Rodney gold
Im just me...
Even with the newest technologies and techniques, digital audio still cannot create exact replications of an original sound wave.
Thats not true , it can...that's why...
Thats not true , it can...that's why...
Which one is most accurate digital or vinyl?
- Thread starter kristoffer
- Start date
TwoSeventy
pfm Member
For me, much of the grinding 'flat earth' stuff is just dispiriting, the very idea that cd is 'perfect' is really not true. Anyway, I just cheered myself up listening to a radio interview featuring Michael Fremer telling a few home truths about the whole fiasco visited on recorded music.
As for the fashionable 'it's all in the mastering'..... What if both mediums are mastered to the same standard, will they still sound of equal quality? I've heard good CDs, but I have never heard a great one.
As for the fashionable 'it's all in the mastering'..... What if both mediums are mastered to the same standard, will they still sound of equal quality? I've heard good CDs, but I have never heard a great one.
adamdea
You are not a sound quality evaluation device
That is an excellent link for anyone who wants to misunderstand digital audio sufficiently to hold nonsense pseudo- technical beliefs which support their prejudices. Although a subscription to Absolute Sound works just as well.
Nic Robinson
Moderator
Absolutely. Utter nonsense.
Robert
Tapehead
No, in that situation the CD will always be better than any turntable if you judge 'quality' as likeness to the original master.
If you apply your own prejudices and sound preferences then the answer is - 'it depends on the listener'.
You clearly enjoy the wet balance of vinyl and therefore CD is never going to sound better than vinyl to you. All vinyl systems add a slew of changes to the sound but they primarily add euphonic distortions in the form of jitter (in the widest definition) and post echo. Humans tend to like these things in moderation and when you remove them complain of things sounding flat.
What intrigues me is why people run scared and are fearful of admitting that they prefer some added spice. It's almost as if the admission dents the audiophile credentials in some way. It's also bizarre that folk will embrace 'error' of sometimes gross proportions in the vinyl source while defending the position of banishing already inaudible tiny distortion in the digital domain to even great inaudibility.
It's all a bit bonkers really.
darrenyeats
pfm Member
Some misunderstanding about digital theory here. As Werner says, it's not intuitive but timing is not an issue for digital sampling. A series of samples can be offset an arbitrary amount of time (coincident with musical events or not) and they will always will encode the same unique waveform - remember the incoming analogue signal is band-limited before sampling.
Personally I've never heard a TT system sound great but I've heard CD systems sound great maybe 2-3 times. My opinion doesn't really prove anything, but the same goes for all of us, I suppose. Maybe my opinion will change in the future.
Personally I've never heard a TT system sound great but I've heard CD systems sound great maybe 2-3 times. My opinion doesn't really prove anything, but the same goes for all of us, I suppose. Maybe my opinion will change in the future.
You may think so but it is a misunderstanding of every salient point and completely wrong...
I don't know what the frequency used is, maybe different for different record labels, but stereo bass is too risky for LPs.
Since one "wall" of the groove is left channel and the other right channel if there was a loud bass sound in one channel the groove would cease to be continuous so the stylus would end up on the shiny bit.
Obvious if you think about it, so stereo bass has never been possible on LP, though it is on other analogue and digital storage methods.
Higher frequencies are OK since the amplitude of the modulation goes down with frequency (and is corrected, to variable levels of accuracy, by the RIAA equalisation curve). The high frequency limitation is the inability to cut high levels at high frequencies because of cutter limitation.
I like this analogy. It is like somebody who has only ever eaten chicken curry claiming unseasoned chicken doesn't taste like chicken.
Prejudices? Possibly, but one of the characteristics of the 'flat earthers' is an assumption that those who disagree are always 'prejudiced'....
I spent many thousands of hours in BBC recording studios, mainly using voice but many other sounds too. That time embraced analogue and digital, and a great many chances to compare the original source with the recorded output.
Digital is wonderfully convenient, but obviously more transparent to source? It was never that obvious to me. But then I must be addicted to colourations and incapable of rational judgement......
Easy way to make an argument isn't it...simply patronise those who have a different viewpoint.
david ellwood
Kirabosi Kognoscente
one of the things digital systems seem to find difficult is the resolution of bass notes.
it clearly isnt a resolution or sampling issue but is a more fundamental failing.
analogue source components are way ahead of the muddy indistinct mush that masquerades for bass in digital systems.
in this respect analogue is definitely more accurate.
it clearly isnt a resolution or sampling issue but is a more fundamental failing.
analogue source components are way ahead of the muddy indistinct mush that masquerades for bass in digital systems.
in this respect analogue is definitely more accurate.
RossB
Former Member
Please explain why. Shannon Nyquist is a mathematical theorem that is absolutely correct within its stated parameters - ie when sampling a bandwidth limited, continuous function. It is not a manual for how to design a DAC or how to reproduce music. Music is neither bandwidth limited nor a continuous function, and to sample it at any frequency requires filters of varying complexity at both the encoding and decoding stages. The quality of the filters determines, to some extent, the quality of the reproduction. However, the fact that real music does not naturally meet the conditions for the sampling theorem to apply is why digital sampling will always be an approximation of the original waveform. As approximations go, it is pretty good. But it is far from perfect - if it was there would only be one DAC chip, one encoding filter, one interpolation filter, and one perfect outcome. Instead, there are many, and they all sound a little different because they are all approximations (generally rather good ones) and not completely accurate reproductions of the original music event. Analog recordings are also not perfect, but no one is pretending otherwise.
