WEBVTT Kind: captions Language: en 00:00:00.070 --> 00:00:04.730 How do you take an existing technological standard, and make it do three times as much 00:00:04.730 --> 00:00:08.290 “stuff”, all the while without really changing it? 00:00:08.290 --> 00:00:13.629 This was the task given to RCA in developing their compatible color television system. 00:00:13.629 --> 00:00:17.890 A black and white television set need only know how bright part of an image is, but a 00:00:17.890 --> 00:00:22.050 color television needs to know how much red is in an image, how much green, and how much 00:00:22.050 --> 00:00:23.050 blue. 00:00:23.050 --> 00:00:27.060 These three combined images will appear in full color, but any one of them on their own 00:00:27.060 --> 00:00:29.480 wouldn’t look right on a black and white television. 00:00:29.480 --> 00:00:33.620 And RCA didn’t have more bandwidth at their disposal as the standards had already been 00:00:33.620 --> 00:00:35.370 established for black and white TV. 00:00:35.370 --> 00:00:39.219 They couldn’t just send three discrete images even if they wanted to. 00:00:39.219 --> 00:00:44.040 So they needed to take the existing television signal, and somehow add more information to 00:00:44.040 --> 00:00:47.880 it, and still make it compatible with existing black and white televisions. 00:00:47.880 --> 00:00:51.990 RCA was working towards a color television system that could do just that. 00:00:51.990 --> 00:00:55.571 They had gotten the fundamentals figured out on their compatible color system, and were 00:00:55.571 --> 00:01:01.870 competing with CBS’s field-sequential color wheel system for FCC approval in 1950. 00:01:01.870 --> 00:01:05.190 You can learn more about that system in the previous video. 00:01:05.190 --> 00:01:09.700 RCA’s system was based on the 1938 work of Georges Valensi of France. 00:01:09.700 --> 00:01:15.370 The National Television Systems Committee, or NTSC, chose RCA’s system as their standard, 00:01:15.370 --> 00:01:16.462 and they lobbied hard for its approval. 00:01:16.462 --> 00:01:21.910 To recap, the CBS system used a conventional black and white picture tube and camera. 00:01:21.910 --> 00:01:25.030 Both of which were fitted with a spinning wheel in front of them. 00:01:25.030 --> 00:01:29.030 The wheel contained repeating sections tinted red, green, and blue. 00:01:29.030 --> 00:01:33.070 If you show each color sequentially and do it fast enough, persistence of vision will 00:01:33.070 --> 00:01:35.689 kick in prevent you from noticing. 00:01:35.689 --> 00:01:39.720 Synchronize the discs between camera and receiver and you’ll get a full color image. 00:01:39.720 --> 00:01:42.900 But to be convincing, the color wheel has to spin very fast. 00:01:42.900 --> 00:01:49.700 CBS increased the field rate from the standard 60 hz to 144 hz, allowing the complete R-G-B 00:01:49.700 --> 00:01:52.350 image to be seen 48 times per second. 00:01:52.350 --> 00:01:58.049 But in doing so, they threw out any possibility of being compatible with existing televisions. 00:01:58.049 --> 00:02:02.740 RCA, meanwhile, was using a new type of picture tube which contained a repeating pattern of 00:02:02.740 --> 00:02:07.100 red, green, and blue phosphor dots in front of a shadow mask, which worked in conjunction 00:02:07.100 --> 00:02:12.040 with three separate electron guns in the tube to simultaneously show a red, green, and blue 00:02:12.040 --> 00:02:14.330 image in one device. 00:02:14.330 --> 00:02:17.910 These groupings are not in any way discrete pixels, which I feel bears repeating from 00:02:17.910 --> 00:02:19.160 my last video. 00:02:19.160 --> 00:02:23.300 They simply serve as a regular pattern with which the shadow mask can prevent the electron 00:02:23.300 --> 00:02:25.600 guns from hitting the wrong color. 