Anyone with a TV knows that high-res recordings sound much better than CDs. My cheap old system can't reproduce very high frequencies very well (although once in a while I'm very satisfied with it), and yet I can clearly hear the superiority of TV-audio over CDs. We typically assume that this is due to the higher sampling rate and resolution of high-res digital, but according to various articles by audio authorities, including High-Resolution Audio: Does it Sound Better? by Goldmund Acoustic Laboratory (posted to Goldmund.com), so-called high-res recordings for consumers use only about 16 bits of the format's 24 bits, and sampling rates over 44.1 kHz are useless at best.
An article entitled Is high-resolution audio really as good as it sounds?, by Ian Paul of TechHive, cites experts who contend that high-def audio sounds better because "sound engineers often put more care and attention into higher-resolution recordings than they do to mass market CD releases." This implies that CDs could sound as good as high-res recordings if the engineers would do a better job of the mastering, but I suspect that the mediocre mastering is deliberate, because the music industry doesn't want to put their best recordings on CD, at least when the corresponding album is new and popular, because it's easy to make perfect copies of CDs. (There are error-correction systems built into the CD record/playback system. Software is stored on CDs, and it can't have any errors. CD-treatments didn't reduce the corrected error-rate - they improved the sound quality of early cheap CD players with poor isolation between the transport and analog power supplies, by reducing laser-servo hunting and the resulting noise on the analog power supply.) So, at first they give us versions made from substandard analog masters, or a poor mastering job, such as by using too much compression or bandwidth-limiting, or by shoving the detail down into the dither. [1] Later, they sell superior versions based on superior source recordings and/or superior mastering, and in some cases inferior versions which supposedly sound better because they're made with high-resolution converters or from the original master, which is used only because it has deteriorated. There are quite a few exceptions to this strategy, such as Jean-Luc Ponty's CDs [2], but I'd rather that it didn't exist at all.
Other experts contend that high-res formats do provide audible benefits. Dan Lavry of Lavry Engineering, which manufactures high-end DACs and pro-grade ADCs, recommends a minimum sampling rate of 60 kHz (36% higher than 44.1 kHz), and apparently a minimum of 18 bits, which practically speaking requires the use of 24-bit formats with 88.2 kHz or 96 kHz sampling-rates. Anything higher than 96 kHz, he claims, only adds problems. Based on my experience listening to CD-grade digital, I tend to agree with Lavry, because the high end on even the best CDs doesn't seem sufficiently "airy" or "open." So it seems to me that we need a new standard consumer format, with a bit-depth of 20 bits and a sampling rate slightly higher than 60 kHz, to which the existing high-res formats could be converted cleanly and fairly easily. Such a standard would minimize file size while providing the highest audio quality.
Good recordings and masterings are typically placed on LPs from the start, because LPs can't be copied exactly and easily, like CDs. Some people spend hundreds of thousands of dollars for a Goldmund Reference turntable, in an attempt to extract all of the detail from their LPs. I've heard a $10K Linn Sondek LP12, which for me was a revelation in palpability, and Linn has recently introduced a new design for the LP12's platter-bearing. So, LPs can sound much better than most of us will ever know.
A record-cleaning machine is absolutely essential for anyone who's serious about getting the maximum detail from LPs, as well as minimizing wear. A "dustbuster" just doesn't cut it. When you buy your first turntable, get a cheaper turntable if necessary to get a record-cleaning machine. You can upgrade your turntable later, but you can't undo record wear.
Tangential (linear-tracking) tonearms seem like a good idea, because they eliminate tracking error and skating force. According to Myles B. Astor, PhD, Senior Editor, Positive-Feedback.com, on AudioNirvana.org (https://www.audionirvana.org/forum/the-audio-vault/analog-playback/tonearms/49834-linear-tracking-tonearms),
"High-end audio seems to be filled with those products that offer the promise of exceptional performance but are held back by some flaw or another. Those products that you know in your heart can be fixed if given a little TLC. Then you find out after spending ages with that component that no matter what you do that issue will never be resolved.
