or Connect
Head-Fi.org › Forums › Equipment Forums › Dedicated Source Components › Confused: Vinyl(Analog) vs Digital
New Posts  All Forums:Forum Nav:

# Confused: Vinyl(Analog) vs Digital - Page 2

Quote:
Originally Posted by GreenLeo

No DAC can reconstruct the original analog signal in all cases.  The DAC, like upsamping algorithms, can only guess it.  You can never use upsampling or increase the sampling rate to change a discrete set of numbers (PCM format) to a continuous graph (the analogue signal).  This is basic Maths.

Practically, when the sampling rate is large enough, the difference between the original analogue signal and the reconstructed signal will sound identical to the ears.  Hence going 24/96 recording should be better than the 16/44 recording.  But I won't say the same thing for the 24/352 because the law of diminishing return always come in.  Probably 24/192 is the sweet point - a wild guess of me.

you are 100% incorrect.  the nyquist-shannon sampling theorem MATHEMATICALLY PROVES that:  "If a function x(t) contains no frequencies higher than B hertz, it is completely determined by giving its ordinates at a series of points spaced $\tfrac{1}{2B}$ seconds apart."  This is not "basic" math, but very rigorously proven information theory.

since a CD is sampled at 44.1khz, the analog signal will be 100% perfectly reproduced between 0 and 20khz

Quote:
Originally Posted by ssportclay

I have yet to hear a digital source that sounds analog which is the way we all hear. I like the way the compact disc is more quiet and dynamic than vinyl which makes digital a fun listen in many respects. The problem lies in the reconstructed digital signal itself. There is always just something phony and synthetic about it that never escapes me. Its like being served fake butter or instant coffee. Technology may prove to iron the bugs out in the future but currently digital is incapable of providing real music as a source. Digital no doubt has much better specifications than vinyl. No matter how you measure them, digital will always come out on top by a wide margin. I have tried very hard to prefer digital over vinyl over the years because of this superiority but it just sounds fake and synthetic. I also buy real butter and brew my coffee with the help of a grinder.

What sources have you heard? I agree a lot of digital sources offer a synthetic feel. Or some sources get some aspects right, but fail miserably on others. But as I previously stated, I have found some sources that aren't so. Have you tried any belt drive transports o non oversampling dacs? I''m curious as to what those may offer. Yet I beg to differ that the majority of the problem lies in the way the original cd is recorded. Too much compression is often used, strings don't have that silky smooth quality that vinyl does, etc... Try listening to a vinyl rip on cd, then listen to the original cd. There's a huge difference.

Quote:
Originally Posted by GreenLeo

No DAC can reconstruct the original analog signal in all cases.  The DAC, like upsamping algorithms, can only guess it.  You can never use upsampling or increase the sampling rate to change a discrete set of numbers (PCM format) to a continuous graph (the analogue signal).  This is basic Maths.

Practically, when the sampling rate is large enough, the difference between the original analogue signal and the reconstructed signal will sound identical to the ears.  Hence going 24/96 recording should be better than the 16/44 recording.  But I won't say the same thing for the 24/352 because the law of diminishing return always come in.  Probably 24/192 is the sweet point - a wild guess of me.

I'm not a vinyl guy and I can't afford it.  However, I must admit that the best SQ that I've heard was a vinyl system.  I still remember that one day I was hearing two system playing the same music: the CD sounds good and I knew that I was in a party, but when the vinyl playing I was in the party.

Dacs were not made to reconstruct an original analog signal. They were made to construct an analog signal from a digital signal.

Quote:
Originally Posted by computerparts

Dacs were not made to reconstruct an original analog signal. They were made to construct an analog signal from a digital signal.

If the analog signal constructed from a digital signal is very different from the original analog signal, what's the point of using the DAC?  We find the timbre of an instrument produced by any other means other than the real instrument (say CD) different from the sound of the real instrument means these two analog signals are different.

I'll give you another argument here.  If the point of using a DAC is just to construct an analog signal from a digital source, then there will be no point to record in 16/44 -- a 8/4 will do.  Some sound will be produced for sure but any original analog signal presented above 2k will be lost.

The existence of some DAC that sounds realer and some don't implies that not every DAC is equal.  The ultimate aim of a DAC should be to reproduce the original signal from the discrete sample (the 16/44 data or whatsoever discrete data) but this aim cannot be 100% achieved.  However, to construct the signal up to a point that it's indistinguishable to human ears should be attainable.

