The Laserdisc FAQ
II) Hardware
This page is "http://www.iki.fi/leopold/AV/LdFaq/Hardware.html".This page is maintained by Henrik 'Leopold' Herranen
The whole FAQ file is available at "http://www.iki.fi/~leopold/AV/LdFaq/".
The Laserdisc FAQ
This page is maintained by Henrik 'Leopold' Herranen
The whole FAQ file is available at "http://www.iki.fi/~leopold/AV/LdFaq/".
Video, or CVBS, or composite video
Karaoke is a Japan bar/nightclub craze in which inebriated patrons sing the lyrics to instrumental pop tunes. Laser Karaoke players are normal players plus five features:
Here I just want to say that if you are living in a PAL country, you probably want a player with NTSC playback capability. Otherwise you'll be in deep trouble trying to find programs for your nice player.
The LD video signal is often referred to as a "composite" signal, which
means that the black and white alias luminance (Y) and colour alias
chrominance (C) information are not sent separately, but in the same signal.
It is the responsibility of the receiver of the signal to reseparate these
original signals. This task is by no means a simple one if you want to do
it well, and so-called comb filters are used to do the separation. If you
don't have a comb filter, or you have a poor one, a picture that should
look like the one on picture on the left side may very well look like
the one on the right side, or even worse.
The following is an article sent by Dave Martindale (davem@cs.ubc.ca), who has done a magnificent job in explaining the basics of comb filters (97-02-13):
If you look at a single scanline, you really can't tell what's luma and what's chroma. You could, in theory, build a comb filter that processed only a single line at a time, though it would be difficult to build. I'm not aware of anyone building a 1-D comb filter for TV use.
But now observe that the colour subcarrier inverts (180 degree phase shift) between two adjacent scanlines in the same field. If you digitize two scanlines with the pixel locations aligned vertically, and if you assume that the image is not changing vertically between the scanlines, then when you add the two scanlines together (and divide by 2), the chroma information cancels completely and only luma is left. Subtract this from one of the input scanlines, and only chroma is left. So the frequency- domain "comb" structure actually comes from spatial-domain subtraction, and the special relationship between colour subcarrier and horizontal line frequency. Thus, all comb filters are (at least) 2D filters in TV.
The simple comb filter above still has problems. The chroma that results is shifted 1/2 line vertically compared with luma. It doesn't handle colour gradients well. A 3-line filter that works on 3 scanlines simultaneously can eliminate both of these problems. There are still circumstances where a 2-D comb filter of any sort generates the "wrong" decoding, so you can do better by having several filter algorithms available and switch between them based on picture content - these are adaptive filters.
Another approach is to observe that subcarrier phase inverts from one frame to the next at the *same* pixel location. You can build a comb filter by digitizing two complete frames and averaging them together. If the image is stationary, the chroma cancels completely leaving only luma. Again, subtract this from one frame to get chroma only. This can give perfect separation of chroma and luma at each pixel regardless of what is happening at neighbouring pixels, unlike 2D comb filtering. This is a "3D" comb filter. Unfortunately, it only works properly with stationary image areas, so some other algorithm needs to be used when there is motion. Thus, in practice, a 3D filter should be an adaptive filter.
An S-Video cable is a cable that can send the luminance and chrominance as separate signals from an LD player to a TV set. Because the video signal on LDs is composite video, there is in principle no sense in using an S-Video cable to connect your LD player to your TV set.
However, if the LD player does some clever tricks to the video signal before transmitting it, like noise reduction, it has to separate the luminance and chrominance information anyway, and it is saner to keep them that way all the way to the TV set. Also, if the comb filter of an LD player having an S-Video connector is better than the one in your TV set (which is very usual), it is wise to use the S-Video cable.
So, if you use an S-Video cable between your LD player and TV set, you will lose only if the player does no clever image manipulating and your TV set has a better comb filter than your LD player. The only case of this I know of is the European 16:9 TV set Philips 32PW9781, which has the best 3-dimensional comb filters I've ever seen (it also has a built-in line-interpolator, which is roughly an equivalent of a line-doubler).
