Vinyl vs DVD-A vs CD-A vs B.S.?

Discussion in 'Mastering' started by Reggie, Aug 16, 2006.

  1. Reggie

    Reggie Distinguished Member

    Dec 20, 2004
    Does anyone else find this article misleading?:

    Especially that little graphic. "Oh, so in digital you can only record and play back square waves?"
    I know for a fact that different recording programs show different graphical respresentations of the digital audio information. SoundForge will show a 20kHz wave as all jaggedy from sample point to sample point. CoolEdit will show the true analog representation of the sampled wave as it follows along the sample points. The peaks and valleys stay intact even though there are only a couple sample points per wave.
    Besides all that, since when is vinyl a more accurate format?

    Howstuffworks, indeed... OK, done ranting. Am I wrong?
  2. corrupted

    corrupted Guest

    Well, that's essentially true. It's blown WAY up in that image, that's not a realistic view of modern digital audio.
    The other thing is that the speakers and wires degrade any steps to the point where it almost couldn't be measured anymore.

    It's just really there to show the difference between the two technologies.

    Analog may not be a more "true" form of recording, but it does work on a pure-sine system as opposed to a sampled signal, so it is a clearer representation of the music. If you were to imagine a "perfect world" situation, where there was no signal loss or degradation, then yes, a vinyl record would more truthfully represent the actual wave. In reality, on the other hand, digital media makes much more sense to the market because it does not break down or wear out... it is just simply read with light.

    So, I don't really disagree with you... those are my thoughts on the topic.
  3. Cucco

    Cucco Distinguished Member

    Mar 8, 2004
    Fredericksburg, VA
    Hey Reggie -

    I just LOVE articles like that. They really do help to propogate bad and just plain old DUMB information about digital. ARRRGGGHHHH. :evil:

    They suffer from one of the most common problems of digital explanation I see on a daily basis - the confusion of bit depth and sample rate and their relation to amplitude and frequency respectively.

    Remember, with 2 sampling points, an accurate sinewave can be recreated. frequencies are moving greater than the possible sampling bandwidth....That's where the brickwall filter comes into play at the end of band...

    Besides, a note changing quickly causes digital to distort?? Huh...?

    Personally I don't put a whole lot of stock into the whole "Analog sounds better than digital."

    The fact is - YES, theoretically analog CAN sound better than digital. But, then you factor in all of the errors associated with analog (Wow, Flutter, Distortion, Oscillation, Noise, etc.) and you now have a severly flawed medium. Just think if all of those things existed within the digital realm. It would have NEVER been accepted on the market. EVER.

    So, to me, the majority of the Analog VS Digital debate is perpetuated by those that are too snobby to realize that their precious "perfect medium" is more flawed than Britney Spears' boob job or those who only read musings of said individuals.

    Don't get me wrong, I love the sound of a good record. ear has to do so much filtering of stuff that I have to "live with" the sound of analog to get the warmth and truth of it.

    Well done digital lacks all of the "harsh, sterile, cold" attributes that its opponents tout as its handicap but also lacks most of the errors and inadequacies of mediocre to even high-quality analog.

    Just some thoughts -
    Donning flack jacket now....

  4. Reggie

    Reggie Distinguished Member

    Dec 20, 2004
    And yeah, that part about the drum hit distorting because of the sampling rate was rediculous. I mean, seriously. Who comes up with this stuff?
  5. Thomas W. Bethel

    Thomas W. Bethel Well-Known Member

    Dec 12, 2001
    Oberlin, OH
    Home Page:
    The whole analog vs digital controversy has been raging since the introduction of the CD. There are still people who point to some of the original releases done on CD as classic examples of what is wrong with digital and they fail to realize that was eons ago and that things have progressed greatly since then. I know some reviewers would still like us to believe that their 1966 mono records of the Beatles sound better than a CD of the same event that has been mastered from the original tapes and sounds (IMHO) quite a bit better.

    I don't know how widely disseminated this information is but one of the reasons that early CDs sounded so harsh and brittle was because the CD players of that time were designed by digital engineers who knew a heck of a lot more about digital than they new about analog and when it came time to put in the analog circuitry the went to am IC audio cookbook (written by Walter G. Jung ), took the general purpose schematic out of the book and presto instant analog audio. Now of course no one would dream of doing anything like that but back then they needed to get the CD players out and copying a circuit from a book what a heck of a lot easier than having to sit down and design one from scratch. A lot of reviewers still remember those terrible CD players and can't get past them to the present.