Everything is bandwidth limited. All that varies is the bandwidth. Musical instruments are bandwidth limited by the materials and size of the components and dimensions of the sound generating parts thereof.
Record players are bandwidth limited at the low end by the way they work and at the high end by mass and geometry effects.
Tape recorders by head characteristics and tape speed.
So there is nothing new about restricted bandwidth. CD is restricted to a bandwidth lower than we hear in the bass to above the point at which the threshold of audibility and the threshold of pain intersect. Whilst this is not precisely the same for all people.
All music signals can be described as a sum of sine waves of various time dependant amplitudes and phases and are completely and in every way compatible with the requirements of accurate encoding and decoding digitally. Music does meet the conditions for the sampling theorem to apply.
Distortion added is vanishingly small.
Different reconstruction filters sound different in my small experience but only one produces an accurate reproduction of the original so, however much somebody likes the others, they are adding some form of coloration.
In terms of all DACs sounding the same my experience would suggest that well engineered ones do.
About 3 years ago I auditioned a range of DACs at home. I did not try any inexpensive ones though. All were properly engineered units, 5 units from 4 makers with a price range from £1100 to £11,000.
Carefully level matched it took hours of careful listening to discern any differences between them, and then only on certain types of sound and I am sure I could not consistently distinguish them blind.
I did end up with a preference but any of them were really good.
david ellwood
Kirabosi Kognoscente
The fifty thousand dollar question must then be why in the face
Of such damning objective measurement does it sound so wooly and indistinct.
RossB
Former Member
I'm not saying it's new. I'm pointing out - as you know - that digital sampling requires that you truncate the incoming signal using complex filtering and exclude some of the musical information. That is the only way you will meet the condition for limiting the bandwidth.
I don't entirely agree with this. As I mentioned in my earlier post, musical events are not just the sum of various sine waves. They are distinct (but often overlapping) events, usually starting with a fast transient sound followed by decay. This will not always - or even usually - appear as the sum of sine waves and is not the continuous function required by sampling theorem. Again, it close enough for an approximation, but only that.
I've got a few dacs here and most sound similar, some sound a little different, probably due to different output stages as well as different dac design. But that's not the point. Even if they did all sound the same, that does not mean that they are all accurate; it may just mean that they are all doing the same thing wrong equally.
I have not spent thousands of hours at the BBC, but was a keen amateur recordist from around 1962 to a few years ago. I started with a valve mono reel to reel recorder and have used various recorders, including cassette (Sony Walkman Pro, Nakamichi CR7E) DAT (Stelladat and a Pioneer DC88) R2R (Ferrograph but mainly Revox) and digital recorders direct to computer (Metric Halo MIO2 and LIO8)
Not all can do the equivalent of "off tape monitoring" but the Nakamichi Cassette, Ferrograph and Revox R2R and the MH stuff can.
Personally I could always tell the difference between the mike feed and the off tape sound of the cassettes and R2R machines. I could not with the digital, so IME the digital recorders have been transparent, or at least sufficiently so for somebody like myself used to the tape colourations not to notice anything.
Also don't forget that both the manufacturing of and method of playback necessary for LPs mean that they are nowhere near as good as R2R tape.
Well in my opinion no audible musical information is removed by the bandwidth limiting.
Whether you agree or not, any musical signal can be described by a sum of sine waves of varying frequency, amplitude and phase with time. The transients, start and decay, are correctly described by the fact that the amplitude and phase vary with time, that is what transients are. There is a limit to the maximum transient rate of rise (and decay of course) in any bandwidth limit, whether it is defined by the size of a double bass or the anti-aliasing filter in a digital recorder, but that limit is outside the limits of audibility of almost everybody apart from a few children.
IME the mike feed to a digital recorder and the output of the ADC/DAC are indistinguishable.
I know not many people have had the opportunity to experience this but I can assure you that, for me, it is. This is the reason, not only the theory that so few are mathematically equipped to understand, that convinced me, years ago, that as a good digital recorder is completely transparent.
I have heard it myself.
I am not speculating on whether the sound of my hifi seems realistic to me. I am stating that I can hear no difference between the input to a digital recorder and its output.
I know what I have heard. That is enough for me.
I expect there are probably only a tiny handful of people on here who have had the opportunity to make this comparison properly themselves.
Everybody else is speculating (sometimes wildly...)
RossB
Former Member
Acknowledging that there is some degree of approximation and some loss of information, but that it isn't (in your opinion) important, is very different from saying that digital sampling reproduces the original waveform perfectly.
I probably wasn't making myself clear. I certainly agree that music can be represented by a single waveform which sums all of the information. However, describing it as a sine wave is not entirely correct; we use this term as short hand because it is the easiest form of wave to describe and operate on. A sine wave of course is a continuous function and can easily be reproduced by any competent dac because it clearly meets the conditions of sampling theorem. However, the waveform of a piece of music is not a sine wave and it is not continuous (in the mathematical sense), and I think you would have trouble defining the function that describes it.
I don't doubt your experience or its validity. I don't even doubt your conclusion - I can readily accept that a digital recording is more transparent in almost all respects. However, I think there is at least one vital respect - ie timing information - which is not as readily apparent but which is important for musical reproduction and which digital recording does not reproduce as well. This minor aspect may be swamped by the accuracy of most other elements of digital recording, but it still prevents digital recordings from having that last element of "realness" that good analog recordings still possess, even if they sound less real in many other ways, such as dynamic range or noise floor.