Answer provided by Allan Jayne (ref):
You will need a processor (scaler, de-interlacer, line doubler, line tripler, etc.) if your video display does not accept the scan line count and frame rate and, if digital the bit rate, of the source device(s) (DVD player, over the air tuner, etc.) Except for the regular analog VGA signal (640* x 480 x 60 Hz) which is the same scan rate as NTSC 480p progressive scan analog output, the standard computer resolutions (scan rates) do not match those of TV and consumer video. *Both for computer and entertainment video, each analog scan line does not have to be seemingly divided into exactly 640 pixels.
2) What is a line doubler, what is a deinterlacer. what is a line tripler and a line quadrupler?
Answer provided by Allan Jayne (ref):
A de-interlacer takes interlaced video and as best as it can, generates progressive scan video frames that look as if they were created or televised as progressive scan from the very beginning. Because progressive scan frames have twice the number of scan lines as equivalent interlaced scan video fields, the terms de-interlacer and doubler are often interchanged. The simplest doubler just causes each scan line to be output twice in the time it took the original scan line to arrive. A tripler generates output frames with three times the scan lines, a quadrupler generates output frames four times the scan lines. If the source is interlaced video, a quadrupler should do good de-interlacing first, then redouble the lines. A tripler starting off with interlaced video should do good de-interlacing followed by scaling with a 3:2 ratio, rather than simply tripling the lines.
3) What's the difference between a line doubler and a scaler?
Answer provided by Allan Jayne (ref):
We can say that a line doubler is a kind of scaler. More specifically a line doubler is intended to create output video frames with exactly twice the number of scan lines while a scaler usually has several ratios of output scan lines to input scan lines per frame.
4) What is film mode? What is video mode?
Answer provided by Allan Jayne (ref):
"Video" mode is the "normal" mode for a de-interlacer in which the incoming content is not expected to have any special characteristics in terms of sameness from one field to the next. The de-interlacer must use its most sophisticated analysis formulas in constructing the progressive frames. In film mode a de-interlacer takes into account the 3-2-3-2 repeat sequence (3-2 pulldown; 2-3 pulldown) of subject matter in successive NTSC video fields produced from 24 frame per second film source. By keeping track of this pattern the de-interlacer can quickly find the matching field to weave when constructing the progressive scan video frames; such matching field either precedes or follows. (For PAL video of a 24 fps film or for NTSC renditions of the few 30 fps films, the repeat pattern goes 2-2-2-2, called 2:2 pulldown)
Answer provided by Jon Lindgren (ref):
When creating DVDs, there are two methods in which video is transferred: from
film, and from a video camera source.
A bit of history is necessary:
When you go to the movie theaters, the film you watch has 24 frames per second
of information. The video you watch at home, however, is presented at 60 fields
per second. To fit 24 frames/sec into 60 fields/sec, a processed called telecine
is used. Telecine breaks each film frame into 2 fields (call them A and B), and
produces a regular cadence (a repeated pattern) to produce 60 fields per second.
The end result is that the 48 fields/sec (generated from 24 frames/sec) is
"expanded" to 60 fields/sec by repeating certain fields. The key
concept here is that the two fields used to produce the cadence are from the
SAME frame of the image (and thus can be recombined later to reproduce the
original frame).
When video is produced from a video camera, however, the output is already 60
fields per second. The difference from film mode (besides the rate) is that each
field is taken from a different point in time (i.e. not the same frame): while
field A is taken at a point in time, field B is taken 1/60th of a second later.
The key concept here is that the two fields are taken from DIFFERENT
"frames" of the image, thus is is impossible to reconstruct a frame
for a given point in time.
When a deinterlacer recombines a signal, it can recombine it in two modes: film
and video. When in film mode, the deinterlacer is smart enough to recognize the
cadence produced by the telecine process. It can see that the two fields
originated in the same frame, and recombine them to produce an exact replica of
the original frame. When in video mode, however, it is not possible to recombine
two fields into an original frame since each field was captured at a different
instance of time. The resulting deinterlaced image can suffer from combing in
areas of the image which moved from one field to the next.
Thus, deinterlacers have two modes: film and video. When in film mode, the
original frame of film can be perfectly recovered whereas in video mode, it is
impossible to recover the original frame. Some companies have technology which
attempt to interoplate between different video fields (such as Faroudja's DCDi
(tm) and KeyDigital's ClearMatrix (tm)). Many deinterlacers will switch between
these two modes automatically, however, some do not.
As a side note, it is still important to have a good video mode on a
deinterlacer even when watching films. Sometimes a bad cut or edit is made when
producing the DVD which interrupts the cadence of fields. In this case, the
deinterlacer cannot reconstruct the original frame of film from the fields
presented, and has no choice except to revert to its video mode. Good
deinterlacers will do this automatically to prevent excessive combing, and will
switch back into film mode as soon as the cadence is recovered. Poor
deinterlacers will stay in film mode too long, and recombine fields from
different film frames causing artifacts.
