DVD: Tomorrow's VHS, or Betamax?

I disagree about vinyl vs CD. Vinyl does not offer richer highs and lows. Some audiophiles claim that there is more ‘warmth’ in a vinyl recording, but I think it’s snobbery.

In almost every measure possible, CD’s outperform vinyl. A biggie is dynamic range, which is much, much better on CD. CD’s typically have a flatter frequency response, and it goes higher and lower than vinyl. But the frequency response of either is typically better than the speakers in the system anyway.

[index finger raised, crushing retort on the tip of the tongue]

[swallow; deep breath; quiet glower]

Ahem. Spending our capital at a rapid clip, are we not, Miss McFlimsey?

I’m sorry, dear, but you really do give me the most irresistible straight lines . . .

Say goodnight, Gracie… :slight_smile:

Omigod, did I really do a little smiley face?

my main plus for cd’s is that i have had the same 5-disk changer for 10 years now and it still works like the day i bought it—no replasment stylus here…

eggo

The interesting thing is that while recordable DVD players are positioned to be released Spring of 2000, a new digital VHS standard will also be available around fall of 2000 I believe. This one will be able to do something that DVD-RW/RAM can not do…record and playback full length HDTV quality films. DVD-RW/RAM doesn’t have the storage space to record 2 hours of HDTV level image and sound. DVD disks bought at the store can, becuase they can be layered, but DVD-RW/RAM can’t record layers. This information is about four months old, so things might have changed, I doubt it though.

>>Being Chaotic Evil means never having to say your sorry…unless the other guy is bigger than you.<<

—The dragon observes

So far we have one DVD player/drive in the house.Came with the last computer. I have yet to see a DVD-ROM disc like an encyclopedia or a game.

The Complete National Geographic Encyclopedia is a 4 DVD or 31 CD set. Riven is now offered on 1 DVD as opposed to the 6ish it took before.

if my memory serves me right… not saying it does… wasnt Beta a higher quality image and recording originaly over VHS??? I thought that why VHS was chosen over Beta was that Sony “who I believed owned beta” copywrited and wouldnt let anyonw record in its format while VHS format allowed anyone and HENSE MOST companies to use VHS and be seen. I.E. just like the minidisk… only sony owned companies can record on minidisk… am I wrong??? I dont think so… but let me know if I am!!!


“Boy, wouldja get a load of the cloaca on that one”? -Cecil Adams “october 8 1999”

I’ve got Riven on DVD.

Satan: in 20 more years, will the CDs still play? Do we even know if there’s a limit?

Somebody else brought up different DVD standards. For movies there are 3 different markets, regions. The cheaper players will only play US versions. It was the same with VHS.

Yaarghh…too many things to comment on!

(1) DVD has already displaced most of the LaserDisc market and part of the VHS market. This is because DVD has a higher quality (arguable vs. LaserDisc) and more reliable quality than VHS–a boon for the rental industry. (Most people use their VCR’s exclusively for movie rentals anyway.) Also, you can put more DVD’s on display at once than either VHS or LaserDisc in much the same way you can display more CD’s than albums.

(2) CD’s have a larger dynamic range, a wider frequency response, and are more durable than records, but records have a more accurate high frequency response. Even well played records hold up reasonably well–the recording is essentially intact although surface damage tends to be very disruptive to most people (ala pops & scratches).

(3) Hopefully DVD audio will displace CD audio. I believe the accepted DVD standard is 24 bits/channel (absurdly large IMHO) at 96K-samples/second (acceptable but low IMHO)…sadly they didn’t go with the mega-sample PCM which apparently had a very clean sound. My biggest beef with CD audio is the lack of accurate high frequency response–I think it’s sad that all the digitally recorded music will sound like crap next to DVD audio.

(4) For only about US$900, you can pick up the ReplayTV device which records up to 14 hours of programming on a hard disk. If the price continues to drop (which it definitely will) and they add a DVD-R drive…who would need a VCR anymore? Plus, if you could download your digital video from your camcorder via FireWire/IEEE-1394 to one of these boxes, edit it, then dump it to DVD-R…oops…I’m getting ahead of myself.

In the end, VHS will probably be rendered to the back burner like cassettes or magnetic computer tape in light of DVD.