Excerpted from Bob Niland's (rjn@csn.org) LD#93: A test tool that never needs calibration Revised: 16 Jan 92
"A Video Standard", the Reference Recordings LD-101 laser video test/demo disc was released in May 1989. The list price was $59.95 and is now 69.95 (but has expanded booklet). LD-101 is available from most LD retailers, all non-club mail-order sources and in the US directly from Reference Recordings, Box 77225X, San Francisco, CA 94107 (415) 355-1892. In Canada, MAY Audio (514) 651-5707 is supposed to stock it ($99CAN+7%GST).
If you own or plan to own a laser videodisc player, are serious about video image quality, and don't own a roomful of NTSC test and pattern generation equipment - investigate this disc.
According to the jacket, most of the video material was created and edited in the digital (video) domain and not converted to analog until disc mastering. Both digital and analog sound is provided (not always the same, either), and the analog trips CX on and off at various times. Mono, stereo, Dolby Surround and Dolby Pro-Logic Surround material is included.
There are eight chapters and around a hundred still frames. Few of the stills are stop-coded, evidently to prevent interruption of the audio program. All of the stills are cataloged in the new instruction booklet.
With this disc, and your TV/monitor's service manual, and CONSIDERABLE CARE if you need to open the TV/monitor...
LD-101 is also useful for testing LD players, TV/monitor/displays and surround decoders, if you are in the market for any of these items.
However, there are several things that the THX program do not cover. One seems to be avoiding aliasing (I will write about aliasing when I have the time. For now you just have to trust me). Another is the actual pressing. Even the best master is useless if the pressing is poorly made. Thus, THX doesn't necessarily mean that the disc looks perfect.
THX also don't assure that you get the film in it's original theatrical aspect ratio. There exist several Pan & Scan THX laserdiscs.
The playback system does not need any special hardware attached to it to get all the benefits from THX discs. THX mastered discs are 100% compatible with all players and TV sets, because it's just a quality assurance program, not a format in itself.
You will need the schematic for your player and access to WWW. On the Web page " http://users.aol.com/chunter/ac-3mod.html " you can find complete instructions on how to make it. The author of this FAQ has seen this modification done on a Pioneer CLD-2950 multistandard player, and it works perfectly! Another version for the Pioneer CLD-A100 is at http://www.rit.edu/~jcd4878/Ac3/ac3.html.
The following narrative of LD care is in life-cycle order. The part 1 scenario presumes that you have just bought a disc. If it is a factory-new disc, or still in print, and can be exchanged, I do NOT recommend the part 2/3 repair procedures (not listed in this FAQ). Exchange the disc.
Disclaimer: These are the practices that I follow. The industry at large doesn't seem to have any. These practices are not warranted in any way whatever, and this series of articles is partially intended to provoke discussion and counter-suggestions.
Prerequisite: you need to know that the video data for side "X" is on the opposite side of the platter from the label for side "X". Also, LDs play from the inside to the outside.
Myth: LDs are indestructable. You can handle them anyway you like. The laser can read right through fingerprints, smudges and scratches.
The myth is false. The laser has a limited tolerance for optical obstructions. If you can see damage (or defects) in the acrylic, there is a strong possibility you will see video noise on the screen. LDs are not as robust as CDs.
If buying a used disc, check its contents. Compare the platter labels (and side count) to the jacket art, and make sure some sleazy customer has not switched old all-CLV analog pan&scan platters for the CLV/CAV digital widescreen ones you paid for.
If the wrap is the traditional taut variety, carefully slice it along an opening in the underlying slip case or box. Peel it off, taking care to avoid ripping any stickers affixed to the wrap. Save the wrap for the moment.
If the wrap is the loose slightly oversize variety common on Japanese imports and IMAGE/Kuraray production, consider using it as the outer storage sleeve. Cut carefully and evenly along the top, removing as little plastic as possible.
Remove any stray dust particles or hair with a soft lint-free cloth. Inspect the disc surfaces for gross dangerous defects, like severe warp, side-to-side mis-alignment and cracks. If any - DO NOT PLAY.
Sample the disc. Turn on "DISPLAY" to show the chapter number and time/frame. Play the sides in order, paying some attention to the program content and chapter.
If it is extensively chapter-marked, skip through by chapter. If not chaptered, play the start of each side, and seek to 11:11, 22:22, etc. When one of the seeks takes you to the end of a side, back up and watch the end of the side. Problems are usually most prominent there.
Find a problem? I have a separate article available, LD#17, "LD: What is a defect?" (For a condensed version, see section 11-1 of this FAQ).