5) What is inverse 3:2 and 2:2 telecine?
Answer provided by Allan Jayne (ref):
In video reception and display this refers to the possessing of a film mode for de-interlacing. This term is actually a misnomer here; the 3-2 pulldown was already present in the incoming interlaced video and should be preserved in the output progressive scan video.
Answer provided by Allan Jayne (ref):
Bob consists of treating each video field independently for de-interlacing purposes; for example in the odd field, even scan lines are synthesized or interpolated from the content of the adjacent odd scan lines. Vertical resolution is cut in half. It is so named because in some implementations, the finished picture seems to vibrate up and down slightly. Weave consists of taking material from the next and/or previous fields to fill in as the intervening scan lines in a progressive scan video frame. Using weaving throughout, subject motion shows up as an obtrusive comb like blurring of the sides of moving subjects (motion artifacts; see question 8).
7) What sort of artifacts does a good line doubler resolve?
Answer provided by Allan Jayne (ref):
A good de-interlacer or doubler delivers video frames with a minimum of (or
no) motion artifacts (see question 8) and with a minimum of (or no) loss of
vertical resolution for the stationary subject matter in the picture.
8) What are motion artifacts?
Answer provided by Allan Jayne (ref):
For all subject motion in interlaced video, one field, say the even scan lines, will show the movement first, then the next field of odd scan lines shows more movement leapfrogging over the previous field, and so on. A mediocre de-interlacer (or doubler if you insist) may construct video frames by blindly weaving together adjacent fields, with the subject motion differences contained within showing up a a ghosted double exposure effect. This is also called combing because the feathered side edges suggest the teeth of a comb. A good de-interlacer delivers video with more solid side edges, without this combing.
Answer provided of Ofer LaOr (ref):
I was actually referring to line flicker (which appears on angled thin lines that appear and dissapear at 60Hz), strange jaggies that are common in interlaced transmissions, moire & rainbow effects that are common when people wear flashy cloths with thin patterns that bounce all over the screen, etc.
9) What makes 72Hz such a magical frequency?
Answer provided by Allan Jayne (ref):
Some viewers notice the slight jerkiness of subject movement on standard NTSC video with the 3-2 pulldown simply because every other film frame is represented by three 1/60'th second video frames or fields and the intervening film frames represented by just two. With a 72 Hz refresh rate (72 fields or frames per second), each film frame gets equal display time on the video screen. Equal display time could also be achieved with 48 Hz (approximately the PAL field or frame rate) but more viewers will notice more flicker.
Answer provided by Ofer LaOr (ref):
...72hz is progressive, it MUST be 72 FRAMES a second (not fields OR frames). 72 is a magical number because each original 24fps frame appears exactly 3 times (it's a clean multiple of 24). 48 is also magical, but it flickers too much for many people. 72Hz is above the flicker range of most people and is the most projector/plasma friendly 24 multiple we know. The next best multiple is 120Hz, which contains both multples of 60Hz (video sources) and 24fps (film sources), so a clean switch between them can be done without introducing new artifacts. After that, 300Hz is the best overall frequency because PAL's 25hz is also a clean multiple.
10) What is the chroma bug and why should I worry about it in my scaler?
Answer provided by Allan Jayne (ref):
The chroma bug (chroma upsampling error) occurs because most digital video (including the DVD standard) has every two scan lines sharing the same coloration. (The light and dark details are still unique for each scan line). All DVD players first decode the video as interlaced, and when the bug exists, scan lines 1 and 3 get the color that belongs to scan lines 1 and 2; scan lines 2 and 4 get the color that belongs to scan lines 3 and 4. This causes discolored streaks to sometimes appear near the tops and bottoms of color patches, and diagonals to sometimes appear more jagged. After the video leaves the DVD player, the color has already been assigned to each scan line and a scaler further on cannot fix the results of the chroma bug, at least without considerable artificial intelligence taking hints from the video content pixel by pixel. Instead, some scalers and de-interlacers do some interpolation to hide the bug. The scaler or de-interlacer should at least have a switch to let the viewer turn the interpolation off.
Answer provided by Ofer LaOr (ref):
The chroma bug was probably introduced by a company who built one of the most basic blocks in the MPEG decoder and sold it to companies who ultimately produce MPEG decoders today. Most MPEG decoders have this bug. Panasonic made MPEG decoders (and DVD players) do not have the bug. The bug manifests itself as thin black horizontal strips occuring every other line, or alternating between two colors near edges of sharply contrasted color objects. A chroma filter or resampler can fix the problem, and such a filter exists in chips like the Faroudja DCDi chips. When these chips are present, the bug will have been fixed in PROGRESSIVE output of the DVD player only! A good place to spot this artifact is in the ToyStory main menu (the Blue text) and the Big Red Button sequence in "The 5th element". Thanks to Hometheaterhifi.com for introducing and following up on this bug for us.
