New to this thread.
As a forumer that is very active on the "turntable" section of Audiokarma, i find this thread exciting.
My take on this, not only as a turntable technician and record lover, but also as an engineer who actually had to go through courses and labs covering PCM analog-to-digital and digital-to-analog conversion, my opinion is the following:
We need first to discern between CD-quality audio (44KHz /16bit) and hi-res audio (88KHz or more of sampling rate, and hopefully more than 16 bit).
And let me start saying that I think digital audio is one of the greatest inventions of mankind. However...
If we're talking about
CD quality audio, this is a medium with the following characteristics:
1. It is a system that is bandwidth limited between 0Hz and (approx) 22KHz. Thus it is limited in slew-rate too.
2. The process of converting from digital to analog requires a reconstruction filter. You can do it in the analog domain, or the digital domain, or mixed, but the end goal is the same: a frequency response more or less similar to this:
View attachment 3201568
That is, a filter that has an insanely ridiculous steep attenuation near 22KHz. This is the often-named "brickwall filter".
This kind of response would make any analog audio fan or even speaker designer or even amplifier designer cringe.
This means the phase near the last octave (10KHz to 20KHz) will be completely wrecked.
And i'm not even talking about the pre-ringing and post-ringing...
There are many kinds of digital (and analog) filters and filtering implementation but the goal is the same, and the result is either you wreck the phase response or you get linear phase, but then what one gets is a non-flat frequency response -- a frequency response that has ripples... Etc. You can never win. Maybe that's why some people prefere non oversampling DACs with no reconstruction filters? Well... i digress... back to our topic.
3. The process of converting from analog to digital requires an antialiasing filter, which will do more or less the same, adding more injury to the signal, that is, more phase distortions, ringing, etc.
The impulse response of such an ADC->DAC system will not really be true to the original.
The solution is obviously using a higher sample rate, in this way, the filter doesn't need to operate close to 20Khz, for example if one uses 192KHz sample rate, one only needs to start attenuating at about 96KHz, well away from the audible band. Thus the impulse response is better, phase is preserved in the audible band, etc etc.
This problem was already known in the 1970s and the main makers of pro digital gear used 50KHz (3M, Soundstream) or 50.4KHz (Mitsubishi), and if they could have used higer sampling frequencies they would've done so, sadly they were pushing the limits of what you could store on magnetic tape. The data rate of digital audio at 50KHz/16bit/Stereo was really pushing the state of the art in digital storage at the late 70s. The AES recommendation for digital audio in the late 70s was 50KHz, and the Sony and Phillips guys were aware of this when designing the CD, however they went for 44.1KHz for other reasons that can be resumed as "Better compatibility with our PCM-to-videotape systems", which used 44056Hz (later 44.1KHz) for the reason that can be resumed as "it's the maximum we can reliably record on consumer Umatic/Betamax videotape".
So, CD-quality audio was already doomed from the birth.
4. And i have not even mentioned: digital quantization distortion... only because it is a problem that can be overcome if everything is done right.
Now, let's compare to the LP record, done correctly. Brace yourselves, because I am going to mention things that the layman often ignores.
Record-to-playback bandwidth: From (about) 20Hz to about 50 (fifty) KHz. Half-speed mastered records take advantage of this bandwidth.
Antialias brick-wall filter: None
Reconstruction brick-wall filter: None
Frequency response of the system does not decay like a landslide after 20KHz; it decays gradually. Thus there is none of the phase aberrations or impulse aberrations that the CD quality audio has.
Frequency response of a very good phono preamp is within +/- 0.1dB, an excellent result already better than almost every speaker out there. And phase is kept mostly unaltered since the (gentle, 6db/octave) RIAA filters introduced at the record time are applied in the exact inverse at the playback time.
Frequency response of a very good MC cartridge is ruler flat up to at least 12KHz, some of them ruler flat to 20KHz. Most have response over 50KHz. Bandwidth of the electrical system of a good MC system is over 100KHz so there is no slew rate problem.
Frequency response of the cutting head used to record the lacquer master is absolutely ruler flat in the audible band, thanks to the use of motional feedback. A really great cutting head (Ortofon DSS 731) reaches 30KHz (at -5dB), this means over 60KHz if doing half-speed-cutting. One audiophile label claimed recording over 100KHz on a lacquer disc, btw.
There are many more differneces that can be mentioned but i'm trying to keep it short.
I also want to say, for people who think that REEL TO REEL ANALOG TAPE is the superior medium, that the direct-to-disc (direct to lp record) recording was already judged to be superior to tape recording back in the 1970s and 80s, thanks to lower background noise, absence of 3rd harmonic distortion, no frequency response "bumps" introduced by the magnetic head, etc etc. If you think a vinyl record sounds good, wait until you hear a direct-to-disc record!
The LP record, thus, is a very high fidelity medium, at least if we ignore the surface noise problems. It can even reveal the sonic signature (imperfections?) of the early digital systems, when you listen to digital records mastered from digital recording systems in the 1970s.
Now, even if you were using Hi-res audio, other things that influence the quality of LP vs "digital streaming" needs to be mentioned:
1. Many times the audio system in your computer is resampling the audio without giving max consideration to sound quality. This can be configured in some systems. So your computer is probably degrading the audio quality. Make sure you can ensure a playback system where the digital file goes straight to the DAC. (Easier said than done...)
2. Are you sure you're getting hi-res audio? There are many cases of so called "hi-res" files that are just files upsampled from 44.1KHz/16bit files.
3. Implementation of digital filters not always is done in the optimal way, for example calculations are truncated due to the need to reduce the computing power used for this, and this can degrade the audio signal.
4. Then there is the other issue, the quality of the digitizing. Did they use the original master tapes? Were the master tapes in good shape? Was this done with care?
Chances are for, a 1950s-1970s recording, that the original LP record was made from tapes in better shape, and/or with greater care.
So, in short, there are many reasons for the LP to outperform digital streaming.
However, the LP shouldn't be able to do better than hi-res audio (i.e. uncompressed 192KHz/24bit) done properly, played through a good DAC. The latter has greater bandwidth, better impuse response, better phase correctness, etc etc.
Phono stages, even the cheapest ones, have already a smoother frequency response than all but the highest-end, highest-performance speakers.
They are also fine with THD. Phono cartridges show mostly only 2nd harmonic distortion, which has been shown to be inaudible even in gross levels like 10%. Never rely only on THD, the harmonic profile needs to be considered too.
It is true that low-end response on a turntable can be enhanced due to mechanical reasons. However most often what happens is the opposite: a weak bass response. And only the truly good turntable systems, with good tonearms and plinths, give you a good, strong, well-defined bass response.