Your beloved "all-analog" post-1980 LP-pressing was probably made from a digital master

 Killer 16/50.4 digital decks with linear-phase anti-aliasing filters were introduced in 1980, and were used for making digital copies of analog masters which were indistinguishable from the analog masters. So, for those albums, time ceased to exist at that time. These digital recordings have been used as source-material for LPs since then, although we've been given the impression that LPs are "all analog." In fact, digital signals ARE analog, consisting of two analog levels with brief analog transitions between them. So, claiming that LPs made from digital masters are "all analog" isn't a lie, although it is misleading. 

Mobile Fidelity customers were outraged a few years ago when a MoFi employee admitted that they had been using digital masters, but they should have been impressed by how good digital can sound, if it fooled them into thinking it was pure, continuous analog for so long. But due to the placebo effect (we hear what we want or expect to hear, at least until the reality breaks through the mental block), a lot of them probably thought that their MoFi LPs suddenly sounded worse. 

MoFi had Theta Digital design and build one of the more unusual audio ADCs ever, using parallel 16-bit Burr-Brown ADCs to achieve a high sampling rate in order to allow the use of simpler, cleaner anti-aliasing filters, and then using a digital filter to reduce the sampling-rate to 44.1 kHz and roll the signal off at about 20 kHz. It was kept confidential, and there's not much information about it on the internet even now, but it seems like it would have had excellent SQ for dubbing analog masters to digital. 

Chesky had its 24-bit oversampling process at the dawn of the digital era, but it was patented and they kept it to themselves for use on their recordings made with a minimalist microphone set-up. The ADC was designed by Bob Adams, who later designed ADC chips for Analog Devices and became an Audio Engineering Society Fellow. Chesky's ADC used 1-bit technology, which generates relatively high output noise in response to clock jitter, but it was a discrete design, which made it easier to obtain high performance. Cymbals are particularly difficult for 1-bit ADCs to get right, but Chesky's reportedly did an excellent job on them. 

Sony's PCM 3324, a 16-bit, 24 track deck which used counter-type converters and sampled at 44.1 and I believe 48 kHz, was a popular digital deck, but its input filters weren't so hot. It was introduced in its original form in 1981, and with the ability to cut and splice tape in 1984. In late 1985, Apogee introduced linear-phase filters for non-oversampling decks, and it transformed them into near-audiophile grade machines. Song X was one of the first CDs made with Apogee filters (which in that case were installed on a 3324 at Power Station, NYC), but digital mixing with 16-bit source material wasn't very good in 1985. The 20th anniversary release is much cleaner, although the first cut was made with Sony's stock filters so don't judge the CD by it.

The 1985 Steely Dan CD fiasco

When some renegade mastering engineers who worked at a record plant where Steely Dan's records were produced got their hands on a PCM-1610 with Apogee filters, they decided to make an unauthorized release of Dan's albums on CD, using what I assume were 3rd-generation analog masters which were improperly stored at the plant. But Katy Lied's original master was Dolby-encoded, so it still sounded great despite being a deteriorated 3rd-gen copy. (The multi-track and the first initial stereo master were dBx-encoded, but the stereo master had problems, so some of Dan's recording engineers made another, but with Dolby, and with the benefit of their experience on the initial master.) The CDs got produced and out to the stores before Roger Nichols, Dan's main engineer, got wind of what had happened, and freaked out. I think he realized that the resulting Katy Lied CD essentially amounted to releasing the original master in CD-grade lossless digital. It's the best CD-release, and possibly as good as the LP.

None of the other circa-1985 Dan CDs are worth getting because all of the non-Dolby analog masters at the record plant besides Gaucho's were so lousy, but the Gaucho CD ended up with a slightly fast speed. Aja's high end is borderline and has a peaky boost. But back then CD players in general were awful, so few noticed how good KL sounds or how bad most of the others sound. 

The original Dan CD-releases were made from the same analog masters, but in 1981 when they were in better shape, using a PCM-1610 with the phase-distorting stock input filters. Dan tried to give us the impression that they were made from the incredible 3M masters which were made from the original stereo analog masters, which they wanted to downplay. At that time, they wanted us to believe that digital was inherently inferior to analog, and that the LPs were made from pure analog masters. Becker and Fagen, notorious for their BS, concocted a story about one of the reels from Aja's original stereo analog master getting stolen from a car in a parking lot, as if that situation would ever exist, and that there was a reward for its return, as a hint that they no longer needed the analog masters after making the 3M copies. 

