SECAM - Technical Details

Technical Details

Just as with the other color standards adopted for broadcast usage over the world, SECAM is a standard which permits existing monochrome television receivers predating its introduction to continue to be operated as monochrome televisions. Because of this compatibility requirement, color standards added a second signal to the basic monochrome signal, which carries the color information. The color information is called chrominance or C for short, while the black and white information is called the luminance or Y for short. Monochrome television receivers only display the luminance, while color receivers process both signals.

Additionally, for compatibility, it is required to use no more bandwidth than the monochrome signal alone; the color signal has to be somehow inserted into the monochrome signal, without disturbing it. This insertion is possible because the spectrum of the monochrome TV signal is not continuous (for most typical video content), hence empty space exists which can be utilized. This typical lack of continuity results from the discrete nature of the signal, which is divided into frames and lines. (Strictly speaking, monochrome video does use the full spectrum, if arbitrary and unconstrained movement of subjects and/or cameras is permitted. Therefore, all of these color systems compromise luma quality to some extent in exchange for the addition of color—i.e. all of these color signals look worse at some time or other than they would if the color signal were absent.) Analog color systems differ by the way in which infrequently used space in the frequency band of the signal is used. In all cases, the color signal is inserted at the end of the spectrum of the monochrome signal, where it causes less visual distortion (only affecting fine detail) in the uncommon case that the monochrome signal had significant frequency components overlapping the color signal.

In order to be able to separate the color signal from the monochrome one in the receiver, a fixed frequency sub carrier is used, this sub carrier being modulated by the color signal.

The color space is three dimensional by the nature of the human vision, so after subtracting the luminance, which is carried by the base signal, the color sub carrier still has to carry a two dimensional signal. Typically the red (R) and the blue (B) information are carried because their signal difference with luminance (R-Y and B-Y) is stronger than that of green (G-Y).

SECAM differs from the other color systems by the way the R-Y and B-Y signals are carried.

First, SECAM uses frequency modulation to encode chrominance information on the sub carrier.

Second, instead of transmitting the red and blue information together, it only sends one of them at a time, and uses the information about the other color from the preceding line. It uses an analog delay line, a memory device, for storing one line of color information. This justifies the "Sequential, With Memory" name.

Because SECAM transmits only one color at a time, it is free of the color artifacts present in NTSC and PAL resulting from the combined transmission of both signals.

This means that the vertical color resolution is halved relative to NTSC. The later PAL system also displays half the vertical resolution of NTSC (i.e., the same as SECAM). Although PAL does not eliminate half of vertical color information during encoding, it combines color information from adjacent lines at the decoding stage, in order to compensate for "color sub carrier phase errors" occurring during the transmission of the Amplitude/Phase-Modulated color sub carrier. This is normally done using a delay line like in SECAM (the result is called PAL D or PAL Delay-Line, sometimes interpreted as DeLuxe), but can be accomplished "visually" in cheap TV sets using PAL-S (PAL simple) decoders. Because the FM modulation of SECAM's color sub carrier is insensitive to phase (or amplitude) errors, phase errors do not cause loss of color saturation in SECAM, although they do in PAL. In NTSC, such errors cause color shifts (hence the "Hue" control on all older NTSC TV sets to adjust the color phase with a constant bias).

The color difference signals in SECAM are actually calculated in the YDbDr color space, which is a scaled version of the YUV color space. This encoding is better suited to the transmission of only one signal at a time.

FM modulation of the color information allows SECAM to be completely free of the dot crawl problem commonly encountered with the other analog standards. SECAM transmissions are more robust over longer distances than NTSC or PAL. However, owing to their FM nature, the color signal remains present, although at reduced amplitude, even in monochrome portions of the image, thus being subject to stronger cross color even though color crawl of the PAL type doesn't exist.

Though most of the pattern is removed from PAL and NTSC-encoded signals with a comb filter (designed to segregate the two signals where the luma spectrum may overlap into the spectral space used by the chroma) by modern displays, some can still be left in certain parts of the picture. Such parts are usually sharp edges on the picture, sudden color or brightness changes along the picture or certain repeating patterns, such as a checker board on clothing. Dot crawl patterns can be completely removed by connecting the display to the signal source through a cable or signal format different than composite video (yellow RCA cable) or a coaxial cable, such as S-video, which carries the chroma signal in a separate band all its own, leaving the luma to use its entire band, including the usually empty parts when they are needed. FM SECAM is a continuous spectrum, so unlike PAL and NTSC even a perfect digital Comb Filter could not entirely separate SECAM Colour and Luminance.

The idea of reducing the vertical color resolution comes from Henri de France, who observed that color information is approximately identical for two successive lines. Because the color information was designed to be a cheap, backwards compatible addition to the monochrome signal, the color signal has a lower bandwidth than the luminance signal, and hence lower horizontal resolution. Fortunately, the human visual system is similar in design: it perceives changes in luminance at a higher resolution than changes in chrominance, so this asymmetry has minimal visual impact. It was therefore also logical to reduce the vertical color resolution.

A similar paradox applies to the vertical resolution in television in general: reducing the bandwidth of the video signal will preserve the vertical resolution, even if the image loses sharpness and is smudged in the horizontal direction. Hence, video could be sharper vertically than horizontally. Additionally, transmitting an image with too much vertical detail will cause annoying flicker on television screens, as small details will only appear on a single line (in one of the two interlaced fields), and hence be refreshed at half the frequency. (This is a consequence of interlaced scanning that is obviated by progressive scan.) Computer-generated text and inserts have to be carefully low-pass filtered to prevent this.