noise. But digital systems are rigid in their design. It is like a book written in French. It can only be read by a person with a knowledge of French, and won’t make any sense to a Chinese reader, even if Chinese were proven to be a more accurate technology for writiing thoughts and ideas.
Digital is always stuck in the algorithms used to produce the standard. In this case it’s the “Red Book" CD-DA standard. It has been amended and improved over the years, but the original choice of sampling frequency (which determines maximum bandwidth and high frequency reproduction) and bit length (which determines dynamic range … the difference between the softest and the loudest sound) can't be changed. That is literally frozen in stone.
DVD sound standards include word sizes of 20 and 24 bits and also increases sample rate to as high as 192 kHz. However, DVD music discs have not found a wide market except for movies and video. LPs, on the other hand, are enjoying a renaissance of demand and supply.
Something else new did develop in the digital marketplace over the last forty years. That something is digital music downloads: iTunes for example, or Pandora, Napster, and Rhapsody. There is more flexibility in this technology because it depends on computer software and these programs can be updated easily and cheaply, even on old hardware.
However, due to the limitations and cost of bandwidth, this technology was rooted in a processed called compression. That’s a mathematical method to squeeze the bandwidth out of the music to make it easier to transmit quickly over narrow bandwidth connections like the early internet. Many, such as iTunes, allow the user to set the level of compression when the user is recording their music and multiple levels of compression may be available from the source allowing the listener to decide the trade-off between quality and quantity. However, the very essence of compression is to reduce something; and that something is fidelity or sound quality. They just hope you won't notice. In fact, they are carefully designed to limit your noticing, and actually do a pretty good job of that.
These compressions were typically “lossy” which means they lose fidelity. Through ingenious design they mostly threw away aspects of the music that you wouldn’t notice anyway, but still they purposely lost fidelity to save bandwidth and storage space. The music is converted and recorded. It is then reconverted on playback restoring some of the music quality. Programs that perform these transformations are called CODECs("coder-decoder") Lossless compression doesn't eliminate any of the musical essence, but lossy compression methods abound since they are better are reducing total file size.
Since much of this listening audience use simple ear buds and listen in noisy locations like vehicles and the street, maybe no one noticed the loss of quality. But some did and some still do.
It is in pursuit of perfection for quality listening environments such as home theaters and quality hi-fi systems that the analog record never lost favor. In fact, it is enjoying a great resurgence. Can those music aficionados really tell a difference in the music quality? Well, the answer can be yes, although often it is more imagined than actually experienced.
In any case, more people are buying LPs than true high-resolution 24 bit/192 kHz files, the ones that can sound better than CD-quality FLAC (Free Lossless Audio Codec) or Apple Lossless files.
Well-recorded old and new LPs, played on a decent turntable, are capable of delivering high-resolution sound. Of course LPs are more likely to be played on a turntable at home, where the listener might actually be able to hear the difference between high- and standard-resolution recordings.
Listening to high-resolution files in noisy environments like a train, car, bus, or plane is like eating gourmet food on a roller coaster. Savoring sound quality is next to impossible in those places, at home with a decent audio system you're more likely to hear the difference high resolution can make.
A record manufacturing company is located a few blocks from Grand Central Station in Manhattan, an example of the modern, high end record producers. I visited them back when I was working in New York, and was given an informative tour.
Their record cutting machine, the Neumann VMS 80,shown in the photo at the start of this article, was described by their engineer as the "Ferrari of lathes." It was built in 1982, but has been rebuilt twice and has many tweaks. This machine cuts the groves in the lacquer master used to press the vinyl records.
I learned that, when inspecting freshly cut grooves with the microscope, one does not to speak while hovering directly over the master, you don't want to contaminate its pristine surfaces. On a bad day the grooves might take on a "streaky, rocky and jagged" appearance. When that happens they’ll have to do it again and cut a fresh lacquer disc. Working with physical tools like the lathe, makes a hands-on aspects of record cutting incredibly satisfying. They just bought an old analog Nagra tape recorder when I was there, and I noted the difference in workflow when not using computers. They weren't interested in the latest technological gadgets, often using older equipment from the golden age of analog audio.
LP mastering engineers are people who literally transform sound into physical wave forms; these engineers are sculptors of sound. To be good at their job they have to be attuned to the machine and materials they're working with. The temperature, humidity, the quality of the blank lacquer discs, and the cutting stylus all potentially affect the sound of the finished product, the LP. So working in the analog "domain" requires extreme attention to detail, as the smallest errors can spell disaster.
I am reminded of other skilled craftsmen who work with their hands from machinist to welders. Their craft is demonstrated in the quality of their work and the products they produce. Reminds me of the small town welder that repaired Robert Pirsig’s chain guard. Robert didn’t want him to try welding the thin metal for fear it would simply melt away, but the craftsmanship (and “quality”) of the tradesman yielded a perfect repair. That was the kind of quality Pirsig was talking about in ZAMM, and it is equally obvious in these workmen with lacquer and vinyl discs.
This level of care and attention to sonic details matches the original efforts of the musicians in their recording studios with multiple takes and late night mixing sessions. Good music doesn’t come from a computer, but from a beating heart and the warm blood flowing in the veins of creator and recorder both. It is no wonder that sonic critics refer to the music from LPs as “warmer” than the cold medium of digital.
This doesn’t mean that digital music work isn’t challenging. Conquering digital formats and the challenges of digital music are different, but still demanding. But even with the best engineering, the vinyl's sound always destroys the peak-limited, 16-bit CD version, it kills it dead. There are very few times when the CD is better than the LP. Dynamically uncompressed music on digital formats like CD and downloads is possible, but just not likely in the current marketplace. I am not an anti-digital guy, but you gotta really love vinyl.
It isn't all doom and gloom for those that rate both music quality and convenience with portability, however. Both SONY and a new comer — to music devices, not music — Neal Young have new, high fidelity, mobile, digital devices announced. The price is high, but for those that care, I assume it is worth it. The high resolution and high quality music catalogs are still a bit sparse, but if these higher priced players become a success, that is sure to follow. So keep tuned.
Nowadays, most vinyl releases are cut from digital masters, so there may be no escape for the purest, but the care and effort that goes into high quality vinyl production can restore some of that human warmth to the music.
No comments:
Post a Comment