
| The Memory Player(tm)... "The most advanced CD playback system ever designed" The Memory Player (tm) plays CDs in an entirely different way than has been done before. This dialog is an explanation of how Memory Playback differs from mechanical CD playback and why the fidelity levels between the two are so vastly different. In order to explain and clarify the unorthodox technologies that were first designed for and first used by The Memory Player(tm), it becomes important to understand how mechanical CD players work and why they are limited to the performance level that is largely unchanged in the CD's 25 year history. This dialog will also aid the non-technical reader in understanding how recording music worked in the analog era, how the CD changed the basics of recording music and especially replaying music, and how in an unforeseen way, it was both miraculous and disastrous from the point of view of pure fidelity, a fidelity that is fairly limited to the lofty demands of the HEA community. A brief history of the CDs evolution through its varying incarnations stressing first bit depths, then sampling rates, then reclocking and now removing the bits, purifying them on to non-mechanical memory and playing them electronically from it, ie: Memory Playback. Continuity and Capsules Packets of Power, Modules of Music, Blocks of Bits, Thanks Mr. Planck Max Planck discovered that all energy/mass exists in packets or blocks. More on this later in the dialog but it does illustrate something important to ALL of recorded music: Sound is quantized. Where digital and analog barely differ is in being recorded. Music is recorded, transferred to a medium of some sort and played by the end user using a reverse technology. This has essentially not changed since Edison's wax. -------------------------------------------------------------------------------------------------> ALL recorded music is quantized. That is, a moment of music is represented by an abstract entity and encapsulated into a given format. This entity can be a magnetized area of a tape, a bump or a valley in a groove of a record or, a bit or a group of bits on a CD. ALL of these representations of moments of music in a period of time have enjoyed enormous success and until the optical CD, there has been relatively little difference between them philosophically. ALL were "formats", a time period structure, filled with music. A block of sound...much thanks, Max Planck. And while we generally have seen improvements with each advancing format and technology, we faced a duality of qualities when the CD dominated recorded music. Fidelity was, for the first time, not the primary motivation anymore but rather, durability and ubiquity became CD's primary virtues. In some very few respects, of course, the CD was superior to records but in the HEA community we found very few who embraced the CD due to its corruption of timbre, artificial tone, harsh high frequencies and loss of subtle ambient information. With slow, thin tape like cassettes and inexpensive ceramic cartridges we expected and understood the loss of fidelity. But the sound you heard was always less, not more, of something. Something went awry in the natural path of the evolution of music recording. Playback technologies were interrupted by a new and unheard of addition to the musical information. It was ADDITIONS, not losses, that changed the development of the CD into the artificial sounding medium we'd soon have forced upon us and one addition in particular: ERROR CORRECTION. Two "additions" in particular doomed CD sound to the artificial nature it suffered for 25 years: 1: CD error correction as touched on above, was birthed by a cooperative effort by Sony and Phillips in August, 1982. The goal being that with dirt and scratches, freezing Winters and burning Summers in your car, fingerprints, physical damage, almost anything and the CD would still play. No other audio medium had ever boasted such a stoic technology before. If I drove a tank I'd never have to worry about digging my car out of the snow but both the fuel economy and trying to find a parking space would be seriously compromised. 