0
\$\begingroup\$

How can the following NTSC video signal be interpreted? Is it using phases or frequencies or amplitudes?

enter image description here

Is this above only the color information of a video broadcast signal?

I have no background on this type of communication. Is H-BLANKING indicating where the data chunk is finished? I guess the active video period carries color information? chrominance and luminance looks the same frequency. How would the encoder know which one is what? How about peak white?

\$\endgroup\$
6
\$\begingroup\$

It's an analog signal in time domain- what you would see on an oscilloscope.

The horizontal blanking interval is when the electron beam would retrace back to the left side of the screen (but it is normally blanked so that there is no trace on the CRT screen).

All the information is transmitted in a single analog signal with NTSC- color, luminance and sync signals. Imagine it displaying one of the upper horizontal lines in this image (Society of Motion Picture and Television Engineers SMPTE color bars from Wikipedia commons):

enter image description here

There is also a vertical blanking interval and equalizing pulses to be concerned with. NTSC uses interlaced fields with 525 lines, with odd and even fields alternating at a bit under 60Hz (59.94006Hz, approximately).

If this seems a bit complex, remember that they had to shoehorn the additional chrominance information into fixed channel bandwidths, remain backward compatible with existing monochrome television sets, and it had to be possible to get the signals out with a handful of vacuum tubes. Making the chrominance bandwidth less than the luminance bandwidth was one of the trade-offs, and was based on the response of human eye (I believe JPG compression also uses this trick). Early vacuum tube color televisions were not very stable which led to the backronym "Never Twice the Same Color" for NTSC.

Here is an illustration of the phase relationship of the color burst reference and the displayed hue (from this site).

enter image description here

\$\endgroup\$
  • \$\begingroup\$ In the active video peroiod there are 9 bursts. But in the plot only chrominance and luminance is mentioned by arrows. Are the rest of 7 bursts other image information? \$\endgroup\$ – user16307 Mar 21 at 15:15
  • 2
    \$\begingroup\$ I believe it's intended to illustrate one of the middle lines in the display of SMPTE color bars. I'll edit above to add. The shape of the signal during the active video period is entirely dependent on the image. Phase is chrominance and magnitude is luminance. \$\endgroup\$ – Spehro Pefhany Mar 21 at 15:18
2
\$\begingroup\$

That signal encodes timing, brightness and colour for an entire line of video.

It dates from the days of the CRT monitor which used magnetic deflection to scan the raster, hence the rather long horizontal retrace time (Changing currents in coils takes time), there is also a vertical retrace interval and a (rather longer) vertical blanking interval elsewhere in the sequence.

Ignoring the colour for the moment (Pretend the burst and little sets of high frequency stuff do not exist) and you have a line of black and white video, the hsync is detected and used to lock the horizontal scan circuit, the front porch sets up the black level and the voltage then controls the brightness as it scans across the screen. This is black and white telly.

The challenge then is to add colour in a way that does not break backwards compatibility with the installed base of black and white sets (They must be able to receive a black and white picture from a colour transmission).

This is done by encoding the RGB components as Y'CbCr where Y is the brightness and Cb,Cr as the blue and red components. The Y component is sent as the baseband signal (and is the bit that black and white sets detect properly, so they more or less just work with no modification).

The Cb and Cr components are modulated onto a high frequency sub carrier in quadrature which is then combined with the base black and white signal. To be able to decode this quadrature modulated sub carrier the set needs a reference for the phase of the sub carrier, and this is what the colour burst provides, the sub carrier is what you see as the little rectangles of high frequency signal riding on the basic luma component.

Early NTSC lacked the colour burst giving rise the need for US tellies to have a 'tint control' which adjusted the reference phase and also the "Never Twice Same Colour" monkier for NTSC among broadcast engineers.

As it turns out the choice of colour subcarrier frequency (and its exact relation to the line frequency) matters as by careful choice you can minimise the patterning caused by the croma components interfering with the luma component (This is the major difference between say NTSC and PAL, for all that NTSC allowed a simpler decoder, PAL had a markedly better picture).

I have deliberately ignored issues of gamma and colour space in this description.

\$\endgroup\$
  • \$\begingroup\$ Disagree sith one part. All US NTSC broadcast signals have had color burst included since forever. The reason US receivers need a tint control is because of the narrow-ness of the channel bandwidth. The color signal is amplitude and phase modulated onto a 3.58 MHz carrier. However, the channel bandwidth is only 6 MHz, so a large part of the upper sidebands is not transmitted. This means the color information is pseudo-single-sideband. \$\endgroup\$ – AnalogKid Mar 21 at 16:37
  • \$\begingroup\$ Without both sidebands, absolute phase relationships are lost. Relative phases are retained; when you use the tint control to make people look correct, everything else falls into place (sky is blue, grass is green, etc.). The Brits get around this with a different coding scheme called Phase Alternate Line (PAL). \$\endgroup\$ – AnalogKid Mar 21 at 16:38
  • \$\begingroup\$ Since PAL was invented after NTSC, they improved it with more BW, more resolution and alternate chroma phase channel. Just as EU has improved mobile RF signals learning from US legacy limitations. Backwards compatibility has so much inertia effects on cost. \$\endgroup\$ – Sunnyskyguy EE75 Mar 21 at 17:08
0
\$\begingroup\$

Is this above only the color information of a video broadcast signal?

