Technically this is the most intelligent post on the forum - I think no one will question it.
Why, oh why, does vinyl continue to blow away digital?
- Thread starter godofwealth
- Start date
- Status
Apropos salami;
This lady comes to the butcher and orders a quarter pound of salami, whereas the butcher asks her; sliced or whole?
She lifts her skirt; does this look like a cd- player!?
This lady comes to the butcher and orders a quarter pound of salami, whereas the butcher asks her; sliced or whole?
She lifts her skirt; does this look like a cd- player!?
As an encoding format and medium, above CD specs, digital encoding is way more continuous (both in amplitude and frequency) than any analog format. Just checking the numbers is easy.
Correlating those numbers back to aural quality and pleasure is not a topic where simplistic views survive.
If anything, I'd really like to put this notion in the past, the 'discontinuous' nature of digital vs 'continuous' analog. It is simple and sexy, a very attractive meme. But also quite useless, wrong, and just fossilized marketing terms that end up shaping a lot of our experience. For nothing. Good example of a facebook or youtube comment section grade argument.
Correlating those numbers back to aural quality and pleasure is not a topic where simplistic views survive.
If anything, I'd really like to put this notion in the past, the 'discontinuous' nature of digital vs 'continuous' analog. It is simple and sexy, a very attractive meme. But also quite useless, wrong, and just fossilized marketing terms that end up shaping a lot of our experience. For nothing. Good example of a facebook or youtube comment section grade argument.
Ok. So, we must consider that all your statements are purely subjective, even if you use technical words?
Wisely used technical measurements can be used to understand and guide our preference. And any one knowing about sound reproduction knows that is no "universal relative sonic goodness in stereo", particularly in the high-end.
Surely. Thanks for confirming that, as written, the thread tittle is absurd.
An interesting take on the discussion, talking about encoding format -- it's not clear to me that thus far we've been talking about encoding format.
Perhaps I am naive and not using the right vocabulary. I'll speculate that you are not denying that digital recording takes samples across time and these samples are time discrete finite 'units' of binary information. Is that correct?
I vaguely understand an "encoding format" as a description of the rules for organizing the binary information and there can be different formats.
edit: You refer to digital encoding and analog format. I'm not sure about what is an analog format or analog encoding. I think of a direct-to-disk audio recording where a 'cutter' etches the audio signal / electrical signal onto a master lacquer.
Last edited:
More or less the same situation/ discussion as when i joined the forum in 2012.
Are we going in circles
?If the sounds were recorded on separate digital tracks, then that salami was never a whole cow.
Correct. We are sampling the signal at a given frequency (time resolution), with a given bit depth (amplitude resolution). Of course. My challenge to you is to consider that an analog format is exactly the same. What about a tape is continuous? There are discrete suspended particles that are magnetized. It is all discrete, just very small. There is a reason why 15ips tape sounds better than 7.5. We have more time resolution, effectively because there are more samples (more particles per unit of time, literally more samples). The same for different width tape of magnetic solutions: you have more or less amplitude resolution (you could resolve more of the amplitude because you have more or different particles to do so). These things can (and should) be measured in the same way: sampling frequency and bit depth. These are not digital concepts, they are general information concepts. They are harder to apply because there is more variability in analog formats than digital (wow, flutter, medium quality, mechanical noise, etc). Our ears, eyes and fingertips also have bit depth for the various information they collect, encode and transmit. Much like analog, they are not a single value for the full spectrum of possible values, but that just makes things complicated (and not easy to memefy, or build catchy marketing around), nothing else.
That sounds like a good description. Notice it applies to tape, vinil, pcm, dsd, mp3, wax cilinders, alphabets, sheet music, ink drawings, html and speech in exactly the same way. You have a 'thing' you want to preserve, you come up with a format and encoding and decoding schemes. Usually the issue isn't the format, but flaws on either the encoding or the decoding processes that screw everything up. Korean isn't an inferior format to English. But your encoding/decoding capabilities for Korean probably render that format pretty useless as far as the objective is you communicating with others.
Of course there are lossy formats (most are...), that suck if you want to preserve the 'thing'. But we deal very nicely with approximations.
This might sound absurdly academic or pedantic, but if we want to discuss these things beyond simplistic terms, this is the effort we should make. Not falling for the simple and comfy explanation.
In this context, they are the same. Your definition was pretty nice: a format is rules + info. So is the cutter running along the lacquer. You are modulating the vibration amplitude with a voltage. You defined the reference voltage, a non-linear equalization curve along frequency, the direction of the vibration, the speed of the lacquer medium, etc. This is your encoding scheme. You'll need to know those to decode it (and the equipment to do it), otherwise it's just etches on a plastic disc. Just like a pcm file, or a bitstream in a usb cable. And just like those, you have a sampling frequency (you can't have a massless cantilever and needle, the lacquer can't be cut at infinitely large speeds) and a bit depth (you can't cut a polymer molecule down the middle, so you have a finite and well defined number of places that the needle can travel trough). This is what I mean by it is all the same if you dig enough.
The really cool part, and what I believe everyone is trying to discuss, is what happens after all of this is said and done, sound resulting from all of these steps and compromises hits our ears, and our brain goes 'ah!'. IMO, the format itself has such a limited bearing in this, that focusing our conversation there is just wasted entropy.
Examples of maybe more interesting topics for why we hear differences:
- identifying differences in mastering
- differences in gear type, topology, quality and design principles
- importance of rituals an expectations
- effect of noise and noise as dithering
- differences in fundamental distortion modes
- ...
science can break down the analog and digital processes and find the common parts, but the subjective consumption of the end product over hundreds of thousands of cases over decades makes it easy to hear the differences. and intuition points to the difference in where the loss is; the math/process involved in the digital conversion. whereas with analog no adc/dac math. even listening to slow tape, or live FM radio, compared to digital it's different. then we get to preferences. and why.