Hey, aren’t you supposed to be at work?

I love my DVD player, DVD is thriving and will continue to do so for a few reasons:

It’s catalog is growing much faster than that of the older LD’s

Large scale rental outlets (Hollywood/BBV) Have embraced it, letting people rent rather than buy DVD’s

Many film companies are releaseing DVD’s for sale prior to the VHS version. For example, The Opposite of Sex came out for rental and DVD for sale the same day, if you wanted it on VHS you had to pay the “rental copy” price of 95 bucks.

And as mentioned before DVD’s have lots of extras.

I think they will continue with the extras mainly because all the cool features prompt renters into buying a disc. Also the films take up much less space on a DVD and it helps them to use up the space. Space will be even less of a concern when more dual layer disks come out. a double sided dual layer disk will store up to 17 gigs of info.

As for computer items, the delux encarts from MS that has an encyclopedia, thesaurus and a bunch of other refence books is available on DVD. The dvd version is on one disk instead of three, offers more video and sound files and has a better picture quality then the other versions.


Just add water, it makes it’s own sauce!

A long way upstream, realitychuck said:

This is not true. There are fundamental differences between laserdiscs and DVDs. First of all, laser disc is an analog format (the audio is digital tho). Obviously the storage capacity is amazingly different (despite being less than half the size, DVD is capable of holding an entire movie on one side, while laser discs almost always require two or more). Also, although both formats feature digital audio, laser disc audio is plain ol’ dolby surround, while DVD is full-blown Dolby Digital.

Live a Lush Life
Da Chef

Audiophiles, and audiophile magazines, are generally about one half step above faith healers and snake oil salesmen as a source for accurate information. One should take things they “prove” with a huge grain of salt.

I used to read a number of “audiophile” magazines - the folks writing these articles by and large have no clue. I must have read hundreds of their comparison tests of various sorts of audio equipment. I don’t recall more than a handful that used anything remotely close to proper experimental protocol. They didn’t seem to understand double blind tests properly normalized for other variables, perhaps because if they did them that way, they couldn’t claim that $50 RCA cables sound better than $5 RCA cables.

One facinating study done back in, hmm… I’m thinking the 80’s?, played the exact same signal through the exact same audio equipment, twice, but one case listening audiophiles were lead to believe by the experimenters they were hearing an expensive high-end amplifier, and in the 2nd, a cheap consumer one. They tended to lavishly praise the “high end” amp and gave a luke-warm reception to the more inexpensive one. Yet there was no difference - they were hearing the very same stuff, twice in a row. People hear what they expect to hear, and for this reason when looking at things like this, the only way to learn anything is to use proper experimental protocol.

Anyway, there is, apparently, a small camp of people who prefer the sound of records to CD’s. But this is due to some combination of the less accurate reproduction of records, and nostalgia (which is not to be underestimated). CD’s are in almost every measurable sense a more accurate reproduction of the original music. Whether that is good or bad, is a matter of opinion, but I would note that a more accurate signal can always be degraded into a less accurate one if this results in a more pleasing sound (as well it might), but the reverse is not true.

This would be a good example of how not to test two different audio components.


peas on earth

I won’t buy a vhs tape after watching dvd. I get so much more from a dvd than a vhs tape. The dvd is great with a surround system and you have all the bonus materials. Some dvd’s include extras for the computer and internet even. The vhs tapes I do have are going bad. I’ll be dead before my dvd’s wear out.

A side note: Thanks to all the teaming millions that helped kill off divix. Who wants to pay every single time you pop the disk into the player?

Granted… often by large margins.

This I seriously doubt, at least outside of some theoretical manner that never occurs in real life.

Even the best phono carts roll of steeply above 20 KHz, and about 99.9% of all commercially produced records were filtered to remove signal above about 16 KHz to reduce wear and tear in the manufacturing process. In contrast, CD’s are remarkably flat from DC up to over 20 KHz. It’s theoretically possible for a record to reproduce signal above 22 KHz that a CD could not, but this is the sort of play-once-test-record-under-lab-conditions situation that doesn’t happen in the real world, and it’s useless at any rate since you can’t hear those frequencies.

