| Film Fanatic May 2004
Taking a Break with Some Definitions
In this month’s "Film Fanatic" I
identify some language I’ve been throwing around in this column and in my DVD reviews
on Home Theater & Sound. This is in no way an exhaustive list; I neither want
to overload you nor make you feel you’ve signed up for a course in rocket science.
What you’ll gain is a high-level view of some complicated terminology in
layman’s terms. I won’t get into the depth that would allow you to talk shop
with an engineer, but it should be enough to take away the mystery.
DVD (Digital Versatile Disc): A DVD is a storage
medium for binary data (0 = off, 1 = on). In this regard it’s not unlike a floppy
disk, a Zip cartridge, or a computer’s hard drive. The difference between DVD, CD,
hard drives, and their floppy-disk cousins is in the way information is written to and
read from each medium. In the case of a floppy or hard drive, binary language is expressed
in magnetic particles that are arranged via a read/write head on a metal
"platter."
In the case of a manufactured DVD or CD, a laser reads
"bumps" on a polycarbonate disc coated with aluminum. This is similar to what a
phono stylus does when it tracks the grooves of an LP. In the case of recordable CD or DVD
(CD-R, DVD-R), instead of bumps, the playback laser reads differences in intensity of a
photosensitive dye that darkens when heated by the write laser.
Compression: Compression is a process of reducing
the number of data. The main goal of compression is to reference multiple redundant
data patterns with a single value. Data compression is typically done to either
increase the transfer speed of data for copying or moving, or to fit more data on a
specific storage medium, such as a DVD.
There are many different CODECS (COmpress, DECompress) --
algorithms that downsize various types of data. Some reside in software applications, such
as Apple’s QuickTime video player or Window’s WinZip file compressor, while
others are embedded in a silicone chip -- for example, IBM’s S-series MPEG encoder.
MPEG-2 Encoding: MPEG (for Moving Pictures Experts
Group), specifically MPEG-2, is used to fit picture data onto DVD. The MPEG-2 encoding or
recording process analyzes picture data from a group of pictures (GOP) by comparing the
pixels of an intraframe (the I frame) to the pixels of a predicted frame (P frame) of
video and interpolating (estimating) the value of a middle, bi-directional B frame. MPEG-2
compression works both within an image and over the course of a GOP. Redundant pixels are
replaced by single values.
During encoding, most of the compression occurs in the B
frame; less compression is applied to the P frame. The I frame is encoded by itself onto
the DVD without any comparison to other frames. This is the reference frame. On decoding
or playback, the P frame is reconstructed with information from the I frame. The B frame
is then reconstructed using data from the I and P frames. The amount of compression
applied to the GOP structure is up to the discretion of the compressionist.
Compressionist: This job is more art than science. A
compressionist needs to have a solid understanding of what elements within an image can
give the encoding process a hard time and the settings that will make a specific image
look its best. The well-trained highly experienced compressionist balances the needs of
the storage medium with the aesthetics and integrity of the image.
Compression Artifacts: CODECs aren’t perfect.
Unless you’re encoding every single bit of data in an image and decoding the same
number of bits on playback, there will be errors in the way a CODEC re-creates the image.
These errors are visible artifacts or flaws in the image. Because the goal of a compressor
is to reduce the number of data by simplifying redundant information, unique information
has a tendency to drive it nuts. High frequencies that are too high for the encoder’s
sampling frequency, or random noise caused by substandard electronics, are the banes of
compressors. Film grain and highlights can cause high-frequency artifacts that appear as
shimmering edges (called aliasing or edge stair-stepping), dancing pixels,
and bleeding color transitions.
Low-Pass Filtering: One way of reducing compression
artifacts is to run picture data through a low-pass filter before MPEG encoding. The video
signal is sent through electronics that let low frequencies pass through while attenuating
the high frequencies. As a result, the high frequencies, where many of the unique data
live, are removed. Like any form of filtering, this process can be abused. Applying too
much low-pass filtering to a signal can remove necessary high-frequency information that
gives definition to edges and important details. When too much filtering is applied,
images look overly soft and lack sharpness. What may have been a raindrop or birthmark on
an actor’s face turns into an indefinable glob of pixels.
Edge Enhancement: Edge enhancement is a form of
processing meant to compensate for the lack of detail in soft images, and is very similar
in function to a TV’s Sharpness control. It creates artificial detail that isn’t
native to the image but is an electronically fabricated edge that can cause some unwanted
side effects, such as halos or ghosting around the outer edges of an element within the
frame. These side effects are typically more apparent on larger video screens.
Anamorphic/Widescreen: The decision to shoot films
in anamorphic or widescreen aspect ratios was made in the 1950s to differentiate the
cinema experience from the then-new television experience. Before TV, films used the same
1.33:1 or 4:3 (fullscreen) aspect ratio -- that is, the image and screen are in a ratio of
four horizontal units to three vertical -- that we’ve come to think of, in the past
half century, as "TV." Widescreen typically comes in two ratios: 2.35:1
and 1.85:1. Anamorphic refers to the geometric distortion used within a
camera’s lens when shooting a film, which allows wide-angled pictures to be shot on
4:3 film stock. This distortion is corrected by a complementary anamorphic lens in a
theater’s projector, to fill the theater’s wide screen. If you watch the end
credits of some older films, you’ll see how an anamorphic image looked in a 4:3 frame
without optical correction. The image is elongated; characters and elements within the
frame look unnaturally tall.
Early DVDs suffered from non-anamorphic transfers, which
resulted in improperly proportioned widescreen images and a reduction in the image’s
horizontal resolution. DVDs labeled "Enhanced for Widescreen" or
"Anamorphic" are encoded so that DVD players will display the image in properly
proportioned widescreen format with optimal horizontal resolution.
The end for now
This list just scratches the surface of how your
home-theater system converts into pictures all those ones and zeros that reside on your
DVDs. It should get you thinking about what goes on inside your DVD player every time you
play a favorite film.
...Anthony Di Marco
anthony@hometheatersound.com |
