Which format is closest to the master: vinyl, CD, DVD-A, or SACD? - Page 4

The sample shown is a 1-kHz square-wave signal, thus absolutely in the audible range. The ringing, however, happens in the ultrasonic range, at 22 kHz. Because this is the resonance frequency of the low-pass filter. However, every frequency below is also affected by the filter-induced tendency towards delayed decay -- the closer to the filter's corner frequency, the more pronounced.

To give an illustration, based on the above-mentioned amplitude modulation, which is the normal response with the CD format before any low-pass filtering. You must know that CD players and DACs with conventional anti-aliasing filters produce a flat frequency response up to 20 kHz without the amplitude modulation. How this? Well, that close below the filter resonance the resonance effect is still quite strong, so strong that it equalizes the amplitude differences, because of the heavy delayed-decay tendency. Now, in an imagined scenario where the original signal in fact does have such an amplitude-modulated shape, a CD player with this filter characteristic can't reproduce it properly -- it makes a continuous sine wave out of it.

In the context of the virtual test configuration: not if the reference point is the reference tape recorder's line out, not the tape itself.

Good point! I own a few albums on both LP and CD, all recorded digitally. On vinyl they sound clearly more analog and organic -- just like vinyl is thought to sound. Which means there's some coloration in play. Whereas the CD should be an 1:1 copy of the master tape.

drminky...

...good thoughts! It has indeed to be considered that many modern recordings are made with high-res PCM, which vinyl records will benefit from, unlike CDs. Nevertheless, I doubt that the majority of the available digitally recorded LPs have such a background. I don't think it's the case with my references:

Alexander Glasunov - Symphonies, 3 albums, Orfeo, 1984
Alban Berg - Violin Concerto/Three Orchestral Pieces, Philips, 1984
Arthur Honegger - Orchestral Works/Symphonies, 3 albums, Erato, 1986
The Nits - Henk, CBS, 1986

All these recordings seem to sound «better» from vinyl in some ways, at least more organic and natural than the corresponding CD version.

I'm interested to know which format is closest to the original sound -- in the perception of humans, not measuring instruments. For this purpose comments such as

- 2" analog tape is superb
- vinyl is beautiful
- DSD (SACD) mostly sounds very good
- PCM at 24 bits (DVD Audio) sounds very good if it is WELL CLOCKED

give not the least clue. Also because these people -- whose trustworthiness I don't challenge -- are just valuing the degree of euphony here, not the degree of deviation from the original.

Isn't this the primary goal of every audiophile: a sound as accurate and neutral as possible, at least a sound with as little flaws as possible? So isn't it interesting to know which audio format comes closest to this goal? If you have to deal with an audio format with inherent flaws, there's the need for compensating (or forgiving) gear from the very beginning, with a high likelihood that this won't be the theoretically best and most accurate gear. I'm an absolute believer in synergies -- another word for compensating effects: flaws compensating for the flaws of the other components in the chain. Wouldn't it be great to know that you can at least rely on what the source format offers?

As to the sonic beauty of vinyl: nothing against liking or even preferring it to digital. But I'm rather sure that it's wrong to equate euphony with fidelity. My own experiences with audio devices have taught me otherwise. Renouncing the preamp and replacing it by a simple attenuator was a sobering experience. Luckily, as a speaker builder, I could mess around with the tuning of my speaker's crossover network and once the sweet spot was found to discover that the setup sounds better than ever. A highly accurate source signal seems to be much more critical in view of the overall sonic balance in its finest facets than a euphonically or otherwise colored signal, such as from the radio or the turntable. Another experiment was renouncing the headphone amp and connecting the HD 600/650 directly to the DAC's line out. The resulting sound was so revealing that I had troubles to accept all the hard edges and the lack of warmth in the beginning. Changing the DAC and finally the cables brought the ultimate revelation: accuracy doesn't mean beauty, but the beauty can absolutely be in the accuracy once you're ready to take the plunge and have the right, synergetic equipment.

Of course there's room left for speculation why digital -- provided that it's indeed more accurate -- is more critical. One reason could be that it reveals all the flaws in a recording caused by the recording equipment (not to speak of bad microphone placement/mixing/equalizing, etc.). I'm sure the microphones used today are still far from being perfect, from delivering a perfectly pure electrical equivalent of the acoustic event, just to name the most obvious issue. The other possible perspective, which doesn't exclude the aforementioned, is the imperfection of the CD format, making the sound heard through highly revealing equipment less favorable than through more forgiving and less transparent gear. I have my own assumptions about the cause for the often lamented «digital» sound, especially in the context of low-res PCM. Apart from the plausible flaws in the players and DACs, such as jitter, converter nonlinearity and quantization noise, the main reason could be the sharp low-pass filter at 20 kHz -- independent of its implementation (analog brickwall filter, digital up-/oversampling filter) and to a lesser degree also valid for time-optimized and filterless designs due to the fact that there has to be filtering before A/D conversion anyway.

I've already explained a few posts earlier what I think sharp low-pass filters do to the sound. But why doesn't e.g. FM radio have a «digital» timbre as well, considering it also suffers from abrupt low-pass filtering? And why can tube-based amplification stages eliminate it to a certain degree? Based on the fact that the sharp low-pass filter creates a resonance with side effects reaching deep into the audible range, the main audible effect is that of smeared transients in the treble -- giving the impression of coolness and glossiness/glassiness. This phenomenon can be compensated by a relatively high amount of harmonic distortion, virtually «restoring» some of transient sharpness in the treble by adding sharp-edged components otherwise completely missing (this can also apply to vinyl in the case of a digital master, BTW). I hope my way of portraying the scenario is passably understandable.

The input signal before recording/encoding to the storage format. Simple as that. I don't care for the technical or musical quality of the recording at this point. Hence the analog master tape as the sole reference (= the original), or more precisely the signal out of the tape recorder's line out. We could discuss the correct way of recording to preserve the intention of the composer, the musicians or the recording engineer or whatsoever, but that's beyond my competence. I'm just interested in the technical quality of the storage format itself in its consequence for the perceived sonic characteristic. I'm rather sure even you wouldn't accept MP3 at 64 kb/s or cassette tape as storage format because of the missing accuracy.


You know, synergy means reciprocal compensation for flaws in a reproduction chain, with the goal to achieve something like an approximated restoration of accuracy -- or what else? Of course I want «musicality» too in the end, but I'm convinced to find it in a reproduction as neutral as possible, not in euphonic coloration. Or do you think an undistorted violin or piano doesn't sound musical (enough)?