24 bits is indeed absurdly large, vastly exceeding the theoretical thermal noise floor at room temperature. Even 20 bits is generally only valuable in the mastering and mixing stages to give some breathing room to avoid digital clipping in the master - if you did this with a 16 bit master, you might only be using 13-14 bits. But when sampled down for the end product from a 20 bit master, 16 bit audio is perfectly adequate - often times overkill. The noise floor limit at 16 bits/sample is about -96 db.

And 96K samples per second is also absurd, unless you’re making recordings for dogs. The Nyquist limit there would be 48 KHz, whereas few adult humans can even hear to 20 KHz, and 12-16 is much more common. This sort of stuff is a marketing gimmick, and a device to sell people new stuff, but there is no other reason for it. Today’s 16-bit, 44K sample/sec technology is virtually never the limiting factor in the quality of audio recordings.

If you are an adult, and even in all likelyhood if you are a child, CDs can accurately reproduce frequencies higher than you can hear. There are a few sources of error near the Nyquist limit, but these are well understood and generally only of significance to people designing filters. If you back off just a little from the Nyquist limit, the filter problems tend to go away, and with CD’s this still leaves you with a higher freq than most humans can hear.


peas on earth

Ahh, bantmof…my infamous trap. True, a CD player can generate a sine wave at 20KHz with remarkable accuracy, however, it can only produce a sine wave at that frequency. This is what I meant by accurate frequency reproduction (I admit I was leading people to a trap by not explaining better, but here goes…all in the name of stamping out ignorance.)

There’s several ways to explain how these errors occur.

First is my simple-minded explanation. Consider an 11KHz sine wave sampled at 44KHz. For each full wave, you take four samples–let’s assume the normalized samples are +1, 0, -1, 0. When played back with the appropriate filtering, you reproduce the sine wave accurately. Now consider an 11KHz triangle wave sampled at 44KHz. Again, for each full wave you get four (normalized) samples of +1, 0, -1, 0. When played back, you get an incorrect sine wave.

Second, Fourier demonstrated that to reproduce an accurate waveform, you always need harmonics higher than the waveform base frequency (I don’t know much more than the basics of Fourier analysis). That triangle wave at 11KHz needs quite a harmonic at “infinity Hz” to perfectly reproduce the sharp corner at each peak.

Third, you can consider slew rates. A sinusoidal wave requires a sinusoidal slope, and at no point does that slope (i.e. slew rate) exceed the amplitude of the original wave. With any other wave, the slew rate needs to be higher than the rate to produce the maximum desirable frequency.

Records have the capacity to reproduce waveforms at frequencies up to 20KHz or so because they can reproduce sine waves up to much higher frequencies (I don’t know exactly, but 40KHz is entirely possible, albeit at much lower levels because the response rolls off). CD’s can reproduce sine waves at frequencies up to 22KHz, but waveforms accurately up to about 4KHz or so (10 samples per wave I think is okay).

If you listen carefully to CD’s, you’ll notice that cymbal hits, “s” sounds, and “sh” sounds are remarkably similar on a CD but sound more natural on a record. For me, it’s this harsh distortion that irritates my ears when listening at clubs and such–much more than when I listen to live music.


Does this belong in another topic yet? :slight_smile:


Hey, aren’t you supposed to be at work?

CD sound is pretty excellent, and I can rest my beer on one and it doesn’t mind, nuff said. I own a recordable MD player; the sound quality is excellent, especially seeing as that I use it mainly when I am travelling. It will never replace the CD IMHO, the big advantage it had from me, apart from being a shiny new toy, was that I could record stuff from CDs with no loss of sound quality. Ergo when I travel I can take one tiny, shock resistant MD player and 3-4 MDs, rather than a clunky CD player with 10-12 CDs.

What is a “theoretical thermal noise floor” anyway?

Well… read on. Your trap needs oiling. :slight_smile:

I know what you were doing - I’ve been down this road many times before. I believe there are some things you are misunderstanding on a fundamental level here. I shall try to explain.

Well, it turns out that I do know more than the basics :-). (I had to take two semesters of this stuff in college; I’ve forgotten a lot, but a lot is still there - and BTW, this stuff makes for really boring classes, trust me :-). Your statement here is perfectly correct, but it doesn’t lead to the conclusion you think it does. I’ll try to explain why.