11) What kinds of connections should a scaler have?
Answer provided by Jon Lindgren (ref):
A scaler should have input connections for all your devices. Usually, a
scaler will provide a composite, an s-video, and a component input. Most also
provide an RGBHV/HDTV passthrough, which you can use to hook your HD receiver
and/or your PC through to your display device.
Although some scalers offer multiple composite, s-video and component inputs,
most do not. In some cases, you may need a quality video switch, which can be
rather expensive. Examine your requirements and determine if it is cost
effective for you.
Optionally, look for scalers which accept SDI if you plan to use it. A scaler
which has a slot for future upgrades may help protect your investment, although
not many scalers support this feature.
Inclusion of newer interfaces (such as DVI and Firewire) is also a plus, even
though there are few sources which produce DVI or Firewire as an output. Things
are changing rapidly, however, so think of the future.
Answer provided by Jon Lindgren (ref):
The two most popular are manufactured by Sage (who own Faroudja and produce
the Fli2x00 series) and Silicon Image (who produce the SIL50x series).
Both are excellent deinterlacers, each with certain advantages. While most
Faroudja chips contain the DCDi (tm) algorithm (which helps deinterlace and
de-artifact video sources), the Silicon Image have no special video algorithms.
Some feel that the Silicon Image have a slightly superior film mode.
There are other proprietary deinterlacers available from different companies.
Many TV makers (such as Sony) include their own deinterlacing engines.
13) Why makes DCDi so popular?
Answer provided by Jon Lindgren (ref):
DCDi garnered its popularity by being the first "commonly
available" algorithm to help deinterlace video (some high end companies
such as Snell & Wilcox and Teranex had algorithms for a few years before
DCDi was introduced, but due to the high price tag of these devices, there were
not available to common folk). Due to the nature of video sources, it is
impossible to deinterlace perfectly. DCDi helps to interpolate the differences
between different video frames, which helps to reduce jaggies and other
artifacts. It is quite noticable on video based sources, such as news broadcasts
and sports.
The "waving flag" video sequence is usually used to show how effective
DCDi can be.
14) What benefits does SDI add to a scaler?
Answer provided by Jon Lindgren (ref):
SDI is a digital connection from a video source to a video
receiver/processor. One major source of noise and loss of resolution is the
analog to digital conversion of a component or s-video signal.
A typical path from DVD player to display is as such:
DVD player -> D/A -> cable -> A/D -> scaler -> D/A -> cable
-> display (possible A/D).
All the digital to analog and analog to digital conversions cause loss of
detail, and can introduce noise into the video signal. If you replace the link
from the DVD player with SDI, you've knocked out one set of conversions:
DVD player - > SDI -> scaler -> D/A -> cable -> display (possible
A/D).
You can reduce this further if your scaler and display support DVI:
DVD player -> SDI -> scaler -> DVI -> display (possible A/D)
The result is nothing short of a remarkable increase in picture quality. Since
signal noise is virtually eliminated, the image is clearer and more detailed.
Also, lack of video noise can help the deinterlacer more accurately recognize
the 3:2 cadence present on film mode material (whereas excess video noise can
prevent the scaler from recognizing the cadence, which generally causes the
scaler to use video mode as opposed to film mode).
15) What benefits does DVI add to a scaler?
Answer provided by Allan Jayne (ref):
All de-interlacers, doublers, triplers, etc. and scalers process the video as digital. If the incoming video was analog, it must be digitized. Every time video is digitized, some loss of horizontal resolution occurs. Therefore DVI (the video is digital as it comes in) reduces the number of analog to digital conversion.
Answer provided by Jon Lindgren (ref):
As noted above, DVI eliminates a possible digital to analog and analog to digital conversion, resulting in a clearer picture with less noise and more detail. In addition, having a DVI input to the scaler may allow greater flexibility in the future, as DVD players and receivers start to incorperate DVI outputs.
Answer provided by Ofer LaOr (ref):
DVI is actually digital video coming OUT of the scaler. SDI is digital video coming in. We also have to explain why SDI is not sold in regular players but only very high end players and through Mods.
16) What about HDCP support for my scaler?
Answer provided by Jon Lindgren (ref):
HDCP is slowly emerging in actual products such as HD receivers. Depending on
your application and how long you expect your scaler to last, HDCP may be
necessary for you. Determining how you will use your scaler and with what
devices will help you decide whether HDCP is a necessary or desired feature.
For instance, if you plan to view DVDs and standard TV only, HDCP probably isn't
necessary. If you plan to use HDTV or any newer high definition variant (such as
DTheater or the up and coming HD-DVD), HDCP is most likely necessary.
17) Do we need scaling for HDTV?
