9 Analogue video

9.1 Analogue video signals

A video signal is an electrical voltage that fluctuates over time. These fluctuations are interpreted by a TV set while an electron beam is zigzagging across the screen to draw a picture, line by line. This is also known as scanning, and a single line as a scan line. The shape of the signal fluctuations tell the TV how bright a certain spot on the screen should be  (luma or Y) and what color it should have (chroma or C).
When a pulse of a certain shape in the signal occurs, the TV knows a line has ended and that the beam should travel back to the other edge of the screen to start drawing a new one. This is called horizontal retrace; the pulse is known as horizontal sync. While it is traveling back the beam is switched off so you don’t see the trace. During the part of the video signal at which this is happening is called horizontal blanking or horizontal delay. For each video line coming into a TV, only about 5/6th is actually being drawn and about the first 1/6th of the total time is spent traveling back to the starting position. During the horizontal blanking, another synchronizing signal is included called color burst. It contains information needed to decode the color in the part of the line you do see. A video source like a VCR produces an endless string of video lines.

A specially patterned video line, vertical sync, indicates that a picture is finished. At this point the beam is located at the bottom of the screen and has to travel up again to be able to begin drawing the next picture. With the electronics of the first TV's this would take some time. This time interval is known as the vertical retrace time. During this time, analogue video line signals will keep coming in at regular intervals from the TV tuner or the VCR, but the beam is switched off, so these lines are not drawn on the screen. In the digital age, all kinds of extra information like teletext is digitally encoded in these video lines. This group of lines is known as the vertical blanking interval area (or VBIabbreviated) and the encoded data as VBI data. The lines that result in the picture lines are called active scan lines.

figure 2: illustration of the waveform

Figure 2 shows a full 625 line PAL frame showing the two fields of a BBC broadcast including horizontal and vertical blanking areas. The yellow vertical bar represents the horizontal sync signal that is present in all video lines; the horizontal yellow/black lines the vertical sync lines that separate the fields; the red bars the color burst and the green bar the Teletext area that can be part of the VBI area. Other digital signals that may be present in the VBI area like closed caption and Video Program System (or VPS abbreviated) are not shown. A Wide Screen Signal (or WSS abbreviated; the white dashed line) is included in the first field, but that is just for illustration.

figure 3: a typical waveform of a NTSC composite video signal (source: Composite/CVBS Interface)

Figure 3 shows the waveform of a NTSC scan line. In the active part of the signal, a low signal results in the scan line being black while a high signal results in it being white. A value in between results in a grey line. The signal in the example would have resulted in a horizontal white line being drawn. To have a small black dot in the middle of this line, the signal would have to dip rapidly to black level and up again. The faster the dip, the smaller the dot would be.  This brings us to another point: to show fine detail, the signal needs to be able to change rapid. Rapid changes in a signal equal high frequencies, since frequency is another word for changes per time unit.

This did give some problems when color was added later. There have been various solutions for it, but the basic principle was the same: a single high frequency was singled out and the color information was superimposed (modulated) on it. This frequency was chosen so high that if the signal was watched on a black and white TV, the color information would only distort very fine details which most people would not notice. This single frequency is called the Chrominance subcarrier frequency. NTSC chose 3.58 MHz and PAL 4.43 MHz.
In a color TV, this single frequency is filtered out and processed separately by a process called Y/C separation. To decode the color information, a reference signal is included in the horizontal blanking part of the scan line. This reference signal is called the color burst and can be seen in figure 4. 

figure 4: the color burst (source: Composite/CVBS Interface)

The hue of the color is determined by the phase difference between the color burst and the color subcarrier. The amplitude determines the saturation. For a more elaborate explanation we refer to the links in the reference section.

Depending on the quality of the filter doing this separation, fine luma detail around this frequency would be damaged to some extend, while some luma fine detail would end up in the chroma signal and course color artifacts. High Y frequencies (= fine detail) are mixed with low chroma frequencies (color crosstalk = rainbow effects) and vice versa (luma crosstalk = dot crawls). Look in the Removal of chroma artefacts and the references therein for a more thorough explanation and examples of these artifacts and how to get rid of them.

