While I agree wholeheartedly with the idea behind that catchphrase - it must be taken in context.
For example, if we're talking about your property, or the borders of a nation, it's true that "the map is not literally the territory".
However, if you ask the lawyers, the map absolutely
DOES represent the territory.
If you want to know where your property line lies, the map will be considered to define it, and arguments that "the reality is right and the map is wrong" will probably not get much consideration.
And, you'll notice, people quite often mysteriously transform from citizens of one country to citizens of another, or from living in one voting district to another, when the maps get adjusted.
In those contexts, the map IS legally defined as the reality, and it's the physical "reality" that must comply.
Also note that, if you want to compare a vinyl album to a digital recording, you aren't anywhere
NEAR the actual territory yet anyway.
In fact, both that vinyl album and that digital audio file are really just
DIFFERENT MAPS of the original territory.
In fact, they're just different versions of copies of another map... the "master recording".
So, in point of fact, it would be more accurate to consider the vinyl recording and the digital recording as an analog photograph and a digital scan of the same hand drawn map.
And, depending on how you feel about it, the "master recording" isn't that original territory either; the original territory is the original real live performance.
The master recording is just the original hand drawn map.
But what if the "master recording" was put together out of separate tracks in the studio and there
IS no "original performance".
Well, then, your vinyl and digital recordings are just copies of a
FICTIONAL original hand drawn map; and the "actual territory" exists only in the mind of the recording engineer.
However, the differences you notice probably come down to
FAR more mundane things.
1)
ALL sample rate conversions involve some filtering. So, if you convert a 24/192k file to 24/96k, they are in fact slightly different in other ways besides their sample rates. Whenever you do a conversion, filtering is applied, and that filtering is different in each program you might choose to use. Some programs simply don't seem to do a very good job (when you run a test signal through them, their output includes obvious junk that shouldn't be there). And most of the more powerful converter programs offer multiple options and adjustments. (Converting a file from 24/192k to 24/96k will produce a slightly different result if you choose a filter with "a cutoff of 22 kHz and a sharpness of 1.2" than if you choose one with "a cutoff of 24 kHz and a sharpness of 1.6".) Even worse, the programs can't even agree on which options they offer, or what to call them (so you usually can't pick the exact same settings on different programs). Starting with a 24/192k "original", then resampling it to 44k and 96k using the same program, will probably get you closer; however, even then, the results won't be identical. The upshot of this is that you can
NEVER make an
EXACT comparison.
2) The internal filtering applied, and lots of other things inside a DAC, may also be different at different sample rates. For a given DAC, things like the filter slope and the oversampling rate will be different depending on the incoming sample rate. In other words, it actually is possible for a DAC itself to sound different with different sample rates (and a DAC could be optimized to sound better at one particular sample rate). To pick a few examples, an aggressive reconstuction filter could render any potential improvements from higher sample rates moot if it filters them out anyway; likewise, a specific DAC could have been tweaked to offer an excellent frequency response at 44k, at the expense of a less accurate frequency response at 96k. (That last sort of tweaking is much more common with niche products that encourage you to make "subjective" decisions; rather than design their DAC to simply be accurate, they've adjusted it to sound the way they want it to under specific conditions.)
3) In most cases you don't control the source. If you purchase a 24/192k version and a 24/96k version of the same album from your favorite vendor, odds are that they produced the album at 192k, then resampled that "original" to make the 96k version. This means that the 96k version will have gone through an additional conversion, and an additional stage of filtering, which may well audibly affect the sound. You may like the result better or worse, but you must assume they are different in ways
NOT specifically related to their differing sample rates. (This is the same problem you have trying to audition an MQA track; you can compare how the MQA track sounds with and without being decoded, but you have no access to the original
UNENCODED version of the exact same version to compare.)
4) Also, if you bought both versions from a
VENDOR, it's
NOT a given that they are "as close as possible to identical" at all. They may have put extra care into making the 24/192k version sound the best, because it's their most expensive version of the product, so they care more about it. In fact, they may have deliberately made the 24/96k version sound worse. This could have happened in a deliberate attempt to ensure that you would notice the difference, or because the recording engineer who did the final adjustments figured that "the audiophile version" should be mastered a tiny bit differently to appeal to a slightly different audience, or simply because they did a final manual adjustment on the top of the line 24/192k version, but didn't bother to do so with the cheaper 24/96k version. This is a common problem when attempting to compare CD and SACD versions of an album that are supposedly from the same master. In reality they usually do sound different, and they usually have been adjusted at the final production mastering stage to deliberately sound different to appeal to different audiences. To put it bluntly, a good mastering engineer who knows that he or she is mastering for SACD, or for vinyl for that matter, is going to make subtle alterations to the sound to make it more appealing to the target audience they expect those particular discs to appeal to. This means that they often
WILL "punch up the treble" on the DVD-A version, because people expect more treble on that high-res file, and "make the SACD version nice and smooth" because that's what they figure "audiophiles" expect to hear from an SACD.
Please note that not all of this is some sort of trickery. The goal of a really good recording or mastering engineer is to deliver a product that sounds good to their audience, and that probably does
NOT mean doing their best to make sure that the CD and SACD version sound as much exactly the same as possible. (It's not
THEIR problem that this makes it difficult for
US to compare the various formats. In fact, they're just offering more choices; if CDs and SACDs sounded exactly the same, then there would be no reason to sell both, right? )
5) Last, but not least, you have to consider the quality of the master, and whether you're trying to reproduce it as is or make it better. Some masters really weren't that good to begin with, and so making a higher quality and more accurate copy of them isn't necessarily an improvement. If you want to make a Doors album sound good, then you probably have to change it significantly, because the masters themselves were not especially good. (And, of course, with every remaster that sounds different, some people like it and others don't). The 24/192k remasters of the Grateful Dead studio albums sound fantastic... at least to me... and significantly different than the original releases. (And there's no way to know how much of that's due to the sample rate and how much is due to the fabulous remastering job they did; probably mostly the latter.) Other 24/192k remasters sound so close to the original release it's difficult to tell which is which. I always tell anyone who asks whether "the 24/192k remaster is better" to read the reviews and treat it just like any other remaster; decide about each based on reviews and the specific merits of
THAT RECORDING AND MASTER. Don't assume that it will be better
JUST BECAUSE it's a high-res file.
Finally (and I didn't think this rated a number), in response to one of drtrey3's comments, recording in 24 bit gives you
OTHER benefits besides those strictly associated with storing audio at 24 bits.
(In actual fact, the difference between the same file recorded "perfectly optimally" at 16 bits or 24 bits would probably be totally inaudible.)
HOWEVER, if you're recording something, using 24 bits gives you
MUCH wider safety margins in terms of recording level.
(You get to keep your record levels a little lower, and so avoid clipping and other distortions, without getting too close to the noise floor.)
This makes it a lot easier to get good sounding recordings from vinyl rips
AND live performances.
I think I can hear the diff between my 192K and 48/96K files. The difference is small, and to me, not worth the premium. YMMV. But the difference between a 24 bit rip and a 16 bit rip of a wonderful old slab of vinyl is obvious. From another room obvious. It does not have anything to do with dynamic range, but there it is. So now I rip vinyl at 48/24 and am quite happy with the results.
The original article reminds me of a true saying: The map is not the territory.
Trey