3M decks used the same DACs for recording (as part of the ADCs) and playback, so they had net zero linearity error. Dan bought a few 3Ms so that they could use the same decks for recording and playback, for the ultimate SQ. In 2006, they played their 3M masters on the decks used for recording them, and dubbed the analog outputs to 24-bit copies, which are used for the LPs. In 2007, the famous Cisco Aja LP was released.

The digital masters for the 1999 CDs were made in 1998 from the analog safety masters stored in Roger Nichols' personal archive, using an Apogee AD-8000 24-bit ADC. They're not bad (although Katy Lied has inverted polarity and some other forms of degradation), but the LPs would be much better. 

High-res ADCs

The next major improvement in digital recording technology took place in 1993, with the introduction of high-end 20-bit oversampling ADCs. In 1997, 24-bit versions were introduced, and more advanced 24-bit ADCs were introduced in about 2005. 

TTs etc.

I didn't take Fluance TTs seriously for a long time, because they struck me as something designed just to look good and make a lot of money, but when I finally took a look at them, I was impressed by their innovative, cost-effective designs, including such things as a heavy molded MDF chassis mounted on soft rubber isolators to kill vibrations, belt drive with regulated platter-speed for belt-drive smoothness and direct-drive's speed accuracy, and a $200 retail cartridge on a $500 TT. Well-Tempered Labs TTs supposedly have high-end SQ without the high-end price, but they aren't cheap. Avoid linear-tracking tonearms, which are high-priced gimmicks with problems with no good solutions. I designed a cheap one and wondered why there are no such designs, until I realized that the high price IS the idea. A simple pivoting tonearm with a fixed head-shell is the best configuration. The use of suction to clamp LPs to the platter is appealing, but it's expensive, and few high-end TTs use it. Belt drive with a speed-regulated platter seems to be the optimal drive system for SQ.

A deep-cleaning LP-cleaner isn't optional, if you want to get the best SQ and longevity from your LPs. Even new LPs should be deep-cleaned before playing them for the first time. For details, seek advice from multiple experts, because bad advice on this subject can be very expensive.

Some CD myths debunked

CD-grade digital has about 20-bit DR due to the use of narrowband Nyquist dither, which has been in use since the 90s. Other forms of dither were used before that. The idea behind dither is to randomize the quantization level of bit 16, and substitute noise for the horrendous distortion of low-level signals caused by quantization and linearity errors. As with analog, low-level details can be heard through the noise, which in the case of narrowband Nyquist dither is ultrasonic. Being able to hear low-level details constitutes higher DR. It took a math PhD 75 pages to explain it thoroughly, so excuse me for not doing so.

The notion that "precision" CDs (such as SHMs and Blu-specs) sound better due to a lower error rate is a gimmick. They couldn't possibly have a lower corrected error rate than standard CDs. (All digital storage systems include error-correction.) SHMs sound better because they're made from better masters, not because of a lower error rate. Blu-specs are typically awful, such as Bitches Brew, which was given an insulting "ear candy" EQ for the BS-CD (2008, 2013, 2023) to give the initial impression that precision CDs sound better. The Amazon "2022" BB CD is one of the best BB remix CDs. I prefer the 2005 Japanese Mastersound release, which was perhaps digitally remastered from the stereo analog remix-master using an advanced ADC. Mastersound CDs are made from high-resolution masters using Sony's proprietary Super Bit Mapping, which includes a type of dither which gives the resulting CDs 20-bit DR, and some other processing. The 2006 "DSD mastered" CD might be better, but not due to "DSD mastering," which is a meaningless buzz-phrase. But the 2006 CD is expensive and hard to find. The 1999's high end, as I recall, is a touch dirty, perhaps because it's more extended than it is on the "2022." Both were made from the same digital master, which I suspect was made from the mixer's output. All of the aforementioned BB-CDs are made from the remix and have the correct polarity.

High-res myths debunked 

Only recording engineers need high-res, so that they don't have to worry about setting the recording levels on live feeds precisely to avoid clipping or losing detail, and can mix digitally without audible round-off errors. But not even they need the absurd sample-rates pushed by some outfits. Pro-grade-ADC designer Dan Lavry claims that anything above 60 kHz actually degrades SQ, although I surmise that 96 kHz isn't too excessive. Apogee concurs with Lavry. For more, Google "Goldmund Does high resolution audio sound better."

Modern ADCs use very high sampling rates at the input, partly to simplify input filters, but these samples are 5-bit samples of differences in the input signal between sample intervals, not 24-bit samples representing actual waveform values. The 5-bit samples are obviously converted to 24-bit samples in the ADC-chips.