References:
Glossary of Video Terms and Acronyms: A great place to start looking for oddball video acronyms.
The Engineer's Guide Decoding & Encoding
Television Technical Theory: Analogue Video
Television Technical Theory: Appendix, The Color Encoding Process
Scenarist Closed Caption Tools package: Capturing closed captions with a TV tuner card and graphedit.
Gary Robson's Closed Captioning FAQ: about closed and open captions.
ITU-R BT.472-3 recommendation: Video-frequency characteristics of a television system to be used for the international exchange of programmes between countries that have adopted 625-line colour or monochrome systems. 
ITU-R BT.653-3 recommendation: Teletext systems.
ETSI EN 300 706 V1.2.1: Enhanced Teletext specification
ETSI EN 300 708 V1.2.1: Television systems; Data transmission within Teletext
ETSI EN 300 294 V1.4.1 (2003-04): About the Wide Screen Signal.
ETSI TR 101 233 V1.1.1 (1998-02): Television Systems: Code of practice for allocation of services in the Vertical Blanking Interval (VBI).
ETSI EN 300 231 V1.3.1 (2003-04): Television systems;Specification of the domestic video Programme Delivery Control system (PDC)

9.2 The 625/50 (PAL) and 525/60 (NTSC) timings

Depending on the frequency of the electricity mains in a country, two scan line timings have been developed: 625 lines/frame, 50 fields per second and 525 lines/frame, 60 fields per second. These are often (incorrectly) referred to as PAL and NTSC timing respectively, referring to the TV systems adopted by Germany and America. (Actually PAL and NTSC refer to the system used for color encoding, and nations with PAL 525/60 and NTSC 625/50 systems DO exist. But in this guide we will follow this incorrect habit too.)

PAL numbers:

• It takes 1 second to draw 25 frames.
• One frame contains 625 lines.
• One line takes 1 second / (25 x 625) = 0.000064 seconds or 64 µs (= microseconds) to draw.
• One field contains 288 active lines. The first line of field 1 and the last line of field 2 may contain a whole picture line,  only half a picture line or even be completely black. For capping purposes and aspect ratio calculations, the PAL frame is assumed to be 576 lines high. If it turns out it wasn’t, you just end up with some black lines at the top and the bottom of the picture.
• While a single scan line takes 64 µs to process, only 52 µs is actually spent drawing a picture line. This is the ‘official’ ITU-R BT.470-6 number. (Many analogue sources will not completely adhere to this, mostly resulting in capping some extra black pixels at the sides.)
• The Chrominance subcarrier frequency is 4,433.618.75 MHz
  

NTSC numbers:

• It takes 1 second to draw 29.97 frames.
• One frame contains 525 lines.
• One line takes 1 second / (29.97 x 525) = 0.000063556 seconds or 63.556 µs to draw.
• One field contains 243 active lines. The first line of field 1 and the last line of field 2 may contain a whole picture line, only half a picture line or even be completely black. Most often,  these two lines will contain only half picture information.  For some reasons this has caused a lot of confusion about the active area of the NTSC frame, while this did not happen for PAL. The NTSC frame consists of 486 lines. It contains at most 2x242.5 = 485 lines of picture info. However, often only 484 lines are complete. Virtually all capture devices capture only 480 lines. Presumably that is because by capping 480, you can avoid both the first line of field 1 and the last line of field 2.
• For capping purposes always use 480 lines. For some aspect ratio calculations  use 486 lines, and regard 480 as cropped from that. Forget the about the other numbers.
• While a single scan line takes 63.556 µs to process by a TV, only 52.6555 µs is actually spent drawing a picture line. This is the ‘official’ ITU-R BT.470-6 number, which is often used by people dealing with digital video. There are many standards for NTSC/M, all giving a slightly different number for the active part of a scan line.
• Assume 52.6555 µs if you need a number for calculations. If  the analogue source used a different number, again this will result in capping some extra black pixels at the sides or cropping a bit off.
• The Chrominance subcarrier frequency is 3,579.545 MHz
  

SECAM:

PAL/M:

PAL-60:

Modern European equipment can handle both 625/50 and 525/60 timings. It cannot handle NTSC color encoding. So if you use European 'NTSC playback' capable equipment to play back an American NTSC tape or DVD, they will produce a signal that retains the NTSC timing, but has both the color encoding and the color carrier frequency converted to PAL. Modern European TV’s understand this hybrid signal and show the video. Some game consoles also produce a PAL-60 signal instead of true PAL.
To capture this, you need a device that understands PAL-60. Almost all devices do. Unfortunately, many drivers won’t give you this option. If your device/driver has the option, set the Video decoder to the PAL_60 video standard, and you can follow the NTSC guides for the rest.