    I love when people show digital as large steps superimposed over a sine wave. What was the sampling frequency in this example may 1000 Hz? Plus they never really explain all the rest of the scenario to their readers so the poor reader is left with the idea that they are listening to bumpy audio when it is digital and pure sine waves when it is analog.

  6. RemyRAD

    RemyRAD Guest

    Sep 26, 2005
    Another thing that many of us don't like but have learned how to live with is the necessity of a "brick wall filter" at 20kHz when recording with a 16-bit 44.1kHz recording system. It's the brick wall filter that does more to screw up the sound than PCM process alone.

    Plus we all have realized that the high frequency resolution is less than it should be in a 16-bit 44.1kHz world.

    A lot of people did not like the way CDs sounded when they were released from analog masters that had been optimized for vinyl release. Who could blame them? The analog master was not necessarily the analog master. The analog master was copied by the disk mastering engineer so that if any of the Stamper plates ever broke, they could create an identical Stamper plate, without having to go through the disk mastering process all over again thus improving consistency, even though it was down another generation.

    Where's my wire recorder?
    Ms. Remy Ann David
  7. JoeH

    JoeH Distinguished Member

    Jun 22, 2004
    Philadelphia, PA/ Greenville, DE
    Home Page:
    Analog tape and processing still has its merits, and I like the warmth, but I am SO over vinyl..... Below is an old rant that I was going to run elsewhere.....perhaps the time has come to revisit it. (Getting out the flame-retardant suit!)

    I've been through enough Vinyl BS over the years to say it is nothing but a subjective JOKE continuing to be foisted on that portion of the gullible public still looking for "vintage" jollies.

    Who wants to buy a format that begins wearing out the moment you play it? Quick; where can I blow more of my money on depleting resources! Gasoline, cars, food, and hey, why not! Let's go back to Vinyl!

    Anyone ever notice the gradual 5 db drop at 15K that happens from the beginning of the disc to the end? Of course not, it happens so slowly you don't notice it. Kinda like cancer or hair loss. (Ever wonder why they always put the best tracks at the beginning of the disc? Ever pick up the needle and start at an inner track? Notice something missing up there?)

    Few know this, but none but the largest of the REAL transcription turntable tone arms can accurately play a disc without distortion and tracking errors across most of the surface anyway. There are really only one or two "Sweet spots", (the very beginning and again about 2/3 of the way in); the rest is inaccurate.

    Am I the only one that hears a CONSTANT low level morass of surface noise whenever an LP is playing? Why do the "golden ears" hear oooh-so-much BETTER stuff coming off a Vinyl record, yet ignore all the artifacts? Talk about subjective filtering!!!!

    What about low-level AC hum? Ground wire or not, most of the time, there's STILL a hum present in any turntable feed, unless you've got a four or five figure $ystem with a preamp costing more than most people's car.

    What about platter rumble? What about constant tics and pops, no matter how well you take care of your collection? (Remember when people stacked them? HA!!!! Little bits of sand and dirt would rub between them causing even more damage.)

    What about off-center pressing, wow and flutter? What about pre and post echo from grooves cut too loud? What about cartridges skipping out of grooves when records are cut too bassy or loud? What about feedback from the cartridge (via the surface of the record) to the speakers when you turn it up too loud?

    I get mad just thinking about it. Vinyl was just a stupid, lousy, lossy format for so many years, and it p*sses me off to no end whenver I hear someone blathering on about how GREAT it is. BULL!!!!

    What people are hearing are old, comfortable, smoothed-over artifacts and rounded-down harmonics that remind them of their long lost youth, and their sense of accuracy goes out the window. Ever wonder why there is so much prep that goes into a 'Pre-vinyl" master tape, from EQ to read-ahead limiting to groove spacing? The compromises visited upon a completed master tape to "Fit" the audio into the grooves was/is galling. The very first thing to go is accurate transient response. ANY master tape - digital or analog - copied to vinyl is a sonic compromise, top to bottom.

    Vinyl always was not much more than dragging a boulder down a tar curb, hoping for the best. Early digital got a bad name with all the pre-equalized for vinyl tapes being used for so many early CD reissues. That stuff is long long gone, and I for one have no desire to go back.

    Don't get me wrong: I have an old standalone LP player with a single mono speaker there I play oldies but goodies from time to time for the sheer nostalgia of it, and I still transfer many rare LPs & 78's for clients (when there's no other version available). It's fun for what it is.

    For restoration work, we spend hours removing artifacts from unopened, unplayed vinyl pressings that have never been touched before, in some cases. The artifacts present in using even the best TT, preamp and never-before-played Vinyl recordings are immediately audible on any mid-to-high-end playback system of today. It's ludicrous, and it amazes me how subjective the so-called purists are who can't hear this stuff.