Answer provided by Allan Jayne (ref):
Ideally no. Ideally the display should be able to show all 1080 scan lines (or rows of pixels) of 1080i HDTV, all 720 scan lines of 720p HDTV as-is, and either 960 scan lines (with an additional doubling) or 480 scan lines for 480p. Unfortunately digital fixed pixel displays such as DLP and LCD, and some CRT displays do not have all three of these display modes. They therefore employ scaling.
Answer provided by Ofer LaOr (ref):
I disagree for two reasons:
Also, see continued discussion at avsforum.com
18) What does "Interlaced" mean? What does "progressive" mean?
Answer provided by Allan Jayne (ref):
Interlaced: The odd scan lines are drawn, then the even scan lines, then the odd scan lines, etc. Progressive scan: All scan lines are drawn in order. Due to limited information transmission capability back in the early days of television, and also for 1080i HDTV today, only enough picture material for a frame every 1/30'th a second was transmitted, and using progressive scan at the same frame rate would have resulted in the top of the picture fading while the bottom was still being drawn. Interlaced scanning was the best compromise.
Answer provided by Ofer LaOr (ref):
When television technology was first invented, progressive scan images were
first used. However, CRTs used to power the first television sets in the 50s
were too slow to display an entire frame all at once. Television engineers could
either display images at half resolution (bad picture quality) or come up with a
clever solution to the problem. Interlacing was the "clever"
resolution - they simply split the picture into two halves (frames) - each
consisting of the odd or even lines in the image. By sending these alternatly,
the picture tube was able to refresh quickly enough and human eyes (through
what's called the Krell factor) were fooled into believing they were seeing a
complete and nearly full resolution image. As computer systems got more advanced
in the early 90s, the advancements display technology for computer screens
leaked into television screens. Televisions are now CAPABLE of displaying
progressive images, but compatibility with legacy systems (recorders, old TVs,
old systems) requires that progressive images be sent through HDTV systems.
As a result, the market was left wide open for scalers and line doublers that
could convert SDTV interlaced transmissions into well formed progressive images.
19) Why does "Progressive" video look better than "Interlaced" video?
Answer provided by Allan Jayne (ref):
With interlaced scan the fading of the odd scan lines as the even lines are being drawn, and vice versa can still be seen. Also the leapfrog effect as motion appears, say first on the odd scan lines, then on the even scan lines, etc. can be seen by some viewers. Progressive scan is intended to give a smoother picture by eliminating these effects.
20) What is a field, what is a frame?
Answer provided by Allan Jayne (ref):
The term field refers to just the odd scan lines, or just the even scan lines of interlaced video. Frame refers to all of the scan lines, whether as one progressively scanned frame or two interlaced scan fields.
22) What is a noise filter? How does it work?
Answer provided by Allan Jayne (ref):
NTSC (and also PAL) video as transmitted has the luminance (light/dark; black and white) information occupying most of the channel space allotted with color information modulated on a subcarrier and somewhat commingled with the higher, or finer detail, portion of the luminance information. To preserve the full resolution of the picture, the color and luminance information must be separated and a comb filter is used for this purpose. WIth no filter, color information entering the luminance circuits gives the picture a silk screened appearance, more specifically light and dark graininess corresponding to the frequencies of the modulated color signal, around 3.58 MHz. Lower priced TV sets use a notch or low pass filter instead of a comb filter that eliminates the silk screen effect but also eliminates the finer picture detail. Meanwhile feeding the unfiltered color information to the color circuits results in discolored swirls, usually pink and blue, appearing where fine picture detail such as on a pinstriped shirt should be. No comb filter is perfect, the less sophisticated ones leave behind some color information in the luminance signal that shows up as silk screening around color boundaries that is also called dot crawl since the graininess seems to move along the color boundaries. Only the most sophisticated comb filters, known as 3 dimensional motion adaptive comb filters, achieve a significant reduction of discolored color swirls. The comb filter is so named because its frequency response, if graphed, suggests the teeth of a comb. Comb filters work because, in composite video, the color video frequencies and luminance video frequencies occupy narrow evenly spaced bands that for the most part do not overlap. Comb filters in consumer video devices are not true frequency band attenuation filters but rather perform phase cancellation by commingling the content of adjacent scan lines. They rely on the phase relationships unique to composite video. They separate color and luminance but will not remove noise.
Answer provided by Ofer LaOr (ref):
Originally, television transmissions were monochrome. When color was
introduced, it had to be added in such a way as to allow old B/W televisions to
receive the transmission. So, the old Luma signal (B/W transmission) had a new
companion - a chroma signal which contains the Pb (difference from blue) and Pr
(difference from red). These signals were combined together and were transmitted
in a frequency that's relatively close to the original Luma transmission
frequency. Old B/W televisions would disregard this signal and new televisions
needed to receive this signal and decode it into color. However, because of the
proximity of the two signals, a filter is required that would separate the two
without causing too many artifacts in the screen. There are a few technologies
that can create a filter that separates the luma and chroma signals. The most
common is a comb filter. Signals in Composite cables use the exact same
specifications (only they are not modulated over an RF frequency that can be
broadcast through the air) and need to be separated in the same manner.
