A/V versus RF and actor/keyword searches

A couple of quick questions, and yes I have tried to do my homework
first and searched the web without finding any answer to these, maybe
I just didn't know what to search for.

Any advantage or disadvantage of using A/V cables versus RF cable
from the cable box to the Tivo box? The TV only has RF input and
I only record at basic quality. But I'm wondering if there are
any quality differences, power consumption differences or ?

Any way to get Wish list items to dig a little deeper? Example,
guy named Michael Persinger who is interviewed on lots of science
and news shows. But he doesn't show up in the list of actor
names, keyword search for his last name never matches. Any way
to get matches for something like this to catch his interviews?

Thanks

Don Taylor wrote:

Any advantage or disadvantage of using A/V cables versus RF cable
from the cable box to the Tivo box?

Major differences in picture quality.

The TV only has RF input and I only record at basic quality.

You're sort of out of luck there.

But I'm wondering if there are any quality differences,

Regular A/V (red+white+yellow RCA jacks) provides significantly
better video quality than using channel-3 RF, and has stereo audio.

S-Video (red+white RCA + black S-Vid jacks) has even better
video quality. In particular, none of the "dot crawl" seen along
brightly colored vertical lines.

power consumption differences or ?

No difference there.

Any way to get Wish list items to dig a little deeper? Example,
guy named Michael Persinger who is interviewed on lots of science
and news shows. But he doesn't show up in the list of actor
names, keyword search for his last name never matches. Any way
to get matches for something like this to catch his interviews?

Find one of the shows he's on, look at the show description, then
press Info (or Enter) to get more details. If his name does not
show up there, then that information is not being provided to the
Tribune Media Services and therefore is not available to TiVo
to match.

-Joe

P.S. Check out www.zap2it.com - it also uses the TMS data.

Any advantage or disadvantage of using A/V cables versus RF cable
from the cable box to the Tivo box? The TV only has RF input and
I only record at basic quality. But I'm wondering if there are
any quality differences, power consumption differences or ?

Give the Tivo the best possible signal and you'll get the best possible
results. If you feed only RF channel 3 into the Tivo you're putting it at a
disadvantage when it records it. Give it s-video out of the cable box and
you at least give it a better chance of getting the best possible picture.
Even using basic recording rates it helps to avoid garbage in, garbage out.
Will you notice it? Possibly. But the difference in cabling costs is
trivial so why not go for the best chance?

Any way to get Wish list items to dig a little deeper? Example,
guy named Michael Persinger who is interviewed on lots of science
and news shows. But he doesn't show up in the list of actor
names, keyword search for his last name never matches. Any way
to get matches for something like this to catch his interviews?

You're entirely at the mercy of how accurate and detailed the data is in the
guide. If it's not mentioned in the show's Info screen you won't find it in
a wishlist. I've not seen most science shows go so far as to detail who's
on the program. Most of the time it's subject matter oriented. You'll have
to search elsewhere on the web and then set up the Tivo. But if it's in the
guide you can use wishlists to find it. If you want better guide data
complain to Tribune and the network broadcasting the show.

-Bill Kearney

Any way to get Wish list items to dig a little deeper? Example,
guy named Michael Persinger who is interviewed on lots of science
and news shows. But he doesn't show up in the list of actor
names, keyword search for his last name never matches. Any way
to get matches for something like this to catch his interviews?



Anybody who wastes their time on the pseudoscience of that man has my
sympathy. No reputable scientist gives any weight to testimonial
evidence.

Bill Kearney wrote:

Give the Tivo the best possible signal and you'll get the best possible
results. If you feed only RF channel 3 into the Tivo you're putting it at a
disadvantage when it records it. Give it s-video out of the cable box and
you at least give it a better chance of getting the best possible picture.
Even using basic recording rates it helps to avoid garbage in, garbage out.
Will you notice it? Possibly. But the difference in cabling costs is
trivial so why not go for the best chance?

This advice is certainly true......although it has ever-less significance for
us, the end users or consumers of video content.

Since the days of the transition from analog to digital satellite transmissions,
we consumers are left at the mercy of the quality the "technological
middlemen" choose to give us. And they *always* opt for quantity vs quality.
This applies equally to cable-provided video fare, as virtually every
cable provider picks up digital transmissions at their head-ends and may
only provide an analog distribution network. And even they are all
transitioning to digital transmission to end users.