Use poly-lined paper-reinforced rectangular sleeves, with center holes that allow reading the labels on both sides. I order them from Starship Audio-Video, but Radio Shack catalog# 42-132 sleeves will do. Reinforced sleeves are particularly important for multi-platter sets that are in single-slot jackets or boxed.
If the disc already has a quality sleeve, but was used or a rental, you might want to replace the sleeve anyway. It may be coated with food oils and salts left by popcorn-eating renters.
Insert the sleeve into the jacket with the sleeve opening UP, not facing out the jacket opening. Don't give the disc a chance to roll out onto the ground and crack.
For transport and loaning, I use a small number of LD-size PVC ziplock bags, obtained from a regional LD sales and rental store. PVC sleeves are on the market under several brand names, such as "LaserShield". I do not recommend the use of PVC (polyvinylchloride) for long-term storage of anything, including LDs. PVC outgasses vinyl chloride over time, which can attack metals and other plastics.
The purpose of the poly sleeves is to prevent damage to the jacket artwork and minimize air circulation if you live in a humid climate. In a normal storage cabinet, the insertion resistance is high enough that unprotected, the jacket art and text will eventually be scuffed away. The zip sleeves prevent the disc from sliding out of the sleeve during transport.
Holding the jacket in the normal horizontal position insert it in the outer sleeve with the sleeve opening up. As with the inner sleeve, the openings are at right angles to each other, preventing the disc/inner sleeve from sliding out accidentally.
If the player has disc support pads in the data region of the disc (and many do), lower the disc vertically onto the pads without sliding or spinning it. Also, periodically inspect the pads for the presence of abrasive debris; vacuum up any found.
If stopping play and ejecting the tray are two separate operations on your player, always wait for the disc to completely stop spinning before ejecting. I have seen CDs destroyed by players that dropped the CD, still spinning, onto in-data-area supports.
Always remove the disc from the player, return it to its slipcase and store it immediately. The disc is warm and dry after play, and will most benefit from being in a proper storage configuration. Under no circumstances leave a disc in the player for any length of time after play, as plastic flow may cause it to sag and develop warp.
From Bob Niland's (rjn@csn.org) LD#17: What is a "defect" on LaserDisc?
Pink Floyd's PULSE and pictures for the Web version added by Leopold.
In the Web version, you can here see an example of a good image. This image will be distorted so that you can see how different defects affect picture quality. If you have only 256 colours in your computer, some defects may become less visible.
Refers generally to artifacts resulting from interference
between two or more competing or adjacent signals. When used alone in
the context of LD, crosstalk typically denotes the herringbone pattern
(below) that results from the laser pickup reading, or partially reading
signal from one or both adjacent pit track(s).
This is usually a player tilt-servo adjustment problem, but is occasionally a mastering defect. The current LD of "Sleeper" (MGM ML101463, batch numbers 97-511A1 and ..B1) seems to have a mild case. The initial PAL release of Pink Floyd's PULSE seems to have lots of colour crosstalk.
\\\\\\\\\\\\\\\\\\\\\\\\\
///////////////////////// <--- This is what cross-talk looks like
\\\\\\\\\\\\\\\\\\\\\\\\\
Other forms of crosstalk within the NTSC signal itself (and not unique
to LD) can result in dot crawl, chroma crawl, cross-color and hanging
dots.
Loss of signal on NTSC LD (missing pits) is a form of noise that
usually results in the visible signal going to white level at the
corresponding point on the screen.
My criteria for "defective" is a disc region with at least two prominent noise specks per field (CLV) or four per frame (CAV) for a least one minute before returning the disc. If the source material was noisy, I wait until the disc noise is worse than the original film noise.
Vertical and diagonal lines, whether stationary or moving, are
almost always scratches on the film. White are on the print or
interpositive. Black are on the interneg or camera neg. Colored lines
indicate a scratch, in tri-pack color film or dye layer print, that
didn't make it all the way to the film base.
White lines indicate disc problems if stationary on CAV or scrolling on CLV. Lines lasting exactly one field are more often disc problems, but may indicate a film splice as well.