A badly designed comb filter can produce an artifact known as "dot
crawl". This is caused because frequency ranges between the luma and chroma
can intersect and one can interfere with the other. Dot crawl is seen as color
dots (in the shape of slanted lines \\\) across the edges of objects. These dots
appear to be rolling up or down in a slant. Improved comb filters or notch
filters can reduce or eliminate this artifact.
Note that a comb filter is only employed when composite or RF signals are
involved. Higher end signals such as SVideo, Component and RGB do not require
this filter.
24) What is a TBC, do I need one?
Answer provided by Allan Jayne (ref):
A time base corrector makes sure that the picture information for each scan line begins at the same horizontal position on the screen despite irregularities that may have been introduced by less them perfectly smooth videotape travel or analog videodisk rotation. Otherwise vertical lines may have bends in them.
Answer provided by Ofer LaOr (ref):
If you use VCR technology you do need one. Scalers rely on exact frame timing, which is not present in VCRs because of the technology. Many high end VCRs contain TBCs already. However, the most common consumer units do not. As a result, if you plan on connecting your VCR to your scaler, make sure that one of these two contains a Time Base Corrector.
25) What is chroma bandwidth expansion? Do I need it?
Answer provided by Allan Jayne (ref):
Chroma bandwidth expansion makes left to right transitions from one color to another more crisp. It does not increase the amount of color detail in the picture. It is probably an item that some viewers may prefer and others not, since no process can determine whether a color transition was supposed to be that sharp in the original subject matter. So named because sharper color transitions require a higher bandwidth for the video subsignals that carry the color.
26) What is so special about a waving flag?
Answer provided by Allan Jayne (ref):
Flag waving is a picture defect consisting of time base errors in the first several and/or last several scan lines, causing the top and/or bottom edges of the picture to waver from side to side, giving a distorted appearance to the picture content there. It may be caused either by irregular videotape movement or by deficiencies in the TV electronics.
Answer provided by Ofer LaOr (ref):
The American flag contains many thin lines (across the edges and red horizontal lines across the section of it). As it waves, the edges of the flag and the red lines shape waves which form during the flag's motion, cause a high number of jagged edges in a relatively small space. These edges are exacerbated when you view them in a progressive image. When this is viewed in "video" mode (rather than film), this is capable of determining resolution - and so is a good estimate of how well "motion adaptive" and "motion compensated" video mode deinterlacing algorithms work. DCDi adds an extra edge as it detects jaggies and smoooths them out. With most other algorithms, this is a good demonstration of bad deinterlacing vs. good deinterlacing algorithms.
27) I have a XXX projector, what is the best scaler for me?
28) What will the scaler "do for me"?
29) What is native rate ("sweet spot")?
Answer provided by Jon Lindgren:
All projectors have a scan rate at which they operate efficiently and produce excellent visual results. For LCD, DLP and other fixed pixel projectors, this rate will be the resolution of their panels (i.e. 1024x768, 1280x720, etc.). For CRT projectors, however, this rate changes depending on the model, its condition, and so on. Note that the sweet spot includes not only resolution, but also update rate - 1536x1152@72Hz may look terrible on an example projector, while 1536x1154@60Hz may look fantastic.
30) Do I need test patterns on my scaler?
31) Are all SDI players compatible with all SDI capable scalers ?
33) What is/are 480i, 480p, 1080i? What do these numbers mean?
Answer provided by Jon Lindgren (ref):
These numbers represent different types (resolutions and rates) of video
signals. In general a higher the number indicates a higher resolution. Note that
a higher resolution does not necessarily mean a better picture, however, as the
capabilities of a signal are only one part in the chain of a good video display
system.
The first part of the term represents the number of vertical scan lines per
frame: 480i has 480 scan lines in each frame, 540p has 540 scan lines per frame,
and 1080i has 1080 scan lines in each frame. Implicit in the number of vertical
scan lines is the number of horizontal pixels: 480 implies a resolution of
720x480, 720 implies a resolution of 1280x720, and 1080 implies a resolution of
1920x1080.
The second part indicates if the signal is interlaced (indicated by an "i")
or progressive (indicated by a "p"). Thus, 480p is a signal which
carries 480 vertical scan lines in a progressive format. 1080i is a signal which
carries 1080 vertical lines in an interlaced format.
Sometimes, you will also see a reference to 480p/24 or 480i/60. This third term
referes to the number of frames (in the case of a progressive signal) or fields
(in the case of an interlaced signal) per second. Thus, a 480p/24 signal has 480
vertical scan lines, sent as 24 full frames per second, whereas 1080i/60 has
1080 vertical scan lines sent as 60 fields per second. The reason for indicating
the number of frames/fields per second is that many systems have different scan
rates: while NTSC typically has 60 fields per second, PAL has 50 fields per
second. Unusual rates such as 24 or 48 frames per second are used by high end
equipment to mimic the frame rate of film (which runs at 24 or 48 frames per
second). See the questions on "jutter" for an explaination as to why
24fps can be better under certain circumstances.