    You couldn't make me go back to vinyl with a gun at my head. And I make no apologies whatsoever for my belief. It's nothing but nostalgia getting in the way of the facts.

    Would I rather listen to a good analog reel to reel vs. Vinyl? Absolutely. A good analog reel to reel vs. today's digital? Maybe; depends on who set it up.

    Whew.....that was harsh, but I guess it was the coffee talking when I wrote it, about a year or so ago! :twisted:
  8. corrupted

    corrupted Guest

    Playing devil's advocate...
    The reason you see an analog signal is because of interpolation and averaging. As opposed to drawing the wave as it's created, the digital system samples a ton of points, and then "connects them with a line", creating an apparently smooth sine wave.

    I love digital, don't think I'm saying I'd rather have an analog setup...
    All I'm saying is, if you don't record the actual sound wave, you can't really say that you have a more true representation of the sound.

    It's becoming a pointless argument as digital technology improves anyway... it's like digital photography. Sure, lower end digital cameras can't get the resolution of a chemical photo, but some of the higher end digitals blow everything else away as far as clarity and resolution. It just takes time and people will all come around.
  9. Reggie

    Reggie Distinguished Member

    Dec 20, 2004
    corrupted "devil's advocate" sampling theory:
    Nyquist sampling theory:
    Yeah, I don't know. What is the definition of recording the actual sound wave? Aligning X number of iron oxide molecules to represent the wave while adding in some extra random noise and distortions not present in the original soundwave? Or using numbers and math functions to reconstruct a smooth wave, given a set bandwidth? Is there an analog format that is free of any bandwidth, distortion, or dynamic limitations? Hmmm.....

    And I'd be curious to hear your further thoughts on how digital photography correlates to digital audio.
  10. ghellquist

    ghellquist Guest

    A little knowledge can be very dangerous they say.

    Anyway, the article does not even come close in describing digital audio -- the description given is on a five-year-old level. The description of analog is at about the same level so you might say it adds up, but it surely does not! The article pictures analog as perfect and digital as the end of the world. We all know both has both shortcomings and pluses.

    Somewhere on the net Dan Lavry has written quite readable stuff better describing how digital actually works, using a mathematical concept known as sinelets. Even I could sort of grasp what he was writing so I guess you would not need any higher education in math to at least follow along. If I find the link I will post it here.

  11. corrupted

    corrupted Guest

    One more "devils advocate" point about the "Nyquist sampling theory"...
    If you sample an unpredictable sine wave at twice the frequency of the highest possible rate for the sine wave... you certainly don't have enough data to properly emulate the exact nature of the wave.

    Think about it this way:
    A complete cycle is one full sweep in the negative direction and one in the positive, and then back to zero. Now imagine one snippet of time containing one cycle of a 20kHz wave. If I were to give you 2 single points on a graph and tell you to plot the sine wave I was thinking of... you'd be hard pressed to do so. Especially when taking into consideration the complex nature of a sound wave that possibly contained.

    In my mind (which is not saying much, teehee), there is no way to reproduce a 20kHz signal by sampling at 44.1kHz. Do I think it's necessary to have 20kHz recorded precisely? Well, no.

    So... I'm not really arguing for anything, I just find that interesting.

    Yea, you could debate it in your own head forever... interesting topic as usual, but I could stress myself out thinking about it. :lol:
  12. corrupted

    corrupted Guest

    Very cool, I will look for Mr. Lavry's info... I'm always up for learnin' more about this stuff!
  13. Reggie

    Reggie Distinguished Member

    Dec 20, 2004
    Dude, the stuff you are talking about, with this unpredictable sine wave, would be filtered out if it is outside of the bandwidth upper limit(half the samplerate). The lower frequency stuff that can make up an odd-shaped soundwave is no problem. The sinc functions can accurately reproduce a smooth sine wave (or whatever shape as long as it is bandwidth limited to fit Nyquist) of any speed under half the sample rate. If you want to get out of range of the slight high-freq rolloff that happens with 44.1K filters, then by all means use a higher sample rate. But don't expect your ears to hear frequencies they can't hear, don't expect your speakers to be able to reproduce frequencies they can't reproduce, and don't expect tape or whatever to faithfully capture all this unpredictable transient stuff you are talking about.

    Here are the Lavry links:
    Link removed
    Link removed
  14. corrupted

    corrupted Guest

    Granted, I may be off topic, but I think you missed my point.