Quote:
Originally Posted by El_Doug

you are 100% incorrect.  the nyquist-shannon sampling theorem MATHEMATICALLY PROVES that:  "If a function x(t) contains no frequencies higher than B hertz, it is completely determined by giving its ordinates at a series of points spaced $\tfrac{1}{2B}$ seconds apart."  This is not "basic" math, but very rigorously proven information theory.

since a CD is sampled at 44.1khz, the analog signal will be 100% perfectly reproduced between 0 and 20khz

I know the Nyquist theorem.  Can detect a 20khz is not equal to to be able to reproduce it faithfully.

A 40khz sampling cannot always reproduce a original 20khz signal .  Just consider a sine wave with frequency 20khz with amplitude 1V.  V(t) will be the voltage of this signal at time t.

Now V(t) = sin(2Pi * t*20k).  For a 40k sampling rate means we can sample at 1/40k, 2/40k, 3/40k , ... etc and V(1/40k) = V(2/40k) = V(3/40k) = ... = 0.  What we have is actually a flat line.  Anyway this highly unlikely that we are sampling exactly at 1/40k, 2/40k, 3/40k.  It's only if one is lucky enough to sample at 0.5/40k, 1.5/40k, 2.5/40k, ...,  then will the 1V amplitude be rediscovered.  Again it's  highly unlikely that we are sampling exactly at 0.5/40k, 1.5/40k, 2.5/40k, ....  Hence this 1V in general cannot be detected just like in general we don't get a flat line.  This are all basic maths and illustrate the point that " Can detect a 20khz is not equal to to be able to reproduce it faithfully"

This is a well known limitation of the Nyquist theorem and hence why Sony Philips use the 44.1k sampling rate instead of the 40k.  A 2f sampling rate can reconstruct any signal that's up to f in frequency is often a misquoted fact.

@GreenLeo: Thanks for writing that up...I was about to post a similar critique of what El_Doug had posted.

Nyquist theorm is only to detect a signal and not to reproduce.

Reproduction theoretically is dependent on the value of the sample and the accuracy of the time at which the sample was taken. And as we all know there are numerous things like jitter in the entire recording to reproduction that can throw the sound quality out.

Having said that vinyl is not without fault. To be able to reproduct the exact sound of a Cymbal, requires the medium to capture not just the frequency but also the amplitude. The needle is not the perfect transducer (zero weight and infinite stiffness) hence its not accurate in reproducing sound as well.

As the current technology stands an accurate digital system is much more likely than an accurate vinyl system. Now musicality is a different thing because for human hearing a fuller and more comfortable sound is more musical.

Quote:
Originally Posted by GreenLeo

A 40khz sampling cannot always reproduce a original 20khz signal .  Just consider a sine wave with frequency 20khz with amplitude 1V.  V(t) will be the voltage of this signal at time t.

Now V(t) = sin(2Pi * t*20k).  For a 40k sampling rate means we can sample at 1/40k, 2/40k, 3/40k , ... etc and V(1/40k) = V(2/40k) = V(3/40k) = ... = 0.  What we have is actually a flat line.  Anyway this highly unlikely that we are sampling exactly at 1/40k, 2/40k, 3/40k.  It's only if one is lucky enough to sample at 0.5/40k, 1.5/40k, 2.5/40k, ...,  then will the 1V amplitude be rediscovered.  Again it's  highly unlikely that we are sampling exactly at 0.5/40k, 1.5/40k, 2.5/40k, ....  Hence this 1V in general cannot be detected just like in general we don't get a flat line.  This are all basic maths and illustrate the point that " Can detect a 20khz is not equal to to be able to reproduce it faithfully"

This is a well known limitation of the Nyquist theorem and hence why Sony Philips use the 44.1k sampling rate instead of the 40k.  A 2f sampling rate can reconstruct any signal that's up to f in frequency is often a misquoted fact.

You are absolutely correct - and this is why we do not expect a CD to reproduce 22.05khz correctly.  However 44.1khz is more than enough to have a critical frequency above 20khz!

Not to mention that even as a young man of age 24, I struggle to hear anything over 18.5k (and my audiologist told me that this is an incredible feat in and of itself).

Quote:
Originally Posted by GreenLeo

I know the Nyquist theorem.  Can detect a 20khz is not equal to to be able to reproduce it faithfully.

This is also 100% true; I was simply rebutting your incorrect claim about mathematical proofs regarding sampling.  I do not believe anyone doubts the practical limitations of signal reconstruction

Edited by El_Doug - 6/9/10 at 11:18am

Quote:
Originally Posted by computerparts

What sources have you heard? I agree a lot of digital sources offer a synthetic feel. Or some sources get some aspects right, but fail miserably on others. But as I previously stated, I have found some sources that aren't so. Have you tried any belt drive transports o non oversampling dacs? I''m curious as to what those may offer. Yet I beg to differ that the majority of the problem lies in the way the original cd is recorded. Too much compression is often used, strings don't have that silky smooth quality that vinyl does, etc... Try listening to a vinyl rip on cd, then listen to the original cd. There's a huge difference.