2: Phillips asked Sony to reduce the strength of their ECC in order to fit the entire Furtwangler's, Beethoven's 9th Symphony---------------------> on a single CD. In order to ADD the 9th Symphony, Sony had to SUBTRACT ECC resolution. Real Estate Realities "Good enough for Music" Every music CD made since 1982 has about 1/3rd of its area filled with error correction bits known as Reed-Solomon Codes and/or ECC----------> (Error Concealment/Correction Codes). To be more accurate, there are actually many other error correction and error concealment codes on a CD and as there is no umbrella term to encompass them all. So we will use the term "ECC" for the sake of brevity of this dialog. Proponents of Reed-Solomon (RS) coding should know that RS codes can be designed to vastly different strengths and properties. RS for DAT tape was vastly different than for CD. Even RS codes for CDs varied depending upon the content. It was discovered early on that the ECC that protected CDs with PROGRAMS on them required a far stronger Reed-Solomon code than the RS code protecting CDs with music on them or they would literally crash the computers. This stronger ECC occupied 50% of the CDs surface. The weaker ECC, only 33%. And music CDs were no different although the designers thought differently. RS coding for "just music" was relaxed while RS codes for CDs carrying programs were very powerful. It was decided that "errors shorter than 25 microseconds will be---> considered inaudible" and "perfected" so a "relaxed" version of the Reed-Solomon code used on the PROGRAM CDs would be used on MUSIC CDs. Look at how much weaker the music protection is! Phillips asked Sony to reduce the strength of their ECC in order to fit the entire Furtwangler's, Beethoven's 9th Symphony on a single CD. A record album barely held 45 minutes of recording whereas a CD could hold over 70 minutes IF the ECC was reduced in quality. One thousand times weaker. This decision proved to be historic. It was this pivotal decision that doomed CD sound to it's artificial nature, one that would create an insurmountable chasm between the vinyl advocates and the embracers of this new technology, for over two decades. The weaker ECC code was known as RS 255/223* and was considered CD Error Correction "See me, feel me, touch me.... and you can still hear me?" Reed-Solomon. Error correction. Error concealment, Interpolation, Parity codes, and more. These error correction codes are the CD standard, and no CD will play at all without them. Digititis or just "digital sound", has plagued CD playback since the first 14bit players were released in the 1980s. They were in a word, atrocious. Despite audiophiles resistance, digital audio seemed inevitable, and in the late 80s', audiophiles gobbled up all the quickly vanishing vinyl, as they didn't know if recorded music could ever sound like music again! The proponents of CD made claim to essentially one, and only one, area of improvement that if improved, could bridge that chasm that stood between the vinyl lovers and CD's musical Nirvana: JITTER While defining jitter has many complex avenues we could take, in an effort to define it succinctly: Jitter can be defined as the error in the timing of all portions of the digital audio chain NOT working in synchronicity. How this worked is as follows: As the CD spins, it sends a "block", a capsule of "bits" to a temporary memory called a "cache". This block waits until the CD in its less than accurate rotation has caught up and is ready to send the next block to this cache. After the CD has uploaded the next block to this cache, the content is played and the process repeated. The caches will have filled and emptied millions of times and the blocks of bits, the very content we seek to hear are fed to varying circuits of little import here with the exception of the device that turns bits into music, the DAC - or, " Digital to Audio Converter ". As you can see at this point, the timing of a cheap piece of plastic spun by a cheap little motor to circuits that require nanosecond accuracy to eventually be delivered to the DAC, exactly in time* with the entire chain that is previous to it, required a not so inexpensive device to "marry them all in a timely manner". To try to reduce this timing error, jitter as it was now called, the CLOCK was created to be the timekeeper of every part of the digital audio system. And, improving the clock became the implied panacea of all things wrong with CD. If the clocking was perfect: NO JITTER. NO PROBLEMS...right? Not as simple, unfortunately. Of course clocking is not perfect nor can it ever be but, it is EXTREMELY good now and as such, has revealed itself as a less important aspect of CD playback than once thought! It was no panacea. Near zero jitter still "sounded digital". Not to say that reducing jitter did not improve CD sound, for most certainly, it did. But that sound, that "digital sound". No reduction of jitter made any significant progress of ridding CD of its dry, airless, acrylic, artificial tone and timbre. Vinyl, with all its faults, noise, clicks, pops and mistracking, sounded like music. CD did not. Why does this nanotechnological wonder, the CD, sound artificial, hard, harsh, sterile and a-musical after 25 YEARS? Blinded by the Laser Light The perfection of the myth: "Bit-Perfection" No clear answer surfaced. We knew of the