Fixed (CAV (left-hand side picture)) and scrolling (CLV (right-hand side
picture)) lines that are definitely media defects may be worth
complaining about. Much of the time, these turn out to be debris
on the disc surface that is easily removed. When otherwise, I
tolerate no (zero) episodes per title that last longer than one second,
and only two episodes per title that last between one frame and one
second each.
Although this can indicate a player problem, it can also result from media defects (and more rarely, mastering problems). I have several Technidisc pressings ("The Prisoner", episodes 1-6) with unreliable chapter numbers and timecodes. This is probably the result of poor signal quality in that portion of the vertical retrace interval that stores chap/frame/time. Since these IMAGE discs were subject to a repressing by Kuraray, I expect to exchange them at some time.
Since the bandwidth of the chroma signal is less than
half that of the luminance signal (in both NTSC and PAL), some minor
mis-registration of colors is unavoidable. In NTSC, intense reds seem
to be more susceptible to blooming and smearing. There are no LD media
defects that I know of that can result only in smearing.
A bad or missing comb filter causes dot-crawl at the edges of strong colours (as shown in the picture). A missing comb filter also makes some high-frequency luminance information to be interpreted as chrominance, as you can see with the vertical lines that are interpreted as color.
If the actual video running time is listed accurately on the jacket, with or without an "electronically time compressed" warning, you have a pre-sales decision to make. If the full film running time is fraudulently listed, you have grounds for return.
Note: PAL LDs from 24 fps film sources are routinely 4% time- compressed. Rather than use 12th-field-repeat, or electronic 104% pulldown, the common transfer process is to simply run the film at 25 fps (PAL is 25 frames/sec, 50 fields). The only way to avoid this is to use a multi-standard player and import NTSC discs.
More information about PAL time compression may be found at http://www.cs.tut.fi/~leopold/Ld/FilmToVideo/index.html#Speeds.
Loss of signal results in white-level
output on NTSC video (not sure about PAL). A moderate-sized point
defect on an LD will often result in many CAV fields having the same
line or lines stuck at white-level, or several seconds of scrolling
white lines on CLV discs.
Such lines are sometimes due to foreign matter on the surface of the disc. On new discs, tiny particles of stray acrylic (from edge trimming in production) scatter laser light quite effectively and can have a very dramatic effect on the video signal. Fortunately, a soft brush removes them. Such defects are almost always limited to the specific individual disc.
Media defects that affect a single pit, or a small
cluster of pits will result in either dramatic excursions of the
luminance signal (too white) on B&W titles, or trashing of the color
subcarrier, with resulting mis-decoding of the color at that point in
the scan line. The visual result is white or color snow, respectively.
An increase in snow over calendar time is the hallmark of laser rot, however, brand-new production can also have snow problems. Suspected causes of snow in freshly minted discs include: pressing discs with a worn or damaged stamper, pressing when the acrylic is not at the correct temperature, peeling the stamper and acrylic apart too soon or too late, metallization layer too thin, and contaminants in the production environment.
If all the dots are white on a color program, chances are they are film damage and not a disc defect. NTSC noise dots tend to be random colors. If the dots are over one scan line high on a CLV disc, then they are not a disc defect, since adjacent screen pixels are not adjacent on CLV (they are on CAV). If the dots have vertical "tails", they are almost certainly print or negative damage.
Streaking, or frame-to-frame variation in the color
of the same scan line (where no motion is occuring in the image)
indicates noise. Due to the coefficients used for the primary colors,
some amount of noise seems to be unavoidable in NTSC reds. High-end
player advancements may minimize this in the future.
However, color noise can also result from manufacturing defects in individual discs, and may also warn of incipient laser rot.
The general rule in the LD industry seems to be that a warp is a returnable defect if a US nickel (coin about .075 inch thick) can fit under the hub or outer disc edge at any point, when the disc is placed on a reliably flat surface.
On used or rare discs, be advised that warps are often correctable. See the "Care & Repair" article, LD#13.
The term "laser rot" is thrown around rather loosely these days. Oftentimes when people have a few speckles, they'll immediately think the disc has laser rot, but that is definitely not always the case. They might just have a manufacturer's defect that causes a few speckles. Laser rot starts out with speckles and gets worse until the disc is, for all purposes, unwatchable or unplayable.