Digital interpolation filters (a.k.a. "oversampling filters") calculate and insert extra samples into the data produced by CDs, to turn the coarse "staircase" waveform from CDs into finer staircase waveforms, so that simple, cheap analog filters can be used to smooth them out completely without introducing any audible distortion. Without digital interpolation, it's expensive to create filters with a sufficiently steep rolloff and linear phase. 

Back when the digital interpolation filters (DIFs) on DAC-chips sounded artificial because they were based on compromised designs to allow them to fit on the low-density chips of that era, 96 kHz recordings might have sounded better than 44.1 because they didn't need as much digital interpolation, so they weren't degraded as much as 44.1. High-end gear used natural-sounding "external digital filters," i.e. high-performance DIFs on separate chips. But external DIFs are rarely if ever used these days because the ones on audio-DAC-chips have gotten so good. Chord's DIFs supposedly sound better than typical DIFs on DAC-chips, but if so I would think that the likes of Audio Research and Bel Canto would use similar designs instead of settling for DIFs on DAC-chips. Audio Research got excited about a new tube socket design a few years ago, so they wouldn't put up with a bad-sounding DIF. 

Why some CDs sound worse than the LPs

From the music industry's perspective, the problem with CD-grade or "better" lossless digital is that it's easy to copy it perfectly and anonymously, and it lasts forever. LPs, on the other hand, can't be copied perfectly and don't last forever, and people are leery of buying used LPs. So, to minimize financial losses due to piracy and used-CD sales, they typically don't put the best masters on lossless digital as long as the LP is selling well, which in the case of some LPs can be decades. The initial CD-release is typically the worst, and each new release is a little better. In some cases, the audio is degraded by simply inverting it. In other cases, such as the 1999 BB CD, the digital master is converted to analog, perhaps processed,and converted back to digital with an inferior ADC. I have a Bach CD which was degraded by mixing a key harpsichord passage in at such a low level that it's practically inaudible. I had a Brecker Pilgrimage CD which was so compressed and rolled off at each end that it sounded like a toy version. So, there are many ways to degrade digital masters for CDs. 

I doubt that the best version of the original Sgt. Pepper's is available on CD even now. 

Lossy recordings

Lossy digital by itself might sufficiently degrade the content so that it might not be necessary to degrade it otherwise to protect it from piracy. So, the best masters might be released in lossy digital, which might explain why some people like lossy releases. HD Radio, which is lossy, sounds quite good if you don't listen too closely. Unfortunately, FM is a wasteland now that Sirius XM, which pays more to content-creators, is available. I haven't heard Sirius, but it might use lossless copies of the best masters. As I recall, Sirius XM receivers use TI/Burr-Brown Advanced Segment DAC-chips (explained later), which are used in some of the best high-end DACs.

Watermarked streams

Each stream is watermarked so that it can be traced back to the purchaser, so that if any copies show up for sale, they're easy to trace back to the stream's purchaser. So, piracy isn't a concern with streams, and good versions can be streamed, although the industry probably still reserves the best versions for LPs, to give us an incentive to buy LPs, turntables, etc.

Un-degraded oldies

But the best versions of albums that aren't big sellers might be available even on CD. I've found a bunch of my old favorites, including Jean-Luc Ponty and Spiro Gyra, that sound just as good to me as the LPs ever did. In fact, the cymbals on the 2016 Pat Metheny Group SHM-CD, with its audio polarity inverted, are cleaner than they are on the LP, because the cymbals are a type with unusually high frequencies, which are hard to put onto and get off of LPs. I had to use my headphones to hear the difference. To invert, I use an Audacity macro consisting of the select-all, invert, and export-as-FLAC commands. But I had to go through a few CDs in some cases to find the one I like.

Steely Dan's Alive in America is available on CD only, and Dan pulled out all the stops for that CD. It was recorded and mixed in analog since digital mixing had problems back then. The mixer's output was digitized directly with an Apogee AD-1000 20-bit ADC, and Apogee UV-22, the original narrowband Nyquist dither, was used to convert it to 16 bits with 20-bit DR. So, Alive is limited only by the playback system. 

Digital playback gear

Then the challenge is finding good-sounding digital playback gear that you can afford. I have a Zen DAC V2, and it's probably the last DAC I'll ever need. After hearing it, I couldn't stand my previous DAC, a Topping D10s. 