    I'm not "taking analog's side" by saying that, I'm just curious as to how one would go about reproducing any wave with merely 2 points, given that you don't have a zero crossing to key the sample from, and the frequency is not held to a clock, of sorts.

    Again, I'm not saying that this matters... I've just heard people telling me that it's "true" without giving an explanation or proof, so whenever the opportunity arises, well, I try to find an answer. I'm not at all unhappy with the quality of digital, I'm just trying to search for the meaning of life with a side of fries.

    I know that most people can't hear much near the 20kHz range... and in fact most people over 25 usually can't hear past ~15kHz. I also don't agree with Nyquist's theory, and I don't believe that digital sampling at 44.1kHz will give you a true wave at 15kHz. So, yea, 20kHz doesn't matter, but if your 15kHz+ was slightly wrong, that might be a downfall.

    All I'm saying is that the theory behind analog recording is more true to capturing a wave than is the theory of digital recording.
    No worries, I guess it's not worth discussing anyway.
  15. Reggie

    Reggie Distinguished Member

    Dec 20, 2004
    Wow, that's a pretty good rant you got going there! :shock:
  16. Reggie

    Reggie Distinguished Member

    Dec 20, 2004
    I think it is totally worth discussing. And I'm not saying that I understand all the math and magic behind it, but here is a test you can try sometime: Get an analog tone generator, put some tones through an A-D at 44.1k/24, screw the phase around by however many degrees you want (in order to try to screw up the sample rate), feed it back out a D-A, put an O-scope on it, and see what you get.
    You seem to be forgetting that each sample represents a sinc function, not just a point along a stair-step of points (as you may see it in your recording software, depending on what it is). Give the creators of digital audio a little credit. I mean, it would probably sound like crap if it was just stairstep plots used to reproduce the frequencies we hear, unless it was oversampled like 100 times. But that would be a waste when you can use sinc functions instead of line graphs. Besides, wouldn't we destroy our amps/speakers if we were putting 15K square waves and sawtooth waves through them all the time? :?
  17. corrupted

    corrupted Guest

    Well, it wouldn't exactly be square or saw waves, it would just be random noise at that frequency. Not exactly random, but not exactly correct either. I'm not implying that it's a stairstep type of plot, but more of an averaged interpolation based on points. So, theoretically, the wave would be close to correct, but not exact.
    About putting a scope on it... that's implying that what you record would be pure and constant at a frequency. Any slight change could throw it out for just a nanosecond until it caught back up with the "groove" of the wave. Now if you're constantly changing amplitude as well, it would be hard to track that all of the time.

    An expiriment of thought: If you sampled at 40kHz, and you were sampling a 20kHz signal, what if every sample was taken at a zero crossing? It would be silent... Is that why the extra 4.1kHz is there?
    What if every sample was taken with no reference to zero because of other frequencies of the sound wave? How would the A/D converter distinguish these from other frequencies? Or, does it just track each frequency individually... I don't think that's the case because it would take a ton of processing.

    It's also possible that there is some magic in these A/D converters that I just don't see.

    Hmmm, now I'm going to have to do a ton of poking around the net looking for test results and stuff... I like sparking the neurons now and then!
  18. Reggie

    Reggie Distinguished Member

    Dec 20, 2004
    Chew on this for a second:

    All the other stuff that doesn't fit into zero crossings and all-- as long as you bandwidth limit (filter) the signal, they are accounted for still by sinc functions or the sum of several sinc functions. I guess. :? Like I said, I don't totally follow all the complex math, so this is getting into the area that I will refer to as "magic." I do know it comes out the other side just fine. 5pm on a Friday is not my finest thinking hour.
  19. mpd

    mpd Guest

    I spent about the first ten years of my career as a DSP engineer working, doing both research and implementation. There is a lot of bad information on the net about sampling. Dan Lavry's papers are pretty good. Another good resouce is Rick Lyons, Understanding Digital Signal Processing. If sampling theory was wrong, they all our our cell phones would be paper weights, DirectTV would be bankrupt, and satellites would be falling out of the sky.
  20. Cucco

    Cucco Distinguished Member

    Mar 8, 2004
    Fredericksburg, VA
    Nika Aldrich's book is also good. Although I do find that it simplifies some things, it is helpful to the layperson (and even some non-laypersons.)

    While I also find Dan Lavry's papers excellent, I find that on occassional subjective issues, he is quite opinionated and can only present the side of the argument which better suits his opinion.

    He's still 10x smarter than I'll be any given day of the week, but I just take issue with a few of his more controversial ideas.


Share This Page