Separate names with a comma.
Deep **** bruh
You are exactly correct.... the music is the signal.
However, it seems to suit some people better to mentally separate "the signal" (as a physical thing) from "the music (as a sort of abstract).
Likewise, in a DAC, the math on which digital audio reproduction is based says that the reconstruction filter is a necessary part of the process.
The way in which a DAC works causes spurious "stuff" to come into existence; the reconstruction filter removes that extra "stuff", and what remains is what you wanted - an accurate replica of the original content.
So, clearly, if you omit a proper reconstruction filter, extra junk remains that should have been removed, and your output will be LESS like the original than it should be.
(The spurious "stuff" is INHERENT in the process; ALL DACs generate it; it is not a flaw or an error; it is PROPER for the spurious images to be generated, and proper for them to then be removed by a filter.)
The fact is that, in simplest terms, what matters is that what comes out is as close to the original as possible - the actual details of how you get it that way really don't matter other than in terms of the final result they produce.
(But, obviously, if you don't do it correctly, then your results will not be accurate.)
However, in the audiophile world, and especially among audiophiles who don't fully understand the technology, there is an "intuitive belief" that, when it comes to circuitry or processing, "less is ALWAYS better".
They will insist, with some justification in "pure philosophy", that "the signal" is merely something that "gets you to the music itself"...
If you read the sales brochures on many "niche audiophile DACs", especially those that don't employ oversampling, you'll often find an inference that, compared to other DACs, "they are simpler and more elegant".
They will try to convince you that, while that nasty reconstruction filter produces an output signal that "specs better", it actually "gets between you and the music" and so you're better off without it.
The reality is that they are starting with extra spurious high frequency information on the output (just like all other DACs)...
Then, since no circuitry has infinite bandwidth, something is acting as a low-pass filter... and, if you want an accurate output, it should be designed very specifically to meet certain parameters.
However, instead of employing a carefully designed filter, that produces accurate results, they are relying on a transformer, or limitations in the circuit design itself, to "informally perform the filtering"....
The result is something that sounds interesting, and may even sound pleasing, but may not be very accurate, and which may sound very different depending on what load you connect it to.
Likewise, they will suggest that you ignore how well Delta-Sigma DACs WORK and focus on all the complicated processing they're inflicting on your precious music.
The implication is that, even though the output seems to be very accurate, all that complicated processing must have caused some subtle harm to the music - if only you listen carefully enough to hear it.
(You will often hear claims that seem to suggest that Delta-Sigma DACs are bad because they seem to deliver very good performance at low cost - as if "great performance really cheap" is an embarrassment or "there must be a catch".)
Similarly, even though DSD is a purely digital format, fans will claim that "it's more like analog" (Sony made that claim in the original marketing literature - because a picture of the digital DSD data "looks more like an analog waveform than PCM data").
A very popular claims was that, "unlike PCM, you can play a DSD signal through a simple capacitor filter and hear your music", or even that "you can play DSD straight to your speaker without a filter at all".
The reality is that, as with other DACs, if you do so, you have omitted the reconstruction filter, and are relying on your amplifier or speaker to act as a filter instead - with predictable poor results.
A similar mythos has built up around "bit-perfect".
Conceptually, it makes sense that devices like digital music players should be bit-perfect if at all possible.
Every time you allow something to alter the bits, you are allowing the signal to be changed; and music players and other devices have a long history of degrading the signal by changing it (think Windows player programs).
However, it's easy to get caught up in the hype, and forget that no DAC is bit perfect (every DAC outputs an analog audio signal - and so discards every single bit of the original digital signal).
Therefore, at that point, all that really matters is that the DAC produces an analog output that is as close as possible to what the digital input signal REPRESENTS.
For example, there's no promise that "an oversampling filter that accurately preserves the original samples" will deliver a more accurate analog output than "an oversampling filter that discards all the original samples and replaces them with all new ones".
(There's no specific promise that either option will produce an analog output that's closer to the original - because there are a huge number of other factors involved.)
To go back to the original phrase........
Many audiophile companies would prefer to convince you that "there's more to delivering the music than simply reproducing a signal"....
(Presumably, some of that "more" is supposed to account for the fact that "two DAcs that measure identical can sound different" and "even if two circuits produce the same identical output one can be 'better' than the other".)
How you look at that really depends on the context.....
In the philosophical sense I agree with you.
Music is "the experience" and the signal is "how the experience gets to your brain".
However,when we're discussing audio technology, the only part we're really talking about is the signal.
(So, when we're discussing audio equipment, we generally would say that "the job of the equipment is to deliver the signal to you as accurately as possible".)
And, yes, there is room for asking whether the goal is to reproduce the signal as accurately as possible, or to reproduce the experience as accurately as possible.