"One such product is linear tracking (air and mechanical). I'm sure that many of us here at one time or another (or still do) in their audio journey mounted a linear tracking arm on the turntable. Yes, in theory linear tracking arms offer the ultimate in LP playback. But in reality, linear tracking arms are held back many issues not the least of which include:
Movement and ability to maintain tangency to the groove [some tangential arms use sensors to detect deviations from tangency, and servo motors to move the arm along the track, which I always thought was a bad design]
Solidity and low octave reproduction [some tangential arms are too light and flimsy]
Susceptibility to warp wow with short arms [some tangential designs have very short arms]
High horizontal mass [a problem when playing records that aren't perfectly round, requiring the entire arm to move back and forth to track the record]
Air flow issues [with air-bearing arms]
Machining issues eg. the best air bearings might have a 1/10,000 of an inch tolerance and finding a material that isn't susceptible to temperature fluctuation is challenging
Freedom of movement [excessive friction]
Inability to maintain azimuth across the record [airplane azimuth: wing up/down - I don't know why this is an issue]
Setup can be frustrating
"Yet, some of the best sound I've ever heard is from air bearing arms such as the now sadly discontinued Air Tangent tonearm. The arm's resolution, sense of spaciousness and soundstage were something to behold. But at the same time, the arm wasn't the equal of a pivoted arm when it came to reproduction of the lowest octaves. The of course, there's the issue of convenience. After a while, who wants to hassle with air compressors, filters, tubing running everywhere in the room. Especially in smaller apartment quarters as opposed to someone with a house who could put the compressor in another room, garage or basement.
"But honestly, I do miss listening to those arms and sometimes I ponder putting a linear tracking in addition to a pivoted arm on my turntable. Maybe one day."
That day might have arrived. Bergmann Audio of Denmark has devised an elegant ball-bearing design (see photo). The design concept isn't entirely new, but the implementation is simpler and more precise than its predecessors, which can be seen on the aforementioned Positive Feedback page. I suppose that the bearings are permanently lubricated with something like a Teflon film on the bearing races. I gather that the bearings aren't sealed because doing so would add friction. The bearings and track could just be blasted with canned air periodically to get rid of dust. It could probably be implemented for a few hundred dollars in carbon fiber (the bearings and carriage would have to be metal, however). It's not yet on the market as of this writing, so perhaps they're still tweaking the design, or perhaps they're afraid that it would destroy the market for the ridiculously expensive air-bearing designs. A lot of people would probably get a turntable with the Bergmann ball-bearing tangential arm just because it's so ideal.
But I'm not interested in the ultimate sound quality for myself - I just want to be able to enjoy music, and I find that one of the biggest factors in being able to enjoy a particular recording, assuming that the record/playback chain is fairly transparent, is being in the mood to listen to the recording. (My speakers leave much to be desired, but I can put up with them until someone comes out with a $1K pair knock-offs of the $60K B&W Nautilus, including the amps and analog electronic crossovers. The cabinets could be molded, the drivers don't have to be the ultimate, and the crossovers and amps could be combined into modules to be placed next to the speakers.)
So, I propose a system which would allow us to obtain good digital recordings, of whatever sampling rate and bit-depth we prefer (assuming that someone is willing to provide them, and as long as there are suitable players/DACs), without the possibility of piracy, and which would allow consumers to be charged a little each time they play their copy, so that the album doesn't have to be sold over and over again as better and better digital versions to keep making money. So, everyone would win, and it would be cost-effective for consumers, despite having to pay a little each time they play something. The source-file would be downloaded in encrypted form and stored in encrypted form on devices owned by the consumer. The files would be encrypted for the intended player, and only the decryption key stored on the player could decrypt them. Nobody would ever have access to the decryption key, which would be stored in EEPROM which would be erased within seconds if any of the following occurred:
A) The player's case were opened (which would require cutting or breaking the case, or special tools which would be illegal for consumers to obtain)
B) Any motion, such as a drill bit, were detected by motion sensors within the case
C) If the temperature got too low or too high, perhaps indicating a plan to kill the internal batteries, which would be highly reliable rechargeable batteries with a reasonable temperature range.
D) If the internal battery voltage got too low, perhaps indicating a plan to let the batteries die in order to access the decryption keys. The player would provide as much warning as possible to connect the AC adapter, depending on how quickly the batteries were being drained, before erasing the EEPROM. Replacement EEPROMs could be installed only under tightly controlled conditions at the factory, where it could be ascertained that nobody had had access to the decryption key on the EEPROM.
The player would also try to send an alert (including GPS data in this situation only, due to the potential for a crime in progress, possibly at some otherwise-unknown location) to the server immediately, or at least as soon as it could establish a link to the server.
There would obviously have to be special provisions for preventing the EEPROM from being erased during assembly or repair, when the case would be open. Perhaps the activation-code (which the server would normally provide, and which would be changed after every activation) could be entered, followed by a code to disable the motion sensors. Naturally, after the player is reassembled, the sensors would be re-enabled. This would be performed only at the factory under tightly controlled circumstances, and be done in a manner which would prevent anyone from having access to the decryption key or the activation code. (If the existing activation code and a new activation code aren't entered by a certain time and date, due to the owner's failure to pay the bill, the player would be deactivated, but obviously the decryption key wouldn't be erased, and the activation code would still be valid. When the owner pays the bill, the current and next activation codes would be sent the next time the player connects to the server. (There could be a button on the player to cause it to connect to the server, instead of waiting for it to do so automatically.) Obviously deactivation doesn't deactivate the player's ability to connect to the server - just the ability to play music. In fact, if the player fails to connect to the server every few hours, a music-industry security official would try to contact the owner to determine what's going on, and the police would pay a visit if the owner doesn't respond. Prospective buyers would be informed of these conditions to ensure that they are willing to accept them, and to let them know that there is no way to beat the system.