So these days, the quality constraint is at the source, so to speak, and
*not* in the hardware at the consumer's end. "Statistical Multiplexing" is
the 'Holy Grail' in this new digital era in which we now live, and this is
the 'physics' which today limits, for the most part, the quality you'll see
on your screen, regardless of what hardware you use in your home.

Here's a primer which describes, in overview form, exactly what I'm talking about:

DSS Technical Fact Sheet
revision 0.1
August 27, 1995

Contents:
1. Advantages of MPEG-2 over MPEG-1
2. Video quality
3. "Resolution" of DSS
4. Decoders produce identical video
5. Bits per program channel
6. Aspect ratio
7. Frame rate of MPEG
8. Statistical multiplexing
9. Composite vs. component
10. Artifacts
11. Deriving channel count
12. Number of satellites
13. DSS and standards



1. Advantages of MPEG-2 over MPEG-1



MPEG-2 offers little benefit for video programs which
originate from film. Most movies and television programs
with high production values (budgets) are shot on 24
celluloid frames per second.



MPEG-2 provides the ability to more efficiently represent
interlaced video signals generated from such devices as tube
or CCD cameras. MPEG-1 was primarily intended for
progressive video.



However, MPEG-2 encoders may possess more mature and
powerful coding techniques that benefit progressive video (MPEG-1).



2. Video quality



Video quality is a function of:



1. source material perceptivity: how much information
(motion and spatial details) are contained within a video
sequence.



2. encoder quality: how well the encoder applies MPEG
syntax to distribute bits throughout the video sequence.



3. bit rate: the higher the bitrate, the higher the
quality (as per Rate-Distortion theorem).



Laserdisc picture quality for complex film source video
(movies) can be approximated at around 4 Mbit/sec.
Interlaced signals can be represented transparently around 6
Mbit/sec. Studio quality video starts at 8 Mbit/sec.



3. "Resolution" of DSS



Most DSS video programs are encoded at 544 pixels/line and
480 lines/frame. 24 or 30 frames/sec are coded, depending
on whether the video contains progressive material (film) or
interlaced.



MPEG-2 Main Profile @ Main Level video decoder chips are
capable of reconstructing pictures of just about any size,
as long as the dimensions are restricted to 720 pixels/line
and 576 lines/frame.



However, for Display purposes, the pixel dimensions are
coded at one of a small set of sizes:



Horizontal: 720, 640, 544, 480, and 352 pixels per line
Vertical: 240 or 480 lines/frame.



The DSS uplink facility in Castle Rock, Colorado currently
extracts its signals from traditional analog NTSC satellite
C-band channels much like cable television plants do all
over the continent. Since the positive lobe of these analog
NTSC signals are bandlimited to roughly 5 MHz, a sample
rate of 544 pixels/line encapsulates full detail.



Sampling rate is not a direct measurement of video quality.
On the other hand it does provide an upper bound to the
amount of detail that can be present in the quantized video
signal. Equivalent "resolutions" for various media are
presented below:



Line density Samples/Line
Cable & Broadcast TV ~335 ~450
Laserdisc ~425 ~560
VHS ~250 ~330
MPEG-1 (SIF class) 264 (upper limit) 352
MPEG-2 ([email protected]) 540 (upper limit) 720



Samples/Line is derived as 4/3 * Line density. Line density is
the traditional figure given for TV monitors and storage formats.



[Note: Laserdisc's advertised "resolution" of 425 lines may be a
more liberal interpretation of the usual -3 dB bandwidth limit point.]



4. Decoders produce identical video



All DSS decoder boxes manufactured by Thomson and Sony have
been fully capable of decoding MPEG-2 video bitstreams from the onset.



It is the encoders at the uplink facility which determine
whether a bitstream is MPEG-1 or MPEG-2.



MPEG compliance implies that any two decoders shall produce
numerically identical pictures when fed the same bitstream.
Some small statistical variances are permitted in the
Inverse Discrete Cosine Transform stage of the video
decompression pipeline, but these should not account for
more than an occasional least-significant-bit of discrepancy
in the final reconstructed video stream. However, encoders
can significantly reduce bitstream patterns which cause
disrepancies.



The Display Process, which is outside the conformance scope of
MPEG, may introduce additional and far more apparent discrepancies
between decoders systems. For example, how two different
decoders convert MPEG's native component YCbCr 4:2:0 video signal
into the output 4:2:2 component Y-C (S-video) or Composite NTSC
signal may differ.