tolerances in Reed-Solomon codes but they were "inaudible". Lasers by definition, do not diffract. We knew the lasers of the quality in CD players suffered diffraction,----> and this too was "inaudible". Jitter was the only acceptable answer to all questions concerning a "perfect" medium. In part, by placing the onus of all things wrong with CD on jitter alone, it prevented the questioning of "Bit-Perfection" as this was unquestionable and always loomed as if it was sacrilegious to even suggest questioning it. We were, for still another decade, fixated on JITTER. The goal was to align the timing from the CD to the DAC and so then clock accuracy was still the primary imperative. But was it? Clocking and reclocking seemed to enter the realm of the absurd. Not unlike the "THD Wars" of the 1970s where ONLY lower THD mattered and for a time, almost no one listened anymore. Instead they read spec sheets. Jitter has too reached the realm of "mythology" with atomic clocks surfacing with incredible accuracy that gave us just another "baby step" in improvement (as Arnis Balgalvis described progress in CD playback once in Positive Feedback magazine in 2007 ), but got paid for with the baby's college fund! Today, clock accuracy and synchronization along the entire chain is so good and the improvements in CD players reclocking are so relatively small, that it seems that neither clocks nor jitter were the cause of "digital sound" as we know it, and we were forced to accept it for 2 1/2 decades. The Memory Player&Memory Playback "The first significant advancement in CD reading in 25 years" Arnis Balgalvis, Positive Feedback magazine, 2007 A kind of fatalist, circular logic of the acceptance of CD became ubiquitous: "It must be too low a sample rate","It must be too low a bit depth", "We're stuck with 16/44", etc. HOWEVER, a radical few, a few who came under tremendous ridicule in audio forums, a few proposed that a new CD playback concept had already become a reality, thanks to the advent of the power of modern computers. The PC did bring to the average audiophile or end user a complex processing power that only ten years prior would cost have cost more than a home. Because of this technology, a once highly controversial and abstract concept became a reality in the Winter of 2003... A CD player that plays ONLY the music bits, (that is, bits representing music, not repair bits added later) and NONE of the error codes, is a concept that is not strange to 32-bit mixers in most recording studios. Yet the mantra echoed in audio forums seemed to literally shout down the dissenters, the thinkers, the creators: "CD is bitperfect and no bits are ever lost because of Reed-Solomon and ECC and if you are hearing any advantages in improving CD reading, they are imaginary". In order to see (and hear) how the 2 1/2 decades of burying one's head in the sand kept CD from stealing the hearts of true of music lovers and audiophiles, we need to explain what Reed-Solomon codes can and cannot do, and what actually IS that sound known as "Digititis". But not all in the audio forums sought to destroy that which they did not understand. In fact, one maverick, a lofty-minded writer, also in Positive Feedback, Clark Johnsen, countered the above traditionalism with this far bolder and more intrepid statement: "Theory told us (or so we were told!) that a certain number of samples, a few simple operations, some well-understood error correction coding (ECC) and so forth would produce a fixed result, at least within the digital domain. Bits are bits! It's all in the numbers!" "But it wasn't a done deal.........Few were forthcoming, and none that addressed the situation overall. None, that is, until Mark Porzilli advanced a narrative so daring, so unheard-of, so ritually unacceptable that many tried (as they often do) to laugh it off the stage." The First audition; "Now, the Memory Player....We are gathered—Rob Hart (former columnist-The Audio Tweakers), Alan Eichenbaum (former columnist-The Audio Tweakers, President; Scaena Loudspeaker Corporation) and myself—in Rob's spacious Fort Lauderdale living room this past December. The occasion: My first private hearing of the Memory Player from Nova Physics, about which I had been first to write. Those views were later expanded by Arnis Balgavis and others elsewhere. In fact, the Memory Player was felt by every auditioner to be quite the finest CD source they had ever heard, if at a price." "We are listening to whatever had been stored on this unit's disc drive by visitors to the Stereophile show. Yes we might upload something ourselves, but you know what? Everything sounds so good, so listenable, so faultless sonically, so absent of enharmonic edgy artifacts, so musically involving, that we almost