Let’s start with the 30,000 ft view in this post - we can go as much lower as you want, within reason (I might have to brush up on some long disused calculus if we go into too much detail). This might be slightly awkward in ASCII, so bear with me here.

  1. First, an arbitrary input signal can be represented as a linear superposition of sine waves at various frequencies. Specifically, an arbitrary time domain signal f(t) can be represented as f(t)= sum(j, Fjexp(-iwjt)), where w is our base frequency. (Recall that a complex exponential can be written as cos-isin, so this exponental is really a sinusoidal summation).

  2. The array of coefficients Fj is the discrete Fourier transform of our input signal f(t). This transformation is reversable; that is, given an input signal, we can construct a (potentially infinite) array Fj, and given this infinite array Fj, we can reconstruct the original input signal with perfect accuracy. That is, we basically have a way to decompose an input signal into it’s component frequencies.

  3. Now, I’m going to gloss over some math - we can go into it if you want. But certain types of input signals have a value Fj=0 for all j>i. This means that these signals have no higher order frequency content. We can take our Fourier coefficients, run them through a low pass filter by discarding j>i, and still perfectly reconstruct the input signal with no error.

  4. Other types of input signals, such as your triangle wave example, or a square wave, or an impulse function, require an infinite frequency summation to reproduce. (I’ll just ask you trust me on the math for the time being). That is, if we run a low-pass filter on the Fourier coefficients, we cannot perfectly reconstruct this signal from the result.

Now, the conclusion you drew from this in your post is that this is a problem with CD reproduction of recorded sound. But this is a mistaken conclusion.

Let’s look at your triangle wave with a period of 20,000 Hz. If you examine the DFT of such a wave, you’ll see that there are frequencies in the Fourier array Fj that extend far (actually infinitely far) beyond 20,000 Hz. So it’s definately true that with a bandwidth limited device, we cannot perfectly reconstruct this triangle wave. We can approximate it to better and better degrees as our bandwidth goes up, but for any arbitrary bandwidth limit, our reproduction will necessarily be imperfect.

So what does this mean for our CD player and our record player? It means that neither of these devices can perfectly reproduce this signal. The CD and the record player represent waveforms in different ways, one digital and one analog. But that’s not the real issue here. The real issue is that both of these devices are bandwidth limited (as are analog audio tape and all other audio recording technologies). The CD is bandwidth limited by the Niquist limit imposed by the sampling rate of the device. In physical terms, a phono cart cannot track a perfectly triangular groove. It has inertia, there are electronics issues in the conversion to an electrical signal, and so on. It just can’t do it. The analog tape suffers from similar physical issues that limit its bandwidth.

So neither the CD nor the record can accurately reproduce this 20,000 Hz triangle wave.

Fourier transform theory tells us that the device that can reproduce it the most accurately is the device that can reproduce a sine wave at the highest frequency. The one follows directly from the other. If the record player was able to reproduce this triangle wave more accurately than the CD player, then it could also reproduce a sine wave at a much higher freq than the CD player can. But in practical applications, it’s limiting frequency is lower than the CD player, and both can be measured accurately. Not only that, but your EAR is a bandwidth limited device, being itself a physical thing with parts that can only move so fast. If you could accurately hear the difference between a triangle wave at 20 KHz and something that was not quite a triangle wave, then you could also hear a sine wave at a much higher frequency, and it could be said that you could hear to, say, 40 KHz instead of 18 KHz. But your ears just don’t have that sort of bandwidth. So the CD’s bandwidth is not the limiting factor.

In many applications, this sort of signal accuracy is measured using square waves at somewhat lower frequencies than the bandwidth limit of the device. By comparing the post-reproduction (no longer quite) square wave with the “perfect” square wave, you can see how much distortion has been introduced by your recording technology. And when you do this for records and CDs, CD’s win, at both high and low frequencies.

With all due respect, I don’t think it’s my ignorance that needs to be stamped out here :slight_smile:

I can go as deep into this as you want - I know a lot more of the math than I’ve put into this post. But I wonder if there might not be a better place for it - maybe Great Debates?

signed,
bantmof the engineer-nerd


peas on earth