It includes {front-porch, H-sync tip, back porch & chroma burst}="H-blanking" standard patterns then starts with "this" video signal test pattern and repeats.

It is just one scope test signal for one line of colour bars part of one pattern,

enter image description here

For details on YCbCr colour information read here
For basic details on contrast (gain), brightness (offset) read here while Gamma is a curve.

( see below for more patterns )

How would the encoder know which one is what? How about peak white?

By understanding all the standard specs for NTSC which could fill a book.

Colour is a continuous vector phase angle and amplitude with a calibrated quadrature baseband modulators of the 3.58 MHz VCXOs after RGB conversion to two differential colour signals. Read the above link.


Every characteristic of this waveform is defined by geometric and harmonic ratios of the frame rate for the time.

The amplitude of the negative pulse is used for PLL sync of H sync.
Level of "Back porch" right after sync pulse has both the "black level" voltage which is actively clamped to some voltage usually 0, and also the frequency of the colour carrier for IQ PLL demodulation of the carrier signal amplitude and phase.

It the Vpp signal is 1.0V what is more important is the Peak white to Black ratio of ~0.7V

This can be adjusted in TV's using Gain for Contrast and DC offset to shift brightness up and down for both peak and black level. Then non-linear correct is done with Gamma gain curves.

ON CRT based TV's the HV flyback Oscillator resonates with the Hsync frequency and the phase shifts the image Left>< Right. In LCD monitors, there are many PLLs which sync to Pixel rate + phase, H rate, Vsync rate, and also detect interleaved 2 fields per frame vs progressive 1 field per frame scans used in computers. These ratios of clocks are all harmonically related.

The H blank time was necessary for Flyback transformers to limit the sawtooth trailing edge slew rate while sweeping the coil current back to correspond to the left edge.

Inside the TV was a sawtooth H sweep synced to the H pulse rate to provide a linear horizontal sweep. The flyback sawtooth return "fast edge current pulse" emits X-Rays due to >35kV/ x microsecond slew rate from the yoke electromagnet which was blocked by the (Pb) leaded-glass content to safe levels. This was a safety reason for the blanking pulse.

This has been replaced in LCDs using addressable counters like SDRAM with except with an analog charge Voltage for each RGB lumina signal and ratio of each from 0 to 127 for 8-bit colour.

enter image description here

Colour decoding is more complex but well-documented on the web. Do you need a link or can you find it on your own?

December 1953 the FCC unanimously approved what is now called the NTSC colour television standard (later defined as RS-170a.

However, when I was in a related business in 1982, to design a TV scrambler-decoder like a little "Moog Synth" , for every known method from line dicing to Sync suppression, we called it, "NeverTheSameColour".

If you have a DSO and to trace this signal in both time domain and quadrature phase (Vector XY Mode) Can you imagine who the Lumina signal fades from White to Black vertically as the Phase steps to each colour bar while the saturation (amplitude of 3.58MHz of the Chroma amplitude rises from 0 to max then back to 0 vertically?

enter image description here

This is one of the dozens of test patterns that I use to calibrate my TV and PC monitor ( same) for desktop and video (separate controls). It is also the best for testing which TV or monitor to buy for detecting distortion nuances not visible on typical photos. I use DPT.exe designed to fix dead LCD pixels but I use to calibrate computer displays. When LCD technology was new, it was common to have one bad transistor in memory and thus was included in the user warranty disclaimer. Now better clean rooms and process controls have improved this error rate to no longer be an issue.

This one is for Gamma where the Chroma, lumina signal of adjacent lines should match perfectly for each colour and B/W exactly in the middle of the screen if Gamma is true for that monitor. enter image description here

This is a coarse 10% chroma step test of your saturation levels. i.e. if the last end two steps are the same, your monitor setting is saturating. But it has other uses. enter image description here

This one drives your Flyback transformer crazy with harmonic jagged edges and blurred colours but no sweat for an LCD. enter image description here

If you see harmonic colour gradient patterns that don't look right with subtle vertical edges, it is because your DAC has monotonicity errors but hard to notice on random images. If you see ghosts, then it is from LCD capacitive memory effects from bias drift ageing. There are many other effects. enter image description here

\$\endgroup\$
  • 1
    \$\begingroup\$ Why do you use past tense? No one uses this anymore? \$\endgroup\$ – user16307 Mar 21 at 15:38
  • \$\begingroup\$ And what is the phase relationship between the chrominance and the burst signal(COLOR BURST) if you have an idea? \$\endgroup\$ – user16307 Mar 21 at 15:40
  • \$\begingroup\$ CRT's are still used with Analog TV by some? , but now ASTC digital TV signals and LCD's now dominate the market. \$\endgroup\$ – Sunnyskyguy EE75 Mar 21 at 15:41
  • \$\begingroup\$ Colour burst defines 0 phase for quadrature colour while the amplitude of the burst defines 100% Saturation of colour. \$\endgroup\$ – Sunnyskyguy EE75 Mar 21 at 15:42
  • \$\begingroup\$ You don't see classic analogue video much anymore, but if you look at the timing diagrams for a LCD panel of the usual sort, you will find HSync, VSync, Front and Back porch, Active video.... About the only thing they don't have is a burst, because you are already in RGB space by that point. \$\endgroup\$ – Dan Mills Mar 21 at 16:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.