Note that many higher rate signals (all progressive formats, and 720 and higher
rates) cannot be carried by s-video or composite signals, but only by higher
bandwidth methods such as RGBHV, DVI or component.
Some common terms you will see are:
480i - this is standard video in the US. s-video connections transmit this type
of signal, and regular non progressive DVD players with component out will
output this type of signal.
480p - this is standard progressive video, commonly found on progressive DVD
players. Sometimes reffered to as 480p/60 and 480p 60fps.
540i - this is "standard" video for PAL systems.
540p - this is "standard" progressive PAL video.
720p - this is a common HDTV resolution, having a resolution of 1280x720.
1080i - this is another common HDTV resolution, somtimes referred to as 1125i.
34) Can my TV/Projector use a scaler?
Answer provided by Jon Lindgren (ref):
Most newer digital TVs have some sort of deinterlacer and/or scaler built
into them. Most older TVs do not have a deinterlacer or scaler built in.
The first thing to find out is if your TV accepts a deinterlaced input. To
accept this, the TV must either accept some form of an RGBHV signal (i.e.
640x480, 1024x768, 848x600 all of which are progressively scanned) or a
progressive component signal (i.e. 480p, 540p, 720p). If your TV only has
composite, s-video or RF inputs, it will not support a deinterlaced input.
For most projectors, you will have to determine if your projector can support
the frequency needed. To display 480p, your projector must be outfitted with
RGBHV or component connections, and able to accept a signal with a 31.5KHz
bandwidth. Similar requirements exist for 720p, 1080i, etc... If your projector
only has an s-video or composite input, you can't use a deinterlacer or scaler.
A second consideration after you've determined that your TV/projector can accept
a deinterlaced signal is "will it look better?" Understanding what a
deinterlacer/scaler does, and applying that technology correctly will provide
you with an answer to this question. Not all TVs and projectors need
deinterlacers, while some absolutely require one. In general, the more advanced
the display, the more revealing of the true video signal it will be, thus the
higher the requirement for some sort of an excellent scaler/deinterlacer.
35) What are "Motion adaptive", "Motion compensated" deinterlacing algorithms?
Answer provided by Allan Jayne (ref):
It so happens that one excellent way of de-interlacing weaves material from
the next (and/or previous) field for that part of the picture representing
non-moving subjects, while interpolating to obtain portions of intervening scan
lines for moving subjects. A motion adaptive de-interlacer analyzes the video in
terms of pixels to try to find the moving subject matter and thus perform this
method of de-interlacing. Ideally the motion adaptiveness switches between bob
and weave dozens of times for every scan line, while less sophisticated versions
have the picture divided up into zones like a checkerboard, using either bob or
weave for the entire content of each square. At least one de-interlacer chip
either bobs or weaves the entire frame and the keen eye sees the entire picture
soften a bit when something starts moving.
One comb filter decoding strategy also relies on a distinction in the video that
corresponds to moving versus stationary subject matter across consecutive
fields. Comb filters using this strategy are referred to as motion adaptive.
Motion compensation in a de-interlacer attempts to guess where the subject moved
to in the next field. It reconstructs a version of the next field with the
subject "moved back to where it was in the previous field" and then
weaves it as the intervening scan lines to make a progressive scan frame out of
the previous field. This, too, is done by analyzing the content in terms of
pixels. The advantage of motion compensation is that the bob method
(interpolation) does not have to be used as much, and therefore there are fewer
areas where the resulting loss of vertical resolution occurs.
Neither motion compensation nor motion adaptive algorithms today are perfect so
some artifacts such as combing will usually be seen every once in awhile.
Answer provided by Ofer LaOr (ref):
8/8a. Actually, we need to identify all the known scaler artifacts (8a is a good start) and explain each one individually.
Answer provided by Jon Lindgren (ref):
If a deinterlacer or scaler has problems or is mismatched with your display
or source, it can produce one of a number of artifacts:
Judder: this happens when one frame/field rate is not easily convertable into
another frame/field rate, or when frames/fields are not shown for equal amounts
of time. Consider 4 frames of film (A, B, C and D) which are from a scene where
the camera is panning off to the right. If one presented the frames in this
order:
A B C D
The motion will be smooth. If, however, some frames are added:
A A B C C D
Some judder may occur because two of the frames (A and C) are being shown for a
longer time than the other two (B and D). This is one reason some high end
equipment will present film material in a 24fps, 48fps or 72fps rate. Since film
is 24fps, it's easy to repeat each frame the same number of times:
24fps: A B C D
48fps: A A B B C C D D
72fps: A A A B B B C C C D D D
When showing film using 60 fields/sec (480i, for example), the fields of each
frame are staggered in what's known as the 3:2 cadence, where some fields are
shown for longer than others (a1 and a2 represent the two fields of frame A, b1
and b2 represent...). This is part of the telecine process used to convert
film's 24fps to video's 60 fields/sec:
60 frames/sec: a1 a2 a1 b1 b2 c1 c2 c1 d1 d2
Looking at this example, it's easy to see that A's fields are shown longer than
B's fields. This may produce judder.
