One sometimes needs to be careful when comparing pre-recorded CD material (or any digital source) vs. vinyl. Here is one major problem with some of the older digital releases -- not 'digital' per se.
This matter is about recordings made between late 1960's and early 1990's. This does not address loudness wars issues or other mastering defects. I am discussing a mastering defect that has been overlooked until recently.
Back in the old days, the tape decks were noisy (unlike digital tape and other recording methods today), and there were means to mitigate some of the noise. The first common version of noise mitigation was the DolbyA NR (noise reduction) system. It compressed the levels somewhere around -20dB to -40dB in 4 different bands (approx LF:0-80,MF:80-3k, HF0:3k-9k and HF1:9k-20k). The issue of HF0/HF1 are a bit confusing because the HF1 band is actually the parallel action of two NR mechanisms. But basically, the gain control/NR bands are as I stated. (DO NOT CONFUSE WITH THE 3k-20k and 9k-20k compressors -- that is what I was speaking about the NR mechanisms.) To get the original signal back, the DolbyA HW could 'undo' the encoding process -- thereby restoring something similar to the original sound -- almost.
So, we have those old tapes recorded with DolbyA encoding. During the transition and afterwards the conversion from the analog tape to the digital media (per Library of Congress procedures) is to copy the base audio tape contents to the digital copy. There is no specification of using the decoding process before recording onto digital. (the process does seem to be that digital copies were created with the NR encoding -- but there WAS a method to the madness -- but it wasn't handled on the other side.)
Next, when producing the digital distribution, the digital copy is transferred from the digital media, mastered, then the distribution copy is ready. One my say "use the DolbyA decoding process", right? Well, it sometimes doesnt' get done... Why doesn't the decoding get done? Because it requires real-time to do the decoding -- first, the material is converted to analog, so XX minutes is required to convert to analog, the signal applied to the DolbyA HW decoder, and then the results encoded back onto digital. That XX minutes and the DolbyA HW setup is skipped from time to time. (I have found maybe 1/2 of the time or so, maybe more or less.)
So, us digital listeners sometimes have DolbyA compressed material (that is what DolbyA sounds like is mostly an HF compression, even though it actually compresses all four bands.) This 'compression' is sometimes stealthy on certain kinds of material, and other kinds it is obvious as day. One of the culprits adding to the stealth is the mastering person who does a -3dB @3kHz or -6dB @3kHz, which mostly hides the DolbyA frequency response perturbation, but still the 'harsh-like' DolbyA sound persists. Places like MFSL seem to do the proper decoding, but normal labels don't always do the full decoding.
What does DolbyA encoded material sound like (esp after the -6dB compensation?) Well, the biggest difference is that the highs sound unnatural, kind of compressed, kind of springy... (DolbyA is a damned fast compressor -- used on Karen Carpenters' vocals *as a compressor*.) Also, the spatial/stereo sound is flattened. This is where sometimes the digital copy gets the 'sound stage' complaint.
I know that there will be claims: "that man is crazy", well... Maybe I might be crazy, but not due to this matter... The best way to demonstrate is by existence proof, and not claiming authority, or this person said 'this', that person said 'that', or my big recording engineer friend said that it cannot happen. Sometimes even those who claim authority can be wrong... I truly am an expert in certain fields, and I have been wrong from time to time also.
I have uploaded some examples -- various recordings -- with matching 'decoded' and 'undecoded' versions. The 'undecoded' were 'leaked' DolbyA copies available on CD or download. I even have a recording from HDtracks that was not DolbyA decoded. The 'decoded' versions were processsed by a DolbyA compatible decoder -- it is a little different from a true DolbyA -- sans intermodulation distortion and a few other advantages afforded by LOTS of CPU.
The issue here isn't the software DolbyA compatible decoder, the issue is that a lot of the complaints about digital sound come from leaked undecoded DolbyA material. My copies were puchased openly and nothing under the table.
Listen carefully to the undecoded and decoded versions, and very likely the difference/improvements from decoding will be heard. Almost ALL vinyl is properly decoded because of previous procedures/processed. Many digital distributions are NOT processed completely.
So, we have the --undecoded, and --decoded versions.
Additionally, there are -Polar and --DHNRDSDA versions of a few recordings -- the Polar versions were done using a true DolbyA unit, and the DHNRDSDA were done by the SW DA decoder, with the anti-IMD (therefore producing more clear/less gritty results.)
EXAMPLES:
https://www.dropbox.com/sh/mduka8faqv1nva7/AAATBBBRIFDht8pVsDN5Dv7Aa?dl=0
John