Probably more than half of the 120 or so discs I own have speckles somewhere on the disc. It's one of the things you learn to live with as a disc owner. But if it's to a point where it distracts me from watching the movie, I return it. One of the qualities of a laserdisc owner is that you CANNOT be bashful about returning items. I have returned dozens of titles over the past three years of owning a laserdisc player. But then again, I'm picky about quality.
Although it is referred to as "laser rot," it has nothing to do with your
laser and it doesn't really rot, at least not in the traditional sense
where there is visible evidence. I suppose the word "rot" came from the
fact that if you watch a rotting movie a lot, say once a week or month,
the picture will get worse with each viewing, with more speckles and skips.
One could say the picture quality was rotting away over time.
Be warned that you probably won't see laser rot the first time you play a disc that eventually rots. You may not even see it the tenth time you play the disc. It could be months or years before the rotting process starts, and once it does, it can rot in a matter of weeks or many months.
If laser rot is a real concern to you, it is a good idea to view parts of your discs at least once a year. But again, this is really a topic that is overblown, and your chances of finding a laser rotted disc is pretty small, with the exception of a few documented titles.
If there is a possibility that air can get through the plastic coating of your disc, then the laserdisc can rot because of oxidation of the data layer. A deep scratch, an original manufacturing defect, or the glue heating and expanding (that's why you should store them in a cool, dry place) can let in oxygen. There's dozens of other possibilities.
Because of the name "laser rot", a lot of people think they will be able to see the problems on the outside of the disc, but this is almost never the case - you have to watch the movie. Only once have I seen a disc that showed external signs of laser rot, and I think that was a manufacturer's defect of the glue getting inside the plastic. It looked like someone had spilled coffee and it had dried out. I touched it, and all I felt was the plastic coating, so it was definitely inside the plastic, touching the metal. There were dozens of skips on the disc and I of course returned it for a refund.
Newcomers to the laserdisc format should not be scared off by laser rot. It was a concern in the early to mid-80s, but most disc manufacturers have seemed to worked out the kinks now. Bob Niland estimates about 2 percent of new discs have laser rot or other serious manufacturing defects. In other words, almost anything made in the 1990s should be OK.
If you're really worried about it, buy from a store that has an unconditional return policy, or that gives lifetime guarantees. Also, some labels/distributors, such as Voyager, give lifetime guarantees on their discs. Probably most companies, if you bugged them enough, would replace a disc for you if the title is still in print.
The truth is the defect rate is probably about the same for LD as it is for VHS or any other movie viewing media. But LDs start out at better resolution and don't usually wear over time (with the exception of occasional laser rot).
Almost all TV sets have vertical size adjustments either outside or inside the set. If you are having problems, and you are bothered by constantly losing vertical information, have your monitor adjusted. It can usually be done in just a few minutes by a technician.
WARNING! Do not open your TV set yourself unless you absolutely know what you are doing. Unlike other home equipment, TV sets may and do have fatal voltages inside even if they've been unplugged for days!
So, what is aliasing and what does it look like? That's the question I try to answer in this chapter. Because of the complexity of the issue, I've used pictures extensively. Unfortunately, this makes the text version of this chapter somewhat crippled.
Let's assume we have an image as shown left on film. And let's assume
it's a high-resolution image and we have to lose 3/4 of the resolution
both vertically and horizontally when we scan the image to video (this
is a typical case).
How do we downconvert the resolution? First, we have to scan it in horizontal lines, and limit the bandwidth so as not to get too much detail horizontally. But, you can scan in many different ways. To scan correctly, you have to either scan with a higher resolution (like HDTV or even more) and then downconvert, or at least you should scan with a dot height of one video scan line. If your scanner is more "accurate", there will be horizontal gaps between scan lines that aren't scanned at all! But where is the harm in that, you might ask. We'll get to the answer shortly.
Here is the simulated result of a badly scanned picture.
Low-pass filtering was used horizontally to get the horizontal
signal properly underscanned. But, the scanner used had such a
sharp scanning point, it scanned slices of 1/4 of the height of
a video scanline. The picture appears very sharp, but you can see
that the almost horizontal lines beneath the smiling man are not
continuous. This phenomenon, being able to see the pixels
because of bad low-pass filtering, is called aliasing.
This picture has been scanned with a beam similar to the previous
one. However, as one scanning line is only 1/4 of the height of
the original picture, 4 times more horizontal lines are scanned
to get the all the information of the original picture. After that,
low-pass filtering has been used also in the vertical direction
before downscaling the picture.