passively accept one selection after another. This is not our usual style!" "The Memory Player, while presenting far and away the finest digital reproduction I've ever heard—including SACD. cj" The Clocking Myth "Time is on our side...yes it is" You all know that a CD is a 16bit / 44100hz medium. This means that every 1/44100th of a second, it plays a block or blocks of bits. This period of TIME is the very nexus of the difference between mechanical CD players and Memory Players. Clock Accuracy and Synchronization "connect" the Player with the DAC and all that is necessary in between. What is important to note is that these genuinely miraculously accurate clocks synchronize the Sample Period with all other parts of the chain. The sample is clocked to the chain of components, NOT the bits in the sample! If it sounds as if you are towing a car without checking if the occupants have left it, it is. The Sample Period is, in CD's case, is a 1/44100th of a second window of time where the laser may seek bits. The bits "float" around during the sample period, a 25 microsecond period of random bit dropping, corruption and bit inversion (from noise) and many times, properly read bits. What this demonstrates that these clocks may synchronize samples containing bits with incredible accuracy, but they allow bits to be read as late as 25 microseconds late, or any number from 0-25 microseconds late. In other words, while the sample is in sync with the rest of the system with great accuracy, the bits are NOT! The bits can be read at any time during that sample period, correct or not, and still be considered perfectly read or "Bitperfect". The sample period is 25 microseconds long and the bits in the sample can corrupt in a myriad of ways, many described below. Analog Analogies "The more things change, the more they stay the same" Alphonse Karr (1808-90) Our universe is not contiguous. It consists of packets of energy that move in quantized time periods. This was first discovered by Max Planck in 1900, the elemental quanta. Science was forever changed. This discovery made things like Relativity, the Atomic Bomb and the CD possible. Because of this discovery, we are forever mindful that there is no true analog because of this and that digital science is as analog as...analog. Here we explore the near identical nature of CD with vinyl and the reverse: As an analogy, if we listen to vinyl and the stylus skips a groove, we don't get to hear those few seconds of music in the skipped groove. Same with CD. A vibration, a speck of dust, whatever and a bit is dropped; we will not hear the music that bit represented. This all changed with the introduction of the error codes and ECC. A dual-nature example: When the groove of the record is skipped or when a bit is dropped, the error correction devices create a simulation of what was most probably in the skipped groove/or dropped bit and inserts where the skipped groove/bit hole would have been. We now hear a composite of a computer-generated estimation of what was there and actual music information, mixed. If it sounds weird and artificial, it's because it is. Had the introduction of error correction not been implemented in CD, we probably would find that CD is really quite similar to vinyl, without the tracking and resonance troubles in the bass. But despite the addition of error correction bits, the original musical information IS still on the CD. And there is a method to extract only music bits and therefore, play only music: Memory Playback Memories of Memory, Caches & RAM Bad memories of great music It is no simple task to explain how complex it is to remove music bits from a CD without removing ECC and its cohorts from a CD. Every time the music stream from the CD is fragmented by the placement of portions of it on memory, it must be reassembled if you will. As ECC required the CD have enough time to look back and add recreations of what was lost, it also required a place to hold this section of bits while this process took place. In other words, in order to use ECC, they created two more areas of potential bit re-ordering, RAM and THE CACHE. -------------------------------> As mentioned above, the cache is a small piece of memory that holds a portion of CD data before it is sent on to the DAC for listening. The cache creates this delay to ensure that the CD has enough time to rotate and upload the next sequence of bits as CD rotation is not very accurate. When uploaded blocks of bits have reached the cache, they no longer can be "corrected" in any way. In fact, the cache is extremely vulnerable to corruption because the rest of the circuitry is "blind" to the cache. It gets worse: The method that caches write the bits to it is in the order that is the fastest, NOT the original order of music bits. In fact, even NOISE can cause a "Logic Inversion"(change a zero into a one) while data is in this most vulnerable of states, "caching."