Video signal system electronics
We will follow the video signal as it takes a trip through the signal system of a typical direct color videotape machine, such as quad or 1" type C. The variations on this theme will be discussed separately.
Video enters the VTR electronics through an input buffer of some sort. Here, manual or automatic video level control is applied to correct for level variations in the video. The video is then low-pass filtered and run through a pre-emphasis network. The pre-emphasis boosts the high frequency components. of the signal, and results in an improved signal-to-noise ratio. After being black and white clipped to prevent overmodulation, the video is applied to the FM modulator. The FM signal is called RF from here on in, because it is well up into the radio frequency range. A sample of the RF is sent to the playback electronics. this signal is called the electronics-to-electronics or E-E signal. The E-E signal is used to monitor input video, and make level adjustments.
The RF signal is sent to the record amplifiers, which boost the signal to a level necessary for recording. The signal applied to the heads during record is usually fairly hefty, due to the high frequencies involved, the narrowness of the head gap, and the high coercivity of modern tape. The signal from the record amplifiers passes through the rotary transformers and on to the video heads.
Playback isn't as simple. The low-level signal from the video heads is coupled through the rotary transformers to the preamps. The preamps boost the very weak signal, and compensates for head resonance, etc. The signal from each preamp is applied to the switcher, which selects which head is playing back at that moment. The switching signal is developed from sensors that determine what rotational position the video head is in.
The switched RF is now applied to the equalizer. This is perhaps the most interesting step in the playback process. Equalization of the RF signal controls high frequency of the baseband video signal. Because the phase linearity of FM signals in this application is critical, ordinary filters cannot be used. They tend to introduce Group Delay, an ineqauity in signal delay through a circuit between the low and high frequencies. Instead, a Transversal filter is used.
A transversal filter shapes frequency response by mixing a signal with one or more delayed versions of the same signal. The phase addition or cancellation alters the frequency response without introducing group delay. This circuit is often called a Cosine Equalizer or Comb Filter. Sometimes, more than one equalizer is used in series. The second equalizer often has adjustable response that changes for each head.
Tricks with the signal
After equalization, the RF signal passes through a switch that selects either playback or E-E RF. After this, the RF now splits into two paths. One path goes into a delay line with a delay of one horizontal line, to a switch. The other path goes directly to the switch, and is the normal path. A level detector senses any drops in the RF level that would indicate a Dropout. A dropout is a momentary loss of RF caused by a defect in the tape coating. If a dropout is detected, the switch is thrown, and RF from the previous line replaces the dropout. In many professional machines, the dropouts are only detected here. The timebase corrector does the actual dropout fix.
After dropout compensation, the RF enters the limiter. After passing through several stages of limiting to eliminate all traces of AM, the RF is applied to the demodulator. The demodulator is usually a pulse-counting FM demodulator. After demodulation, the video passes through a low-pass filter to eliminate any residual RF. The video is then de-emphasized in a network that is opposite of the pre-emphasis network.
After an optional AGC stage, the video is ready for the outside world, which usually involves a processing amplifier and timebase corrector of some sort. More on that later.
Direct color systems, like the one described, are simple and produce high quality recordings. However, mechanical instabilities make the output signal too 'jittery' for color. The mechanical instabilities introduced by a videotape machine are called Timebase Errors. The usual way this error is corrected is to use a Timebase corrector or TBC. These very complex devices are discussed later.
Timebase correctors are too expensive and complex to put into consumer machines, although this is slowly changing. Another means to correct the chroma jitter was needed that would not require thousands of dollars in electronics. The solution was a system called Color-Under. In the color-under system, the chroma (color) information is separated from the luminance (Black and white), and down-cnd white), and down-converted to a lower carrier frequency. (Chroma information is suppressed-carrier analog quadrature AM centered at 3.579545 MHz in NTSC.) This signal is recorded as an AM signal below the FM luminance information carrier on the video tracks. On playback, the luminance timebase error is measured, and the chroma is converted to the correct frequency with the jitter corrected. Some loss in chroma frequency response is incurred in this process.
chroma information is separated from the luminance
After the video input buffer, the chroma information is separated from the luminance information with a filter of some sort. In newer machines, this is often a Comb Filter. (A comb filter is a type of transversal filter that separates luminance and chrominance by using delayed video from the previous line. It takes advantage of the phase reversal that the chroma carrier experiences every horizontal line. More on this later.) A Voltage Controlled Oscillator (VCO) is locked to horizontal sync, and runs at some multiple of horizontal sync, such as 40 X Fh. A second crystal oscillator produces a signal at the color subcarrier frequency.
These are mixed together to form a color conversion signal. Before mixing this with the chroma signal, the phase of this signal is inverted every other line, every other field. The color conversion signal and the chroma are mixed, resulting in a low frequency chroma signal. For consumer beta and U-matic, this frequency is 688.373 kHz. (43.75 X Fh) For VHS, this frequency is 629.040 kHz.(40 X Fh) This signal is recorded along with the FM luminance on the tape.
On playback, the low frequency chroma is separated from the FM luminance just after the RF head switch. This chroma is run into an AGC stage, which removes level variations due to head wear, etc. The 40 X Fh VCO is locked to playback horizontal sync. This compensates the the color conversion signal for horizontal rate timebase error. A second VCO creates the 3.579545 MHz color subcarrier frequency. These are mixed to form the color conversion signal. Before mixing with low frequency playback chroma, the color conversion signal is inverted in phase every other line, every other field. The low frequency chroma is mixed with the color conversion signal, and the original chroma signal results.
The burst in the recovered chroma signal is compared to a crystal oscillator running at the proper burst frequency. Any residual frequency error causes the second VCO in the color conversion signal generator to shift frequency to correct the error. Thus, all color instability is removed.
The consumer VCR formats
The consumer VCR formats use azimuth recording to ensure the FM luminance does not crosstalk. However, this does not work on the low frequency chroma signal because it's frequency is too low to be affected much by azimuth errors. So, there is significant crosstalk present on the chroma signal upon playback. But, remember how the phase of the color conversion signal was reversed every other line, every other field on record and playback? This causes the chroma on every other field to not reverse phase every horizontal line like it is supposed to. This is restored to normal during the up-conversion, but the crosstalk component has it's phase reversal canceled out. The crosstalking chroma is now easily removed in a comb filter. Professional formats employing color-under recording (U-matic, for instance) do not need this comb filtering system, as the guard bands on the tape prevent chroma crosstalk.
The chroma and luminance video are reunited, and the result is a color picture that is stable without a timebase corrector. The price paid? Reduced resolution in the picture due to the separation and recombination process.
It should be noted that the actual signal processing is a bit more complex than indicated here. Also, older color-under formats, such as 3/4" U-matic, which have significant guard bands between their tracks, have simpler color under circuits. Other color-under schemes have been devised and used. Many of these employ a pilot tone recorded below the FM luminance signal that is used to measure and correct the timebase jitter error.
One other circuit that is present in some newer consumer VCRs is the second noise canceler. This is yet another transversal filter that removes luminance noise by summing the current and previous line. Any high frequency information that is not common to both lines is filtered out. This removes some very objectionable noise while having little effect on detail in the picture.