7
DVD Technical Notes
Video
One potential way to cut corners in 525P output is to adopt
relatively inexpensive frame memory. However, simple frame
memory systems can expose the signal to motion blur on two out
of every five frames. This becomes clear when we review the 3-
2 pulldown process. In simple frame memory reverse conversion,
the player combines the present field with the previous one — no
matter what cinema frame it came from. If the previous field
came from the same cinema frame, all is well and a good 525P
picture results. But if the previous field came from the previous
cinema frame — an event that occurs regularly on two out of
every five frames — then images that were captured 1/30 second
apart will be artificially combined on the television screen. The
result will be a blur of any moving objects in the video picture.
Such motion artifacts would be completely unacceptable in the
design program of the DVP-S9000ES. That’s why the player
undergoes the full 3-2 reverse conversion process.
Many televisions capable of 525P have internal line doubling or
scaling circuitry capable of converting conventional 525i inputs
into 525P display. Sony’s own such circuits include the Digital
Reality Creation
TM
(DRC
TM
) and DRC Multi-Function (DRC-
MF) systems. With seemingly similar capabilities in both the
DVD player and the television, it’s only natural to ask which is
preferable. In most cases the 525P output of the DVD player
will provide superior results. There are two reasons.
First, only the DVD player can perform the transformation on the
DVD’s digital signal in the digital domain. The alternative is to
convert the signal to analog, transfer this analog signal to the
television, reconvert the signal back to digital and perform 525P
conversion. This exposes the signal to the losses and distortions
of an additional analog-to-digital and digital-to-analog
conversion. So the DVD player has a natural advantage.
Second, when designed properly, DVD player will observe the
correct 3-2 inverse conversion, respecting the integrity of the
original film frames. The overwhelming majority of outboard
devices are not sophisticated enough to detect and maintain this
integrity. Here again, a well-designed DVD player surpasses
most outboard devices.
Sony’s progressive scanning outputs represent a new benchmark
in home video perfor-
mance. But Sony’s design
goals for the DVP-
S9000ES required even
more. Sony built a new
MPEG Image Processor to
undertake three crucial
functions:
1. Motion Adaptive Field Noise Reduction.
2. Block Noise Reduction.
3. Clear Frame Still Image Performance.
Data compressed formats such as DVD are susceptible to noise.
In the video signal, noise appears as tiny flecks or specks of
unwanted color. Typically, the circuits that reduce noise also
suppress fine picture detail. Viewers are asked to sacrifice the
ultimate in resolution for the ultimate in low noise picture clarity.
Many designs attempt to overcome this limitation by comparing
the pictures from several video fields at once. Pixels that
correlate from one field to the next are considered accurate.
Pixels that vary are
considered noisy. This
method performs
beautifully as long as
the images are still.
But because different
fields can capture the
image at different
times, the noise
reduction system can
easily misinterpret
movement as noise.
When this happens, the
noise reduction circuit
can create ghost
images, unwanted
motion artifacts that
may be more annoying than the original noise. In the past, Sony
has overcome this by exempting areas of screen movement from
noise reduction. Of course, this allows some video noise to
reach the television — another compromise.
Achieving 525P: DVD Player vs. Television
Fig. 4: Allowing the television to transform the signal to 525P (top) exposes the signal to
the losses entailed in additional D/A and A/D conversions. Performing the conversion in
the DVP-S9000ES (bottom) simplifies the signal path for an image with lower noise and
lower distortion.
High Performance MPEG Image Processor
Motion Adaptive Field Noise Reproduction
High-quality Progressive Output from Film
Photo 1: Sony’s MPEG Image Processor LSI
performs three important functions to optimize
image quality.
Fig. 5: Conventional digital noise reduction
assumes that any difference between two fields is
video noise. This effectively reduces the
background noise, but can create new problems.
In this example, the movement of the car is
interpreted as noise, resulting in an unwanted ghost
image — a motion artifact behind the car.