If there are any differences in perceived video quality
between two boxes, it is mostly likely the result of the
NTSC video generator chip.



5. Bits per program channel



Currently, Direct TV assigns about 3 Mbit/sec towards the
video bitstream of Pay Per View programming. Some sports
channels receive the full 6 Mbit/sec. Most other
programs are coded at 4 to 5 Mbit/sec.



The 240 watt/transponder DSS bitstream has a payload rate of
30 Mbit/sec. Six programs share the bitstream. For example:



2 pay-per-view movies @ 4 Mbit/sec
2 variety channels @ 5 Mbit/sec
2 sports or high priority channels @ 6 Mbit/sec.
-------------------------------------------------
Total 6 programs @ 30 Mbit/sec.



6. Aspect ratio



MPEG video decoders are capable of handling a variety of
aspect ratios, often eliminating the need for letterboxing
in the source video.



Source video can be coded at its native aspect ratio,
leaving the decoder box to perform the necessary format
conversion for the target display device.



For example, a movie produced in 2:1 Cinemascope can be
represented with (anamorphic) 720 pixels/line x 480
lines/frame. For a 4:3 display the MPEG decoder will
perform an internal Pan & Scan operation by extracting 480
pixels/line ( 4/3 * 1/2 * 720 = 480) followed by a scaling
of the 480 pixels through a simple interpolation process to
match the display rate of, e.g. 720 pixels/line.



A 4:3 display need extract 544 pixels/line (4/3 * 9/16 * 720
= 544) from a program encoded at maximum resolution
(720x480) for 16:9 displays.



Thanks to Pan & Scan operations embedded within MPEG decoders
and/or display devices, potential vertical detail need not be
lost to the black horizontal bars of letterboxing.



7. Frame rate of MPEG



Currently, a legal [email protected] video bitstream can be only 23.976,
24, 25, 29.97, or 30 frames/sec.



MPEG decoders have a built-in frame store which permits the
decoder to perform 3:2 pulldown itself, eliminating the need for
the encoder to include redundant fields or frames in the
bitstream. For example, a progressive video bitstream coded
at 24 frames/sec will be mapped to the constant interlaced
30 frame/sec Display Rate by replicating every 4th coded
field. (5/4 * 24 frames/sec = 30 frames/sec).



All 24 or 30 frames/sec are coded in legal MPEG bitstreams.
None are skipped or interpolated. Frames are not homogeneous.
They are broken into non-overlapping 16x16 pixel regions
known as "Macroblocks." Each macroblock can arbitrarily be
coded from scratch or use 16x16 regions from previous
decoded pictures stored in the frame buffer as reference.
Additionally, macroblocks which employ reference pictures can
add refinement detail.



Refinement information may further propagate into future frames,
reinforcing the strength of the video signal.



8. Statistical multiplexing



Statistical multiplexing allows video programs multiplexed
onto a common data carrier to share bandwidth. For example,
when a scene change occurs in one sequence, it is assigned a
high instantaneous bit allocation priority. A scene in another
program at the same time with limited motion would be assigned a
lower priority.



The coding gain of statistical multiplexing is, by nature,
statistical. If all the video programs of a common carrier
exhibit a strong thirst for bits at the same time, then little
statistical multiplexing is of little benefit.



Research groups have reported gains between 1 - 2 dB (10 to 30
percent).



The new Digital Video Disc (DVD) format lead by Toshiba and Sony
acheives a more predictable form of statistical multiplexing.
Since the sustained data delivery rate of the high density Compact
Disc itself can be controlled within a range of 1-10 Mbit/sec,
the bit allocation can be spread or normalized over the entire
program length (e.g. 135 minutes for a movie). The 135 minute movie
will average 5 Mbit/sec, but can for a necessary time sustain 8
Mbit/sec
during difficult scenes. The encoder must pre-analyze the frames
which correspond to the time period in which the bitrate is averaged
over.



9. Composite vs. component



MPEG video signals are exclusively component YCbCr (CCIR Rec. 601).
This permits programming to remain component over the entire
signal path (film to TV monitor).



The video signals encoded onto Laserdiscs are in fact
composite analog NTSC (or PAL for Europe) and are therefore
subject to the traditional crosstalk artifacts.



10. Artifacts



There are two completely different classes of compressed digital
video artifacts: bitstream error artifacts and coding artifacts.