NTSC 443:

NTSC 443 is another hybrid. It uses NTSC timing end color encoding, but the color carrier frequency has been shifted from 3.58 to 4.43 MHz, the PAL color carrier frequency. Some game consoles can output this. Once you have set the decoder to the correct video standard, you can follow the NTSC guides.

9.3 Signal types and sources

There are several signal types which are of interest for capturing: RF, Composite, S-video and Component. In this subsection the pros and cons will be discussed, in order decide which type you should use. Some characteristics:

S-video or Y/C


S-video or Y/C carries the luma and chroma components as separate signals. Since no Y/C separation is needed, there will be no comb-artifacts. If it is possible to use such a S-video connection for capturing you should try it, because it will probably result in better quality.

There are exceptions to this rule: devices that store video in composite form, like laser disks but do have an s-video output have to perform Y/C separation internally to produce this signal. If the internal comb filter is of lesser quality than the comb filter in your capture device, you would be better off using the composite video signal instead. So always test!

A special case are VHS VCR tapes. Some people think VHS stores a composite signal while S-VHS stores Y and C separately on tape. This has led to people mistakenly referring to s-video outputs as S-VHS outputs. In fact, both formats store chroma at a lower frequency band than luma (color under). For normal VHS, there is a very slight overlap between the chroma and luma bands on tape, which was eliminated for S-VHS. Still, on playback both formats produce separate Y and C signals. To produce a composite signal, a standard VHS VCR will join these again. Your capture device will have to separate them again, producing the above mentioned artifacts.

A S-VHS recorder however will keep Y and C separate, even when playing back normal VHS tapes. So if you can get your hands on one, use a S-VHS recorder even when you only want to digitize ordinary VHS tapes.

A final argument for preferring s-video for grabbing lies in the way capture devices operate. Most devices have two Analogue-to-Digital-Converters (ADC). When capping analogue, only one of those is used. When capping Y/C however, each signal gets it's own ADC which increases the sample resolution and precision. There simply will be more data for the device to work with.

Component variations: RGB (Europe) and Component Video YPbPr (US)

Component variations: RGB (Europe)	Component Video YPbPr (US)

Component video splits up the video information in three signals. In America, Component Video or YPbPr is prevalent. It is made up of the luma component (Y), the red difference sample (= R-Y = U) and blue difference sample (= B-Y = V). Again, there is no Y/C separation and hence no comb-artifacts.  In Europe, RGB was chosen instead. It carries Red, Green and Blue as separate signals. (Sometimes  it uses 4 signals, R,G,B and sync.) In Europe, the most common form for a RGB connector is the SCART connector, though RCA and BNC connectors are also found on some equipment. Since consumer equipment rarely can capture these signals and most analogue sources cannot output them anyway, the component signal types are not of interest to us.

A few words about the SCART connector: this connector is just that, a connector. Depending on the device it can contain composite, s-video, RGB, sound and various control signals, both in and out and in any combination.  Having a SCART does not imply having RGB or s-video. For instance, most European standard VHS VCR's have a SCART connector that only delivers composite and sound. Using the SCART adapter below will not give you s-video in that case; the result would be that the composite signal ends up on the Y pin of the s-video lead, resulting in a flickering black and white picture.




RF (Broadcast)

RF is the signal used for TV broadcast. A composite video signal and a mono or stereo audio signal are modulated onto a radio carrier wave. To decode this you need a TV card. The TV card will split of the carrier wave and decode the audio signal. Typically, he resulting audio signal is fed into the line in of the sound card using an internal or external loop-through cable. The resulting video signal is composite and suffers from the same limitations.

References:
Television Technical Theory: Analogue Video
Cybertheater - Cybertheather - Video Signal Formats
Cybertheater - Cybertheather - Guide to Comb Filters: Y/C separation

footnotes:
(1) Closed Captioning – Service that provides decoded text information transmitted with the audio and video signal and displays it at the bottom of the display. See (M) NTSC EIA-608 specification. Transmitted on line 21 of NTSC/525 transmissions and line 22 of PAL/625 transmissions, it contains subtitling information only. These captions show sound effects (e.g., "phone ringing" and "footsteps").
(2) Video Program System (VPS) and Program Delivery Control (PDC) - Information is included in the video signal to automatically control VCRs and overrule the timer if a TV program starts on a different time. Hidden in line 16.

Back to VirtualDub postprocessing

Back to AviSynth postprocessing

Back to the Index: HOME

English version last edited on: 07/08/2004 | First release: n/a | Authors: Version4Team | Content by Doom9.org