Anatomy of a Video Signal
Print
this Page
You can think of an image as a two-dimensional array
of intensity or color data. A camera, however, outputs a
one-dimensional stream of analog or digital data. The purpose of the
frame grabber is to acquire this data, digitize it if necessary, and
organize it properly for transfer across the PCI bus into system
memory, from where it can be displayed as an image. In order to
understand this process and be capable of troubleshooting display
problems, you need to know the exact structure of the video signal
being acquired. This document gives an overview of common analog and
digital video formats. For information on any non-standard video
signal, please refer to the camera manufacturer's
documentation. Table of Contents:
Standard
analog video signals are designed to be broadcast and displayed on a
television screen. To accomplish this, a scheme specifies how the
incoming video signal gets converted to the individual pixel values
of the display. A left-to-right and top-to-bottom scheme is used as
shown below:
At an update rate of 30
frames/sec, the human eye can perceive a flicker as the screen is
updated. To minimize this phenomenon, interlaced scanning is typically
used. Here, the image frame is split into two fields, one containing odd-numbered
horizontal lines and the other containing the even-numbered lines.
Then the display is updated one field at a time at a rate of 60
fields/sec. This update rate is not detectable by the human eye
(remember AC lighting operates at 60 Hz). Cameras that output
interlaced video signals are usually referred to as area scan
cameras.
For some high-speed
applications, you want the display to update as rapidly as possible
to detect or measure movement accurately. In that case, you might
want to update the display without combining the odd and even fields
into each frame. The resulting image frames would each consist of
one field, resulting in an image with half the height and twice the
update rate as the interlaced version. This is called
non-interlaced video,
and cameras that output signals of this type are referred to as
progressive scan
cameras.
Line scan cameras are a third type; they output one horizontal video
line at a time. The frame grabber collects the lines and builds an
image of a predetermined height in its onboard memory. A variation
on this is a mode called variable height
acquisition (VHA). In this mode, the frame
grabber collects video lines into an image while another input
signal remains active. When the signal becomes inactive, the
resulting image is transferred to system memory. Line scan cameras
are often used to image circular objects; for example, if you were
to rotate a soda can in front of a line scan camera, you could
obtain a flattened image of the entire surface of the can. Line
scanning is also useful for conveyor belt applications, where parts
are moving past a fixed camera. Often a detector is used to provide
a trigger signal to begin the acquisition when the object reaches
the camera. In the VHA mode, a second detector can be used to signal
the end of the object, terminating the acquisition. This is
extremely useful for applications in which the objects to be imaged
are of variable or unknown lengths.
See Also: Conveyor
Belt Applications
An analog
video signal consists of a low-voltage signal containing the
intensity information for each line, in combination with timing
information that ensures the display device remains synchronized
with the signal. The signal for a single horizontal video line
consists of a horizontal sync signal, back porch, active pixel
region, and front porch, as shown below:
The horizontal sync (HSYNC) signals the
beginning of each new video line. It is followed by a back porch,
which is used as a reference level to remove any DC components from
the floating (AC-coupled) video signal. This is accomplished during
the clamping interval
for monochrome signals, and takes place on the back porch. For
composite color signals, the clamping occurs during the horizontal
sync pulse, because most of the back porch is used for the
color burst, which
provides information for decoding the color content of the signal.
There is a good description for all the advanced set-up parameters
for the video signal in the on-line help for the Measurement &
Automation Explorer.
Color information
can be included along with the monochrome video signal (NTSC and PAL
are common standard formats). A composite color signal consists of
the standard monochrome signal (RS-170 or CCIR) with the following
components added:
- Color burst: Located on the back porch,
it is a high-frequency region which provides a phase and amplitude
reference for the subsequent color information.
- Chroma signal: This is the actual color
information. It consists of two quadrature components modulated on
to a carrier at the color burst frequency. The phase and amplitude
of these components determine the color content at each
pixel.