The keys to the Zen's SQ is that it uses a "femto-clock" (a clock with sub-picosecond jitter specified in femtoseconds) and TI/Burr-Brown Advanced Segment DAC-chips (ASDs), each of which is a combination of an "R2R" DAC for bits 1-6 and a delta-sigma DAC for bits 7-24. The R2R DAC is split up into 63 individual segments (which combined with zero gives 2exp6 = 64 codes). This allows the use of a randomization technique which gives the segment-DAC 25-bit linearity by converting its linearity errors into ultrasonic noise. It would be impractical to use this technique for more than six bits, so a DS-DAC, which is inherently linear, is used for bits 7-24. The advantages of R2R/segment DACs is that they have low jitter-sensitivity, meaning that they don't produce much output noise for a given amount of clock jitter, and thus have a clean high end. DS-DACs have relatively high jitter sensitivity, but because the DS-DAC section in ASDs covers only the bottom 1/64th of the total amplitude range, its output noise is scaled down by a factor of 64 and is minuscule at the ASD's output. So, ASDs have the advantages of R2R and DS-DACs, without the disadvantages. They're good enough for Audio Research and Bel Canto, who make some of the best DACs in the business, without the stratospheric prices of some high-end DACs.

For best SQ, I leave the Zen powered-up 24/365. Reviewers leave digital gear powered-up for at least 24 hours before reviewing it. It doesn't always make a difference, such in the case of my Topping D10s, but it does make a difference for the Zen. Nobody seems to know why, but it does.

The 20-foot USB3 cable between my PC and my Zen, which I bought from Amazon just for that purpose, would cut out from time to time. I assumed that the connectors were the problem, so I reseated them, and this usually fixed the problem, or so I thought. After about two years of this, the cable cut out, and reseating the connectors no longer fixed it. So, I replaced the cable with three short cables and was blown away by the improvement in SQ, compared to the SQ with the long cable when it was "working." As a result, I realized that the actual problem was that an internal connection was barely making contact, and that reseating the connectors was just wiggling the cable and restoring this connection to its normal data-degrading state. So, I recommend getting a USB cable tester to identify any problems quickly. Any standard USB cable in good condition should sound as good as any other. It's OK to use long USB cables for CD-grade audio data because its data-rate is nowhere near the maximum data-rate for USB cables.

Music players

The only music players that have actual bit-perfect capability and can be trusted to retain that capability, without someone sneaking in any dirty tricks to interrupt playback or degrade the audio, are the Android players USB Audio Player Pro, Neutron, and HiBy. I tried HiBy and gave up on it, but some people apparently love it. Neutron reportedly has optional transparent DSP (including polarity inversion, which is a simple logical operation built into some DAC-chips, including some TI/Burr-Brown Advanced Segment DAC chips - invert each sample and add a 1 to the LSB), and the Poweramp player also supposedly has transparent DSP but not a bit-perfect mode. A few player-developers have tried to use Android's built-in bit-perfect mode, but found problems with it, which explains the lack of players that use it. You'd think that Google could devise a simple bypass for Android's audio system.

Now that Google is working on an Android/Chrome OS for PCs or Chromebooks, it will be possible to run Android players on something besides portable devices.

System maintenance
 
I have an old Denon analog receiver and live in a dusty area, so dust gets into the receiver through the ventilation slots on top of the receiver. This dust forms a layer on the circuit board, which acts as a parasitic circuit. Once a year, just after the monsoon season hits (which settles the dust and increases the humidity, and thus the parasitic effect of the dust-layer on the circuit board), I use an air-gun to blast the dust off the circuit board, and I reseat all of the connectors in the system, including the ones inside the receiver and the AC power connections. If I'm feeling energetic, I also cycle all of the breakers out to the power meter (which I do twice a year in any case). You could also use alcohol and a small paint brush to loosen any residual crud on the circuit board, and then let it dry and hit it again with an air gun. I never cease to be amazed at how this annual maintenance transforms my cheap system's SQ.

Speakers

My dream speakers are LX-Minis with digital crossovers with external Zen DACs. They're a sort of B&W Nautilus for those who can't afford to shell out about $80K for a pair of speakers. If I wanted to play LPs, I'd get the analog crossovers.

Ground loops

Everything in an audio system should be plugged into the same outlet, or the same outlet-strip etc. Otherwise, strange problems might occur. For example, I plugged my PC into one outlet and my audio gear into another, and the music player installed on the PC would cut out when the refrigerator turned off. Eventually it occurred to me to examine the set-up for ground loops, and found one. So I plugged the PC into the same outlet as the audio system, and haven't had that problem since then. Another strange problem also disappeared at the same time.