I am 56 years old and I have gone through many machines trying to get CDs to reproduce high-fidelity. It seems that everyone is looking for the next great DAC in an attempt to get their sound straightened out. Digital always sounds digital even though specifications approach perfection and theory seems without fault. Vinyl from analog masters never sounds digital. It isn't perfect either but as a whole, it will show off a very good system far greater than any digital source due to it's superior sound quality. I actually listen to far more digital throughout a typical day. Digital is technically superior but it's inferior sound quality is it's only weakness.

It depends on the recording, if you hear mainly stuff fully made on a computer then digital will sound always better, if you hear analog recordings with a top notch vinyl rig then this will sound better. Personally i like the analogue vinyl sound wich is more weightful, blacker, just more real music and not so thin as digital in comparison regarding true music. To get near this analogue feeling  with digital there´re are ways like NOS DAC´s with good jitter rejection and ripped vinyl in hires.

There are people who try to match their ultra high end vinyl rigs with digital and shop the best of the best DAC´s and just can´t match their vinyl rigs. Check out the Greece community

http://aca.gr/event08-9.htm

Edited by Fidelity Castro - 6/9/10 at 4:05pm
Quote:
Originally Posted by El_Doug

...

I was simply rebutting your incorrect claim about mathematical proofs regarding sampling.  I do not believe anyone doubts the practical limitations of signal reconstruction

Thanks for your kind intention.  Please show me where was my  incorrect claim about mathematical proofs regarding sampling.

Also, please read again your post (#16) which stated that  "since a CD is sampled at 44.1khz, the analog signal will be 100% perfectly reproduced between 0 and 20khz".

My simple example posted in post 20 has illustrated that the quoted statement is simply not true.

Quote:
Originally Posted by CrazyRay

Oh man....that made me squirt coke out my nose! LOL!

I could post a nice paragraph on the topic of digital v. analogue but I am laughing too much. However, I will sum it up with the following:

IT'S ALL IN THE MASTERING!!!

It's interesting how fast this thread went south haha.

Well if we're going to invoke the math, let's get real.  The entire model of sound as the fourier transform of the amplitude-time curve is only an approximation to the physical universe in which we live and listen.

Here is the (well-known) argument that supports this:

Music has zero amplitude before it begins, and after it ends.  The time extant of the signal is finite.  The fourier transform of a square wave has infinite support at all frequencies, thus recorded music has no upper limit in the frequency domain.

Or if it does, then -- using the same logic -- the music has been playing for all time, and will extend for all time.

Both of these conclusions are obviously absurd -- music starts and stops, and is of limited frequency.  What is wrong ... this is a paradox?

Not really -- the error is the model.

Mathematicians have "solved" this paradox by deriving families of curves with transforms (like amplitude-frequency pairs) that avoid this bizarre problem ... but they don't yield any insight into audio engineering so that no one uses them.

For those playing along at home, this "paradox" I believe was first pointed out by Fuchs in 1954 and the solution, prolate spheriod wave function analysis, was developed at Bell Labs and published by Landau and Pollak in 1961.  I actually met Pollack when I worked at Bell Labs.

Who cares?  Not me anymore!

Listen to what you like the sound of.  No reproduction method will match what you hear in your headphones with what you would have heard live in the studio or hall.  Pick the inaccuracies that please you the most, and make you feel the way you would if you listened "live", or perhaps feel better.  If the music never existed in the analog domain, then there is no standard to even compare it to -- everyone will hear it differently.  Listen to alternatives, and pick the one you like.  There are many pop and rock tunes where I strongly prefer the 5.1 ripped from the DVD-A mixed down by SOX to stereo, 24/96, played thru FB2K, USB-to-SPDIF converter, and my DAC than I do to the audiophile vinyl version played back on my co-workers \$114,000 turntable rig.

Most likely, as LFF just said, it's all in the mastering.  But you can decide for yourself by doing something very pleasurable -- listening.

Screw CD’s and Vinyl.
Go directly to the source!!!!

Edited by CrazyRay - 6/10/10 at 10:59am
New Posts  All Forums:Forum Nav:
Return Home
Back to Forum: Dedicated Source Components
• Confused: Vinyl(Analog) vs Digital