Bitstream error artifacts result when the signal received by
the dish becomes too weak. These artifacts include: picture
freeze, black or strangely colored blocks throughout the picture,
audio loss. You can create bit error situations by gradually
obstructing the dish until the noise threshold is reached, beyond
which no video or audio completely disappears.



Coding artifacts including ringing along edges of objects
and blockiness. Unlike the bit error artifacts above, all decoder
boxes see these artifacts the same. In theory these artifacts should
never exist in the first place. It is the responsibility of the
encoder to balance video content with bitrate.



DSS is currently plauged by an artifact most prevelent during
rapid lighting changes (scene fades). This is a result of the
encoder intelligence, not MPEG, and will be corrected with time.
Cost is a major factor behind slow upgrades: like most
high-end MPEG-2 broadcast encoders, each CLI Magnitude encoder
at Castle Rock (one per video program) costs around $100,000.



11. Deriving channel count



DSS is comprised of 32 transponders, each with an average of
6 programmes. Not all 32 transponders are used for video
programming (e.g. the 30 musical channel service).



The usable bandwidth of the high powered Ku band spans a
frequency range of around 450 MHz. It is a North American
convention to space each transponder every 27 MHz, which in
analog terms corresponds to one television channel. This
yields a total of 16 transponders (450 MHz/27 Hz/transponder
= 16). When broadcast from space, the Ku band's microwave
signal is coherent enough for the same frequency space to be
used twice: each at opposite polarities. Hence the
transponder count is doubled to 32.



A spacing of 27 MHz may seem somewhat liberal when
considering that cable television and terrestrial television
channels are spaced every 6 MHz. However, to keep the power
levels low, analog video satellite signals trade off
bandwidth for power through the method of wide band FM
modulation. This permits a wideband FM modulated carrier
of, say, 15 dB to convey a narrow baseband signal of more
than 45 dB.



One DSS carrier is modulated per 27 MHz transponder channel
at a rate of approximately 20 million symbols/sec. The 4-
QPSK modulation scheme assigns 2 bits per symbol, yielding
40 Mbit/sec per carrier. The carrier-to-noise ratio needed
to achieve nearly error-free decoding is a little less than
10 dB.



Out of the total 40 Mbit/sec, only 30 Mbit/sec carry usable
payload data. The remainder 10 Mbit/sec is comprised of
various channel coding overheads such as Reed-Soloman error
correction bytes and Viterbi codes. The lower-powered 120
Watt transponders have a payload rate of only 23 Mbit/sec.



Compact Disc itself has a channel rate much higher than the
effective payload rate reserved for the 1.5 Mbit/sec digital
audio bitstream. 0.5 Mbit/sec are consumed by Reed-Soloman
bytes, and another 1.5 Mbit/sec are consumed for EFM coding.



12. Number of satellites



The number of satellites needed to provide DSS is purely an
implementation issue. The power requirements for a 32
transponder DBS service are too great for the solar panels
of a single conventional space craft. International
Agreements (UN, et al) limit the signal strength (240
Watts/transponder), frequency use (spacing of 27
MHz/transponder including 3 MHz of guardbands, etc), and
spacing (9 degrees) of the satellites along the geosationary
orbital belt (a.k.a. "Clarke Belt").



Therefore it is more accurate to say: "There is a potential of
32 transponders, each with X Mbit/sec of channel capacity,
per ORBITAL SLOT."



Currently, DSS is already in possession and use of all 450
MHz of allotted Ku band space. A few transponders still
remain at the lower 120 Watt level prior to the activation
of the 3rd satellite, so there is potential of raising the
overall channel count by another 25% or so (whilst keeping
the same average bitrate per program).



13. DSS and standards



DSS makes use of several standards, but is not fully
compliant with the European Digital Video Broadcasting (DVB)
satellite convetion. DSS was designed before the
ratification of DVB.



DVB specifies the channel coding methods (e.g. symbolic
rate, QPSK modulation, Reed-Soloman error correction,
Viterbi, packet interleacing), transport layer (MPEG-2
Systems Transport bitstreams), and elementary stream layers
(MPEG-2 Video, MPEG-1 Audio).



DSS elements are almost identical with the exception of
subtle details such as transport packet length (136 bytes
vs. MPEG-2's 188 byte).



The DVB spec also addresses cable and terrestrial
broadcasting.


--
Valentin Guillen

http://www.thuntek.net/~vguillen