00:02:25.600 --> 00:02:29.200 This new shadow mask picture tube was the key to compatibility. 00:02:29.200 --> 00:02:33.220 If all three electron guns ran at the same intensity relative to each other it would 00:02:33.220 --> 00:02:37.360 function exactly the same as an ordinary black and white television, making an image using 00:02:37.360 --> 00:02:40.050 a repeating pattern of horizontal lines. 00:02:40.050 --> 00:02:44.400 But if you control them independently, the picture tube can now make a full color image 00:02:44.400 --> 00:02:49.220 on its own, while still building the image in fundamentally the same fashion as a standard 00:02:49.220 --> 00:02:50.220 television. 00:02:50.220 --> 00:02:55.090 However, RCA’s color system required a more complicated camera with three separate video 00:02:55.090 --> 00:02:59.590 tubes, each of which was behind a filter to provide it only with red, green, or blue light. 00:02:59.590 --> 00:03:03.650 The three separate image components provided by the tubes were combined and encoded in 00:03:03.650 --> 00:03:07.930 such a way that the transmitted signal appears to be a black and white transmission. 00:03:07.930 --> 00:03:12.470 A black and white TV would display a perfectly acceptable image from these transmissions. 00:03:12.470 --> 00:03:16.740 But a color television set could recover the color information from the camera using some 00:03:16.740 --> 00:03:17.740 clever tricks. 00:03:17.740 --> 00:03:22.130 Now, I’m going to do my best at describing how NTSC color is encoded. 00:03:22.130 --> 00:03:26.670 It’s very technical and hertz to think too much about, so I’m going to tell you some 00:03:26.670 --> 00:03:31.650 basic info and do my best to explain how it allows the TV to form a color image from an 00:03:31.650 --> 00:03:33.690 apparently black and white transmission. 00:03:33.690 --> 00:03:37.870 The beauty of NTSC color was how it dealt with the outputs of the three separate camera 00:03:37.870 --> 00:03:38.870 tubes. 00:03:38.870 --> 00:03:42.900 Instead of deal with them directly in an RGB encoding scheme, the outputs from the tubes 00:03:42.900 --> 00:03:48.550 were used to create two separate intermediary signal components: luminance, a brightness 00:03:48.550 --> 00:03:51.740 component, and chrominance, a separate color component. 00:03:51.740 --> 00:03:55.420 The luminance part of the signal was created by adding together the red, green, and blue 00:03:55.420 --> 00:03:57.300 channels from the camera. 00:03:57.300 --> 00:04:00.760 This would be compatible with existing black and white televisions, as it was simply a 00:04:00.760 --> 00:04:02.880 measure of overall brightness. 00:04:02.880 --> 00:04:07.090 The most genius part of NTSC color was that the chrominance component was almost completely 00:04:07.090 --> 00:04:09.920 hidden using some clever encoding methods. 00:04:09.920 --> 00:04:14.200 No extra bandwidth was required because by using a little math, a color television can 00:04:14.200 --> 00:04:18.970 derive the chrominance component from within the luminance component itself. 00:04:18.970 --> 00:04:20.720 So first, some methodology. 00:04:20.720 --> 00:04:24.820 The luminance signal is created by adding together the the R, G, and B components from 00:04:24.820 --> 00:04:27.530 the camera in this ratio. 00:04:27.530 --> 00:04:29.450 Luminance is referred to as Y. 00:04:29.450 --> 00:04:35.080 Chrominance is conveyed as two signals, I and Q. I is created by taking 60 percent of 00:04:35.080 --> 00:04:39.190 the Red signal, subtracting 28 percent of the Green signal, and further subtracting 00:04:39.190 --> 00:04:41.570 32 percent of the Blue signal. 