Huge rewards could be offered during the development phase for cracking the security, so that the best hackers would do their best to crack it.
The player would have analog outputs only, to ensure that nobody could build a device that would provide access to the unencrypted data. So, even the basic player would have an excellent DAC-section, which isn't difficult with modern DAC-chips. There could also be high-end models.
The player would keep track of what's played on it (the player would contain an FeRAM or MRAM - nonvolatile memories with better data-retention than flash memory at high temperatures, although with other problems such as relatively tiny capacity and relatively high write-current requirements). This information would be uploaded to the server at the end of the billing-cycle, and if the bill weren't paid in a reasonable amount of time, the player would be deactivated during one of its more frequent security-related connections to the server. The activation-code would be changed for each billing-cycle, and would be unique for every player.
When it's time to contact the server, which would be controlled by an internal clock (which would be synchronized to internet time whenever the player is connected to the server), the player would establish a secure link to the server by means of an IoT module. Although very little data would be sent over this link, it would have to be very secure because it would contain the authorization codes (current and next) and usage-data. The details would have to be worked out by a circuit designer who is familiar with each aspect of the overall design.
The manufacturing process would be automated and monitored by security personnel to ensure that nobody would ever have direct access to the decryption keys.
To convince people to switch to the new system, a representative selection of its music would be available in decrypted form for free.
Although quantum computers might some day be able to easily crack some types of asymmetric encryption (i.e. with separate encrypt and decrypt keys), there are already some types which quantum computers can't crack, and experts are working on others.
The server would serve files from all record labels, and be protected by a bullet-proof firewall, or a series of firewalls, which would limit inputs to the server to a specific format, and validate every input to ensure that it conforms to a valid selection and an activated player, and that the registered owner is making the request, such as via two-layer authentication and perhaps an iris scan, to absolutely prevent hacking.
There would also be a limit on how much someone could be charged per billing period, so that people can feel free to play as much music as they like, without worrying about receiving a massive bill. However, the amount would be sufficient to pay a fair amount to the music industry for any amount of music that someone would likely play in a billing period.
It would be convenient and inexpensive for the vast majority of people to download these files, such as free access to wi-fi hotspots for this purpose. Eliminating the need to download the music every time you want to listen to it is one of the main advantages of this system. High-res streaming requires a heavy-duty internet connection whenever listening, and universal high-res streaming would put a heavy load on the internet. If and when high-capacity NRAM becomes available, it will become feasible to download vast amounts of data to a wi-fi-connected phone at extreme rates, and store it indefinitely even at 200F. The files could be copied or transferred to a couple of large drives (primary and backup), which would serve as a personal music archive.
Some might object to this system by noting that the analog output could be digitized and stored as a regular FLAC-file, and played with something that wouldn't levy per-play charges, but the inconvenience, loss of quality, and lack of significant savings would prevent most people from committing this form of theft. Besides, this can already be done with LPs and high-res streaming.
Music could still be released on CDs for those who don't like the idea of paying a little each time they listen to a high-quality digital recording of their favorite music, and would rather pay more for a lower-quality recording on CD to avoid paying a small amount each time they listen to it.
Notes
[1] Dither is noise with a special frequency balance, with an amplitude of about 1 LSB p-p, added after mixing, etc. to mask the grittiness of 16-bit digital recordings, and to give the impression that there's 20 bits of dynamic range, so that for example fade-outs don't have audible cutoffs. With dither, bit 16 becomes a sort of DSD-bit, which allows it to represent details smaller than it normally would, but limited to low-frequency details, with high-frequency details converted to noise, which I believe explains the flakiness of the high-end on many CDs.
[2] The original CD-release of Ponty's album Open Mind (which includes outstanding solos by Chick Corea on synth, and George Benson on electric guitar) was apparently digitized with a Sony PCM-1610 16-bit, 44.1 KHz digital processor with stock input filters, so that the CD has a dry high end and mediocre imaging. However, the 1990 release was apparently digitized with a PCM-1610 with Apogee's superior input filters, which produced recordings with a detailed, liquid high end and good imaging, and the transfer to digital was otherwise done very well. The analog master was also excellent, so the CD is very clean and juicy. The music is great for unwinding. The 1990 release, which is sold by Amazon and perhaps others, is apparently distinguished from the original release only by differences of a few seconds in the listed duration-times of the cuts (the catalog number and packaging are the same for both releases). The 1990 release's listed duration-times are as follows: 1-8:05; 2-6:05; 3-4:57; 4-7:17; 5-5:14; 6-7:40.