Note that judder can also be created by the display device. Certain technologies
(such as some LCD and DLP projectors) have a fixed rate at which the panel can
refresh itself. Most LCD projectors, for example, are fixed at 60Hz. If the
projector is presented with a 72Hz signal, the projector must drop or
interploate 12 frames or fields (depending on the signal) per second to fit the
72Hz signal into the 60Hz refresh rate of its panel. The act of doing so can
cause certain frames to be displayed longer than others, which can produce
judder.
Sometimes, bad programming or bad firmware (on the deinterlacer/scaler or
display device) can also cause judder.
Judder is most commonly seen one slow, smooth camera pans.
Tearing:
This artifact looks like you get two different frames on the screen at the same
time, one on top with the other on the bottom. Essentially, what happens is that
the image being displayed is changed midway through the process of displaying
the frame. This causes the first part of the image to be from one instant in
time, while the latter part is from a later instant in time.
This is generally caused by one of two things: a mismatch of signal rates (as
described in the latter part of "judder") or bad programming/firmware.
Combing:
This effect is seen when small horizontal "ridges" form on the left
and right edges of objects in motion. This is primarily caused by the
deinterlacer's inability to merge two fields together properly (as is the case
with most video mode sources). Due to the fact that one field is from a given
point in time, and the second field is from 1/60th of a second later, the two
fields can not be merged into a single frame perfectly.
It is also caused by viewing interlaced material without a deinterlacer. In this
case, each field can plainly be seen. Certain camera movements make combing
quite obvious.
Combing is most commonly seen one slow, smooth camera pans, although it is
visible elsewhere depending on the viewer. Combing is always horizontal in
nature, never vertical. It is not easy to see on smaller TVs (i.e.
>20"), but is usually easy to pick out on larger displays which do not
deinterlace.
Answer provided by Allan Jayne (ref):
Jaggies (jaggedness; stairstepping) are always present on diagonal lines or edges when a picture is represented as scan lines or pixels. For mediocre de-interlacing, plain line doubling or outputting each scan line twice being a reasonable worst case, jaggedness is coarser than the scan line spacing.
Answer provided by Jon Lindgren:
Luma (also referred to as "Y") basically refers to the black and white portion of an image, while chroma (also referred to as "C") refers to the color information. In the beginning of television days, all images were black and white (hence the bulk of the TV signal was luma information). Once color TV was announced, a second piece of information (the chroma) was added on top of the black and white signal, mostly to maintain compatibility with the existing black and white TVs of the day. Together, luma and chroma give you a color image. Note that there are subtle differences between luminescence and luma, and between chrominescence and chroma (one is a mathematical transform of another, taking gamma into account). For the purposes of this FAQ, we will simply refer to luma and chroma.
39) What is the difference between composite, s-video (Y/C) and component connections?
Answer provided by Jon Lindgren:
From question 38, we know that a color signal is comprised of two pieces of information: the luma information, and the chroma information.
In a composite signal, the chroma information is mixed with the luma information (and is separated by a comb filter within the display or processor). While this means that the resulting signal is easy to carry (only one connection is needed), it also means the signal has bandwidth constraints. Also, separating the luma and chroma information is not easy, and can lead to artifacts. Although advanced comb filters have been designed (such as 3D comb filters), it is still difficult to do perfectly. Composite connections are found on almost every TV, VCR, Laserdisc and DVD player. A composite connection can not carry high definition video.
In an s-video signal, the chroma information is sent on a separate pair of wires from the luma information. This means that the two signals don't need to be separated (so no comb filter is needed on s-video connections), and allows for a higher bandwidth. The result is a clearer picture. Although an s-video connection is a single plug, there are 2 pairs of wires run through it (one for luma, one for chroma). This type of connection is also called a "Y/C" connection (Y=luma, C=chroma) because the signals are not combined, but transmitted separtely. An s-video connection can not carry high definition video, but is a large improvement upon the picture quality available with a composite connection.
Component connections take the luma/chroma separation one step further, and actually separate color information into primary colors. The connection consists of a Y connection (again, the luma signal), a Pb connection (the percent of blue in the signal), and a Pr connection (the percent of red in the signal). Due to how the human eye reacts to images, the Y (luma) channel is actually the amount of green in the signal, giving us three color channels of R, G and B. The main advantage of a component connection is higher bandwidth - it is possible to carry high definition video over component connections. This type of connection is also called YPbPr and YCbCr (although the latter is incorrect, and actually refers to the component connection in a digital form; it is, however, used interchangeably in the consumer markets).