Would "a home theater system" be more accurate if it made your floor sticky and wafted the smell of popcorn into the air?
And, if you wanted to light up a Rembrandt, would it be more accurate to use lights that accurately displayed the colors of the paint, or to match the light in Rembrandt's studio?
However, in the specific context I'm referring to - which is "descriptions of audiophile gear".......
We often see various phrasing that suggests that the music is some sort of an intangible part of the signal.
As a rough analogy, they seem to believe that there is some sort of "soul" involved that is connected to but separate from "the signal".
For example, we hear of gear "that specs well but doesn't convey the music well"....
And we often hear descriptions of gear "conveying the emotion in the music" (or failing to do so) as if it were something separate from simply reproducing the signal accurately.
And we hear descriptions of things like "rhythm and pace".
The dictionary definitions of those words tell me unambiguously that, if the signal is playing at the correct speed, then the rhythm and pace must be correct.
However, audiophiles will insist that certain equipment can somehow get them wrong - even though acknowledging that the speed itself is dead on perfect.
(I'm inclined to suspect that, with claims about "rhythm and pace", some other characteristic, perhaps an odd frequency response or some form of distortion,
"makes the music sound less lively", which people then MISINTERPRET as affecting the actual rhythm or pace of the notes...
they're really claiming that "it sounds as if the rhythm is wrong", but simply not explaining their experience very precisely.)
As I mentioned above, there is a sort of grey area between conveying the signal and conveying the experience.
For example, a live concert may sound a certain way, but the recording you purchase may sound quite different.
Then, a particular piece of equipment may add a significant amount of second harmonic distortion.
And this may, at the same time, make the reproduction of the signal obviously LESS accurate, but make the subjective listening experience MORE like being at the actual concert.
So do we "rate" it based on the reproduction of the signal, or on the reproduction of the original experience?
My personal impression is that we should rate the accuracy with which the signal is reproduced... and leave it to the mastering engineer to get the signal right.
After all, that added distortion that makes a certain recording sound better, is equally likely to make another one sound worse.
But that is clearly a matter of personal preference.
@KeithEmo I completely agree. It isn't how we get there, but that we get there that matters. I couldn't give a hoot how complex, or how simple a topology is. As long as the signal/music reaches my ears and is perceived within the finite ability of my hearing brain as accurate and pleasant, I'm content. I also like your point that even if a technology may be objectively less accurate, if the user likes or prefers that result then the device is successful in bringing musical pleasure to life. Just like operating systems, I don't care Mac or Windows, blah, blah, blah, this is better than that, I don't care, if they both get me to where I want to be in essentially the same manner and time, the differences are just frosting, the cake is the results.
I don't take a position on DS, R2R, NOS DACs, I'm sure I would be able to find implementations of each type that I completely enjoy. I listen to music, what is behind the audio signal reproduction is transparent to me.
I also completely agree with your point that just because technology has given us the ability to produce capable DS chips inexpensively, that isn't an indication of inferiority or a scam. Nobody complains that the multi-core and advanced CPU architectures of today are by way of comparison much less expensive to produce than when the commercial activity began years ago. People just accept the massive performance improvements.
I think that between the experience of music by the brain and the right conveying of the signal by a dac, there is an area that is not touched often by engineers or audiophile here... My own experience is that ALL piece of gear interact with one another at a level that is not taken into account by many people.... And this is : first the very big difference that mechanical isolation, and second the electro-magnetic interaction and isolation between pieces of gear, finally and thirdly, last but not least, the cleaning of the electrical grid of the house...After my experiments I dont trust anymore any reviewer, and I think that the more important factors in audio are these 3 cleaning methods +room treatment …. My NOS dac does not have any of the limitations in my system that many reviewer wrote about... The reason is these cleaning methods in my audio room and the lack of these cleaning methods by some reviewers.... I am right with you about the marketing methods by some...Thanks for your posts...
Here's where I part company with your view - the WORK of an audio DAC is to deliver a signal which listeners are able to interpret as music. To take one example, ES9023 (a popular cooking S-D DAC with decent enough measurements) doesn't do that for me.
I take it that in your view the 'work' of a DAC is to deliver immaculate measurements?
Just out of curiosity, to your mind, if a DAC measures really well, what are the mechanisms by which it changes a signal to something different? If it measures well, what else should we go with to determine how it will do? And assume that the DAC chip that measures well is implemented properly. Where will it fail in terms of audio signal reproduction?
Clearly with DACs that don't deliver a signal that results in audible satisfaction (various S-D DACs I've tried do not) then there's something being added to the signal which wasn't in the original. My current hypothesis is that its noise being added, modulated by the signal. So I reckon we need to extend the current suite of measurements to include relevant tests for noise modulation.