As you can see, in this small size the difference between this picture and the incorrectly scanned picture isn't very big: the badly scanned picture looks a bit sharper, but the correctly scanned picture has its horizontal lines better intact. But what happens if we display these low-resolution video images on a big TV screen? Let's see.
Here, from left to right, we see the enlarged versions of the bad and
the properly scanned images. Now, you might start to argue that the
leftmost picture does have a better overall look, but now here we
have two gotchas: line flicker and your eye's signal processing.
First, in the badly scanned picture, there are very many abrupt changes from white to black or vice-versa in the vertical direction. In the properly scanned version, there's almost always a grey pixel between the black and white areas of the picture. Why is this important? Because of the interlaced nature of NTSC video, any abrupt changes in contrast will introduce so-called line flicker. In the badly scanned version, the two lowest horizontal lines would flicker like hell.
Second, if you watch your TV set from a distance big enough that you can't quite see the scanlines, or you can barely see them, or if you are using a line-doubler or line-quadrupler that uses ideal signal reconstruction to reproduce the image, what you'll actually see is as follows:
To me, there is no question about which one of these two pictures more
accurately represents the original, high-quality movie frame. In the
properly scanned picture, all round shapes (head, eyes, mouth, legs)
are preserved, as are the almost horizontal lines beneath the
smiling man.
Although this is an artificial picture example, all this holds true for real-life pictures, too.
Aliasing seems to be especially a problem of Panavision widescreen transfers (some especially bad examples of popular films are Star Trek: Generations, The Thing (not the SE), Criterion's Halloween, The Star Wars THX versions (just look at those flickering stars)). Other movies, even if matted to 2.35:1 (for example Apollo 13, where stars are not flickering like in the Star Wars movies), don't usually show significant amounts of aliasing. My educated guess for all of this is that anamorphic movies are scanned in the original film-frame 4:3 ratio, and are squeezed vertically to fit the screen at a later stage. If this squeezing is not done with proper low-pass filtering, it's easy to introduce aliasing.
Fortunately, some newer Panavision transfers, like Braveheart, don't introduce significant amounts of aliasing, so there may be hope for us after all.
____ ___/ | | 20 | 1. Audio OUT R 11. G IN | 19 | 2. Audio IN R 12. - | 18 | 3. Audio OUT L 13. R or Chroma GROUND | 17 | 4. Audio GROUND 14. IR IN | 16 | 5. B GROUND 15. R or Chroma IN | 15 | 6. Audio IN L 16. Blanking | 14 | 7. B IN 17. Video GROUND | 13 | 8. Mode selection 18. Blanking GROUND | 12 | 9. G GROUND 19. Video OUT | 11 | 10. - 20. Video IN | 10 | 21. Screen GROUND | 9 | | 8 | 21 is the metal surrounding the connector | 7 | | 6 | | 5 | | 4 | | 3 | | 2 | | 1 | |________|21
A more detailed view on the pins:
A simplistic working cable can be built as follows:
Video Television 19 Video OUT --------------------------------- Video IN 20 17 Video GROUND -------------------------- Video GROUND 17 4 Audio GROUND -------------------------- Audio GROUND 4 3 Audio OUT L ----------------------------- Audio IN L 6 1 Audio OUT R ----------------------------- Audio IN R 2
____
/ \ 1. Chrominance GROUND
/ 2 3 \ 2. Chrominance
| 1 4 | 3. Luminance
\ == / 4. Luminance GROUND
\____/5 5. Screen GROUND
If you want to make a simple S-video -> S-video cable, you just connect all
5 pins straight.
To make a cable between an LDP with a S-Video connector and a TV set with an S-Video input with a SCART connector, the following cable would do the trick:
LD Player (S-Video) Television (SCART) 1 Chrominance GROUND --------------- R or Chroma GROUND 13 2 Chrominance -------------------------- R or Chroma IN 15 3 Luminance ---------------------------------- Video IN 20 4 Luminance GROUND ----------------------- Video GROUND 17 5 Screen GROUND ------------------------- Screen GROUND 21Of course, if you want your audio to be spoiled with the lousy loudspeakers of your TV set, a more complex cable is required and the RCA audio outputs of the LDP must be routed to the SCART connector.