40) Why is a component connection superior to other analog methods?
Answer provided by Jon Lindgren:
The main advantage a component connection has is its bandwidth, which allows it to carry high definition video. Color information is also more accurate when compared to s-video connections.
41) What is the difference between component and RGB[HV]?
Answer provided by Jon Lindgren:
The main differences are the electrical format and the color space. Electrically, slightly different levels are used to indicate different color intensities. Different color spaces are also used (similar to how one might use red, green and blue for your monitor, but cyan, yellow and magenta for your printer). They are easy to convert from one format to another using a device called a transcoder.
Answer provided by Allan Jayne:
Component video is approximately red, white, and blue! Also for component video, sync. signals are usually carried with the luminance (white; Y) while for RGB, horizontal and vertical sync. are often carried separately on fourth and fifth lines but may also be combined with the green.
Any three colors could have been chosen as the "color space" or three "components" from which red, green, and blue must ultimately be obtained or derived; however Y, Pb, and Pr are very easy to work with. By varying the intensity (as the electrical level) for each component, all other colors are represented. In addition, Y, Pb, and Pr best fit the video model for the best picture quality for a given bandwidth. The luminance carries the maximum picture detail requiring "full" bandwidth while the other two components use less bandwidth taking advantage of the human eye's lesser sensitivity to loss of color in fine details. Using RGB, for example, would require full bandwidth for all three colors, making it more difficult to record video on DVD or transmit it even as HDTV. A device called a transcoder is used to convert from one color space to another.
42) What connection should I use from my DVD/Laserdisc/VHS/S-VHS player or cable/DSS box?
Answer provided by Jon Lindgren:
To answer this, it is necessary to figure out how the device stores video signals. Note that using s-video or component isn't always better, depending on how the device stores and plays your video:
Laserdisc players are composite in nature - the video signal is digitally stored as a composite signal. If you believe or know your player to have a superior comb filter, then try using your laserdisc's s-video output. If unsure, use the composite. Chances are that the comb filter on your video processor is better than the comb filter in your laser disc player, so keep that in mind.
VHS and S-VHS players differ laserdisc by storing video as separate luma and chroma (Y/C), and thus benefit from using an s-video connection. If the composite connector is used, the output signal from the tape must be combined from a separate luma/chroma signal to a composite signal. Usually only S-VHS units provide an s-video connection, but if your regular VHS machine has it, then give it a try.
DVD players are native component devices in nature. On the DVD itself, the image information is natively stored in a component colorspace. Component video connections are a natural medium to use with this device. If your processor/scaler or display doesn't support component inputs, try the s-video connection. Use the composite connection only as a last resort.
Cable boxes and DSS boxes are very different than any of the above. Certain devices are best used on different types of connections. For instance, the Scientific Atlanta 3100HD box upconverts non-high definition material to 1080i when viewed through its component connection, but it doesn't do a very good job of it. On this box, standard definition material is best viewed through the composite or s-video port, while high definition material is only available via the component connections. Your best bet is to ask around to find out what is best used under each situation.
43) Do I need a video processor?
Answer provided by Jon Lindgren:
The short answer is this: if you see video artifacts, then you may benefit from a video processor. Video artifacts usually become bothersome as the viewed image becomes larger, as video was never designed to be blown up to large sizes.
Many projectors, RPTVs and plasmas have some sort of scaler and/or deinterlacer built into them. In some cases, they are excellent quality (for example, some projectors include electronics from Faroudja or Silicon Image, both of which are excellent). In most cases, however, the integrated video processing is rather lackluster, so most large venue setups will benefit from some sort of video processing. What you need depends on what type of display you have, and whether you are bothered by any atrifacts you see.
If you see combing, jaggies, pixelation, or other artifacts, chances are a good quality video processor will help resolve them to some degree.
Also, remember this: a video processor cannot add information to the signal, so if you have a poor source to begin with you will most likely have a poor quality signal to end with. The old adage of "garbage in, garbage out" fits well here.
44) Do I need a scaler, deinterlacer, or doubler?
Answer provided by Jon Lindgren:
While a scaler, deinterlacer, doubler, tripler, etc.. are all video processors, they fall into different categories and are used for slightly different applications.
Scalers are typically used with fixed pixel devices, such as LCD TVs, plasma displays, and fixed pixel projectors. As all of these devices have a fixed grid of pixels which the video image must be mapped to, a scaler is the perfect fit for these types of displays. Line triplers and quadruplers usually do not match the frequencies of fixed pixel devices, so they are generally poor choices for fixed pixel devices.
Traditionally, doublers and their kin (triplers and quadruplers) are used with CRT products. The main reason for this is that CRT devices do not have a native pixel structure, so they aren't in need of a scaler. Note that this is not always the case - people do use scalers for CRT products with excellent results, as a scaler can help more accurately hit the "sweet spot" of a CRT projector.