Another aspect of the video
signal is the vertical sync (VSYNC)
pulse. This is actually a series of pulses that occur between fields
to signal the monitor to peform a vertical retrace and prepare to
scan the next field. There are several lines between each field
which contain no active video information. Some contain only HSYNC
pulses, while several others contain a series of equalizing and VSYNC pulses. These
pulses were defined in the early days of broadcast television and
have been part of the standard ever since, although newer hardware
technology has eliminated the need for some of the extra pulses. A
composite RS-170 interlaced signal is shown below, including the
vertical sync pulses. For simplicity, a 6-line frame is
shown:
It is important to realize
that the horizontal size (in pixels) of an image obtained from an
analog camera is determined by the rate at which the frame grabber
samples each horizontal video line. That rate, in turn, is
determined by the vertical line rate and the architecture of the
camera. The structure of the camera's CCD array determines the size
of each pixel. In order to avoid distorting the image, you must
sample in the horizontal direction at a rate that chops the
horizontal active video region into the correct number of pixels. An
example with numbers from the RS-170 standard:
Parameters of interest:
- # of lines/frame: 525 (this includes
485 lines for display; the rest are VSYNC lines for each of the
two fields)
- line frequency: 15.734 kHz
- line duration: 63.556 microsec
- active horizontal duration: 52.66
microsec
- # active pixels/line:
640
Now, some
calculations we can make:
- Pixel clock (PCLK) frequency (the frequency at which each pixel arrives at the
frame grabber):
640 pixels/line / 52.66 e-6 sec/line = 12.15 e6
pixels/sec (12.15 MHz)
- Total line length in pixels of active
video + timing information (referred to as HCOUNT):
63.556 e-6 sec * 12.15
e6 pixels/sec = 772 pixels/line
- Frame rate:
15.734 e3 lines/sec /
525 lines/frame = 30 frames/sec
The
following table describes some characteristics of the standard
analog video formats in common use:
Format |
Country |
Mode |
Signal
Name |
Frame Rate
(frame/sec) |
Vertical Line
Resolution |
Line Rate
(lines/sec) |
Image Size (WxH)
pixels |
NTSC |
US, Japan |
Mono |
RS-170 |
30 |
525 |
15,750 |
640x480 |
Color |
NTSC Color |
29.97 |
525 |
15,734 |
PAL |
Europe (except
France) |
Mono |
CCIR |
25 |
405 |
10,125 |
768x576 |
Color |
PAL Color |
25 |
625 |
15,625 |
SECAM |
France, Eastern
Europe |
Mono |
 |
25 |
819 |
20,475 |
N/A |
Color |
 |
25 |
625 |
15,625 |
Digital
video signals are produced by cameras in which the signal is
digitized at the CCD array, rather than at the frame grabber.
Applications that call for the use of digital video usually include
some or all of the following requirements:
- High spatial resolution (larger
images)
- High intensity resolution (bit
depth)
- High speed
- Flexibility in timing or scanning
characteristics
- Noise immunity
The timing signals for digital video are much simpler than
those for analog video, since the signal is already digitized. They
include a pixel clock, which times the data transfer and can be an
input to or output from the camera; a line enable to signal the
beginning and end of each video data line; and a frame enable to
signal the start and completion of each frame:
These signals, as well as the
data itself, can be single-ended (TTL) or differential (RS-422 or
LVDS).
There is no standard scanning
pattern for digital video signals, so the digital frame grabber
needs to be configurable in order to be compatible with all the
different scanning conventions available. One important factor in
the type of scan is the number of taps a camera has. Some cameras can
output two, four, or more pixels in parallel. For example, a 32-bit
frame grabber (having 32 data I/O lines) is capable of reading four
8-bit pixels simultaneously. So, the frame grabber needs to be
configured to place those four pixels in the proper portion of the
image. The camera documentation specifies the exact order in which
the image data will be delivered to the frame grabber.
Related Links: Anatomy
of a Camera
Print
this Page
|