------------------------------> And if ECC et al was the only area of digital audio corruption, we could stop here but in fact, there are different types of memory that can preserve or corrupt the order in which the bits are written! For example: The type of memory used in CD player's caches and a computer's RAM* can destroy original bit sequences and reorder bits to populate more quickly, (*Random Access Memory...need we say more!) rather than the original music bit order. This circuitous path is somewhat illustrated in the block diagrams indicating how caches and RAM must "decompress" (reversing the incorrect bit order in which it was written) before the information is usable. With sequential memory, it is written to more slowly, but it was never out of order so there is no "compressed" information to try to return to the original bit order.-------------------------------------------------------------------------> These error correction codes allow a recreated error correction bit to be created as much as 25 microseconds LATE and still fall under the title of bit-perfection. If we convert the TIME these codes are created when a bit drops, and the time the replacement bit is actually played into actual distances of late playback, the differences can be.................confounding. As if you're hearing the CD for the first time (see "Testimonials"). Any bit recreated within that much error in time is considered perfect or bit-perfect, as the lingo goes. So the "perfection"they extol is proven by the math itself as "perfect up to a point" or within a tolerance which of course MUST be the case with everything in Nature. To assert that ANYTHING in mathematics is "perfect" is contrary to the nature of mathematics. Math is not infinite. It creates explanations within a range of what is considered acceptable but NEVER, "perfect". Real-World Perspectives: Being late is not better than never To put that time error in perspective, sound travels at a specific rate in air at most elevations where humans dwell, so the "lateness" of bit recreation and the amounts it can reorder bits and still qualify as "Bitperfect" can be directly translated into physical distance. If you had a 2-way speaker with a tweeter and a midrange or mid/woofer,crossed over at say, 5000hz, the bit error timing, still within BitPerfect parameters could suffer enough delay to equate to moving the tweeter 6 INCHES behind the midrange! And the "lateness" can change randomly from 0" to 6"--------------> Serious, random phase shift. Imagine this happening on your speakers as you play CDs, because it does. But it gets worse. Our brains compensate somewhat to continuous sound aberrations, and we adapt. For example: Carnegie is soft, Fischer is bright, we know. Adapt and the concert sounds wonderful. If you've ever sat in front of a bright speaker for hours you'd notice that everything sounds dull until you have time to readjust. These ECC codes exhibit this TIME error RANDOMLY. So to return to the tweeter analogy, the tweeter would in effect, move from 6" behind to 1" behind to 3" behind, or any value in between - completely RANDOM, and our brains cannot "lock" on to this completely unnatural phenomena. It just sounds artificial, because IT IS. We now can see why we simply cannot resolve random changes in arrival times of portions of the music. They sound synthetic, artificial and create harmonics (as do all sounds) of these artificial fundamentals, adding still more artificially created sounds that have NO relation to the music bits as they were created after the music has been played. They blur the music bits as they pollute them, and the artificial nature of them is omnipresent. A well-pressed CD that is pristinely clean will of course generate less of these codes but regardless, they are literally generated continuously as a CD plays. It has been paradoxically suggested that "ECC is why CDs are bitperfect, yet ECC is rarely invoked". -------------------------------------------> This contradiction has been suggested many times. It makes no sense, so it would make no sense to address it. What IS important is to stress that there are MANY error correction codes that are invoked literally on every block of bits. Parity Bits and Sub-coding are just two. We explore these and others in greater depth in the "Technical Guide" (for the Non-Technical Reader) and we make note of them here to alert the reader that there are many types of error correction codes and still more errors generated by the wrong types of memory as well as the dissection of the bits and their reconstruction never being quite "perfect".Nothing is.