Last edited:
I have two questions:
- Why did you leave out the CD? Why don’t you include CD as well? How did you come to the conclusion that digital is way more continuous than analog above CD sampling rates?
- What is the sampling rate and bit depth equivalent of analog — 88/24, 96/24, etc.? How did you arrive at that conclusion?
This is cool
- Why do you think we need to 'find' the common parts? Didn't we build all of these processes as any other engineering topic? Don't we, by default, know what these things do (from a mechanicist pov)?
- Given the complexity of both of these processes how confident should we be on our intuitions when discussing details?
The subjective consumption over decades creates listening habits that create preferences, nothing else. Younger people that are not "trained" to the analog distortions will find them more easily. There are technical reasons why old people with more limited hearing abilities prefer the old recordings, recorded and mastered to complement analog recording.
Well, proper use of intuition supposes that we use knowledge to confirm our findings. My intuition says that current top digital can be more transparent than analog vinyl and the technical arguments confirm it.
Surely. But FM radio is sometimes digital since long.
Why what, sorry?
1. redbook is cutting it very close to even the limits of what we should be able to hear, so I just use it as a cut-off point. It is just useful short hand, of course, please give me some headroom here.
2. at what frequency? One of the characteristics of most of our analog systems are that very few things are constant or even linear with frequency. this is very true for bit depth, shorter interval for sampling rate.
I'm actively trying to not generalize. I'm arguing these things are more complicated than simplistic arguments make them sound.
I can’t believe I’m reading this — it’s completely illogical. How do you know that CD is very close to the limits of human hearing in terms of resolution? How can you be so sure, how can anyone be sure that CD actually delivers that level of resolution? This discussion has been about resolution from the beginning — and it still is.
Please stop and read what you’ve written.
Then please read your prior post below
This sounds pretty general to me.
Last edited:
I'm taking this conversation exclusively in good faith. I don't have a general preference, I don't have anything to defend, except the nature of the discussion itself.
16bit? A nice ball park for our hearing dynamic range is 120dB, in ideal conditions. That translates to about 20 bits. With 16 you get in about 96dB, which is enough to cover most human experience, especially in less than ideal conditions (that include perfect hearing, low ambient noise, trained subject, warble test tones, etc).
So is it OK to say CD Redbook format is cutting it close? Just that, not that it is a bad, flawed, inferior, whatever format.
So you see both things are different, but the operating principles are just the same, if you dig deep enough. How could they not be?
I didn't make a conjecture, I just stated (loosely, I know) general knowledge. How close to the limits are we? Well, let's take the familiar 20Hz to 20kHz interval. With 44.1kHz sampling rate you have just 2.05Khz headroom there, where you usually want to implement low pass filtering. That is cutting it close. We usually don't hear up to 20kHz linearly, but even if we go for 15kHz (I reach 15@0dB and 18@-4dB in my left ear and -3dB in my right, at the moment), we don't even have an octave of available space to operate.
16bit? A nice ball park for our hearing dynamic range is 120dB, in ideal conditions. That translates to about 20 bits. With 16 you get in about 96dB, which is enough to cover most human experience, especially in less than ideal conditions (that include perfect hearing, low ambient noise, trained subject, warble test tones, etc).
So is it OK to say CD Redbook format is cutting it close? Just that, not that it is a bad, flawed, inferior, whatever format.
What do you mean? A CD adheres to a format and performance specification. It wasn't found at the bottom of a pit, and we're still trying to figure out what it does. It was literally designed and built to encode and deliver at a specific resolution.
What do you mean by generalizing analog mediums using specified pcm resolution? I agree they are different things, but the measuring units stay the same. Dynamic range is expressed in dBs or Bits (it is the same, we just use them differently in different contexts), and time resolution is expressed in Hz. These are not pcm terms, they are valid for everything. Like weight or speed of a car. It doesn't change with the car color. Also the digital speed meter on your car has a certain bit depth (probably 8 bits, so it can go from 0 to 256 in whole numbers, I dunno) and a sampling rate (probably 1Hz or so). It is not sound, but the terms are the same. On your older car, with the analog meter you had the same. There was either an induced voltage or tension on a spring with a rotating magnet somewhere on the drive train (so your sampling rate varies with speed, from 0Hz in stationary to some Hz in motion). The stiffness and mass of the spring or the inductor characteristics, together with the needle bearing friction means you have a bit depth, as they can't react to infinitely small speed variations.
So you see both things are different, but the operating principles are just the same, if you dig deep enough. How could they not be?
Thank you for the Digital Audio 101 lecture, but I don’t need it. In my opinion, it’s been completely unnecessary—not just for me, but for almost everyone here.
When you bring up basic, primitive information about digital audio as if it has anything to do with musical engagement and focused on technical specifications, I don't know how to explain why people prefer analog's continuous capturing to digital's quantized conversion.
I didn’t intend to lecture, but I like to discuss things under well defined terms. You made questions and I answered. You made statements and I answered as best as I can.
From what I wrote you could read that I’m writing these precisely to try and convince others that these differences are both artificial and don’t have much to do with musical engagement.
Following what I’ve been writing, it should be natural that you can’t explain that preference with this type of classifications, because they are mostly unrelated imo. But of course you’re free to keep doing the same thing over and over and expect a different outcome. I’ll be doing something else though.
- Status
 | Steve Williams Site Founder | Site Owner | Administrator |  | Ron Resnick Site Owner | Administrator |  | Julian (The Fixer) Website Build | Marketing Managersing |