To be fair, before we do that even we need to prove that noise modulation is somehow audible and discretely detectable by the brain, and if it is, at what point does it become audible. Measureable does not mean audible, people really over-estimate how sensitive the hearing system in the brain is. It is certainly amazing, but it is still limited. We can't assume something has an impact simply because it might. I am not saying that there is nothing to your point, but these are things that need objective data. Not trying to grind your gears mate so hopefully you don't take this as a personal attack, just a conversation among curious people who enjoy audio.
And we always need to remember the most important criterion for a good theory, if it isn't falsifiable/testable, it isn't a valid theory.
Yes, I agree, but they also under-estimate how sensitive we are to certain aspects of the 'music'. So for example, using Fletcher-Munson to estimate what's going to be audible and what's not, when those curves are built from the audibility thresholds of single tones, not music. To me, that's a huge leap of faith to assume single tone audibility thresholds should be transferable to music and needs detailed justification which so far I've not seen.
Some of the most promising work in audio perception comes from Al Bregman's 'Auditory Scene Analysis' but IME numbers guys seem strangely reluctant to enter into discussion of it.
As regards your comments on 'needing to prove', I disagree. 'Proof' exists in mathematics only, there's already evidence that some S-D DACs (I'm not extending the discussion to S-D in general) don't deliver the goods sonically.
Interesting post, and thanks for the thoughts. I think we'll have to agree to disagree on the need for evidence. While I enjoy the subjective very much, ultimately I know that for any actual phenomenon that is experienced subjectively, it is ultimately tied to biomechanical process systems in the brain and body. People are just biological machines, there is nothing going on that isn't subject to the laws of physics, audio perception is just another process, nothing more, nothing less. There may well be things happening that aren't currently measureable, but because we don't know if they exist at all, or if they do, how to measure them, it is useless to speculate about them in terms of how they impact the perception of sound.
And how do we determine what if any impact these, lets for now call them anomalies, have on music? Even if we determine that some "noise" artifact travels along with the rest of the signal, all the way, what if our brain never evolved the mechanisms to detect and interpret this noise because it is of no consequence? Even if something can be detected again, what is the sensitivity of the brain and what is the threshold of meaningful audibility?
I look at it this way, we know what we can hear, and clearly the VAST majority of a well reproduced audio signal is not noise, it is meaningful information. Even if there is a miniscule amount of noise that journeys along with it, are we to believe that this tiny, tiny, tiny little fraction of the sound for some reason has a big impact during decoding and perception? If so shouldn't everybody be sensitive to this noise? Our brains are going to be so similar that to assume a difference is not easy to support. And suppose we discover that a tiny fraction of the population is sensitive to this noise, should we call a technology that works flawlessly for the majority a problem? I think from a biological/evolutionary perspective the people with the sensitivity would by definition have the problem due to the rare sensitivity they possess. It is like a rare genetic disorder that makes a person sensitive to any sunlight. Would we say the sunlight is the problem when clearly the vast majority of people do not share that level of sun sensitivity? No, we would say that the person has a rare disorder, not that the sunlight didn't work properly. So given how prevalent DS DACs are, and that most people don't seem to find them to be an issue, I don't think we can simply assume that the technology is the issue. It might be, but we haven't established that. It would be more scientifically plausible to suggest that those that really are measurably able to detect a DS DAC (blind listening trials) have a sensitivity.
So far I'm really enjoying the AZG. I noticed that it is quite peculiar in that it is DS NOS DAC. All other DS DACs I'm aware of use upsampling/oversampling. It is also of note that the product that replaced AZG, the Platinum, implements DSD and is upsampling all input to its highest DSD reading capacity.
Dos anyone have more info about DS NOS? it seems like a very uncommon application.
Yep, no disagreement there though I'd say they're biological rather than biomechanical.
There we diverge, there isn't any evidence in support of this claim as far as I'm aware.
Yep - I guess we diverge here because you probably think we already know all the 'laws of physics' however in my view they're still a work in progress.
Incidentally if you're interested in exploring viewpoints which diverge from the mainstream 'man as machine' meme, you could do worse than turn up some of Robert Rosen's work.
As regards noise modulation in S-D converters, we do already know it exists, it shows up in some current measurements just we need more work to be able to identify it happening and quantify it when music is used as stimulus.
Not sure if you're implying this but...does this mean all people (biological machines) process audio the same -- assuming you control for their different hearing capabilities? If so, then wouldn't we all like the same headphone sound signatures? Seems to fly in the face of the various folks that love/hate the Grado/Beyer/Denon/Fostex/Focal/Audeze/HiFiMan/Senn/AKG/etc. sound.
Am thinking once it gets past the inner ear, there's a fair amount of variability to how it is 'processed' (interpreted) by our brains.