-----------------------------> It has been said that "Reed-Solomon error correction is perfect so if the Memory Player sounds better, it is not for this reason". They are mathematically, subjectively,objectively, wrong. See the "Technical Guides". Streams of anti-Memory Player rhetoric weaken in the audio forums as the Memory Player owners grow in number, as more the rave reviews continue to come forth, and as the copycats grow in number and release their own "memory players". The traditionalists are losing the fight. It is a fight as predictable as human nature itself as big egos strain to protect the small knowledge that they fear losing. As for Nova Physics, after 4 years of study and 2 years of production, we now believe, and categorically assert, that it is not and never has been the RESOLUTION (16/44.1) of CD that was the dominant difference in sound between digital and analog audio. Consider this: BOTH use a quanta of something to represent recorded music, BOTH try to play as much of that original information as possible, BOTH never reach perfection but come close to 100% and BOTH have no fundamental reason to sound essentially different aside from resolution and strengths where electronic playback will not suffer mechanical playback limitations. It is Redbook CD which suffers the most from that which ECC and other error correction technologies creating bits that were not present on the master, pollute the original and do not represent anything natural in the recording of the music. The Memory Player(tm) with RUR(tm) "The most sophisticated CD playback system ever designed" A brief retrace for continuity's sake: A music bit simply represents a moment of music in time. If ONLY music bits could be heard, the only difference between analog and digital would be almost nothing but resolution as nothing artificial exists anymore. Even vinyl is quantized as is all of nature. To that end, enter The Memory Player(tm). The Memory Player(tm) extracts ONLY music bits from a CD, and deposits them on non-mechanical, SEQUENTIAL memory to avoid jitter from spinning ROMs and hard drives. It's pure, electronic memory. It can't move so it can't shake, it can't shake, no jitter. Measurable, almost ZERO laser (read) jitter. It also suffers no "Bit-Reordering" from the use of Random Access Memory or other memory that populates with bits according to the nearest vacancy on the memory instead of using the next, sequential space (1,2,3...etc). How it accomplishes this is when The Memory Player(tm) misses a bit, the location is notated, and the laser returns to THAT area and rereads the CD instead of using ANY form of error correction or recreated bits. This technology known as "Read Until Right"(tm) or "RUR (tm)", was developed at Nova Physics in 2004 to replace all forms of error correction. It can literally reread 150X to try to locate a dropped music bit, as there is no error correction anymore. It will move the laser in many different positions "looking" for lost bits until it finds them, reads them and deposits them on the memory in the original sequence. If it fails to find a dropped bit, an instant of silence occurs, not unlike the record skipping a groove although these instances average under 2%. Once complete, the music bits only play from non-mechanical, sequential memory. The latter, not being mechanical, as it cannot vibrate, has NO jitter. As it populates in numerical order, ONLY music bits are heard as they were heard in the studio. The rereading has captured virtually all of them, and more often than not, you are virtually listening to the master tape when you play a CD from memory. On many CDs, clean and with no scratches, The Memory Player will reach 100%, or virtually the master itself. But even Read Until Right (RUR) can drop a bit but what you do hear was on the master. You hear nothing artificial or created afterwards. In a very real sense, it IS "analog playback of digital music". Odd, but if you think about the process, it's virtually identical to a record or a tape. "Nothing is new, just newly discovered". So in a sense, The Memory Player (tm) is not new. And neither are the pronouncements of denouncements of self-appointed "experts" trying to reduce high technology world to a level they can understand. |






















| "In summary, Error correction IS the primary difference between digital and analog sound for once it is removed, all that's left is... ...MUSIC." |
| ++++ = ? |


| Irving Reed and Gustave Solomon |













| "When observation conflicts with knowledge, it is then time for the observation of knowledge" |






| YELLOW Music Files TURQUOISE DLLs BLUE Corrupted Memory Area RED Defragmented Files VIOLET System Files WHITE Vacant Memory Area |