00:04:41.570 --> 00:04:47.750 Q is created by taking 21% of the red signal, subtracting 52% of the green, and then adding 00:04:47.750 --> 00:04:50.180 31% of the blue. 00:04:50.180 --> 00:04:54.190 Yes this is all confusing and I don’t even want to try and work out how these ratios 00:04:54.190 --> 00:04:59.430 fit into each other but the fact of the matter is so long as the TV set can see I and Q, 00:04:59.430 --> 00:05:04.270 it can bring back the original RGB values seen in the camera by applying a bit of analog 00:05:04.270 --> 00:05:07.490 algebra to the known luminance, that’s Y. 00:05:07.490 --> 00:05:12.240 With I and Q signals, a television can use the same ratios the camera did to produce 00:05:12.240 --> 00:05:17.640 I, Y, and Q to bring back R, G, and B, and it will then be able to adjust the output 00:05:17.640 --> 00:05:21.910 of the three electron beams in order to follow the original RGB ratios the camera picked 00:05:21.910 --> 00:05:22.910 up. 00:05:22.910 --> 00:05:27.120 If the television doesn’t see I and Q, it will simply fire all three electron guns with 00:05:27.120 --> 00:05:29.760 the same intensity to produce only black and white. 00:05:29.760 --> 00:05:32.200 Now for the REALLY complicated bit. 00:05:32.200 --> 00:05:34.380 You might have asked where I and Q are coming from. 00:05:34.380 --> 00:05:36.450 He keeps talking about these apparently nonexistent values! 00:05:36.450 --> 00:05:38.220 What is he going on about?! 00:05:38.220 --> 00:05:42.590 Well, I and Q are hidden within the luminance component. 00:05:42.590 --> 00:05:47.950 Using a process called QAM, or quadrature amplitude modulation, the luminance component 00:05:47.950 --> 00:05:54.200 can serve as a vessel for I and Q. To actually extract them, the luminance carrier is multiplied 00:05:54.200 --> 00:05:58.930 by a 3.579545 MHz reference signal. 00:05:58.930 --> 00:06:03.480 The phase of this signal is used in determining the hue of the color, so it was critical that 00:06:03.480 --> 00:06:08.660 the television receive a perfectly timed reference of this frequency from the transmission itself, 00:06:08.660 --> 00:06:10.980 otherwise the hue would be all wrong. 00:06:10.980 --> 00:06:15.670 Just before each line, in the back porch as it’s called, is the colorburst. 00:06:15.670 --> 00:06:18.290 This is a brief transmission of that carrier. 00:06:18.290 --> 00:06:22.600 The television set’s own crystal oscillator will lock onto and synchronize with the carrier 00:06:22.600 --> 00:06:25.240 to ensure the colors are decoded correctly. 00:06:25.240 --> 00:06:28.970 The colorburst doesn’t actually include any color information, but it’s vital to 00:06:28.970 --> 00:06:32.110 the television’s ability to correctly extract the color. 00:06:32.110 --> 00:06:36.580 If the television sees the color burst, it will begin demodulating the I and Q signals 00:06:36.580 --> 00:06:41.600 with the help of that crystal oscillator, now in a phase-locked loop with the colorburst. 00:06:41.600 --> 00:06:44.900 Actually, this is easier to explain in graph form. 00:06:44.900 --> 00:06:47.520 A black and white television signal looks like this. 00:06:47.520 --> 00:06:51.540 These low points are the horizontal blanking intervals between the lines, and this squiggly 00:06:51.540 --> 00:06:53.390 bit is the part you see. 00:06:53.390 --> 00:06:57.030 The brief section between the blanking interval and the start of the visible portion is called 00:06:57.030 --> 00:06:58.780 the back porch. 00:06:58.780 --> 00:07:02.830 Along the visible portion, the higher the line goes, the brighter the image is drawn 00:07:02.830 --> 00:07:03.940 on the screen. 00:07:03.940 --> 00:07:07.820 Remember, the image is made of lines, which can you learn more about, here. 00:07:07.820 --> 00:07:13.320 Now, a color transmission looks nearly exactly the same, but this little squiggle, the colorburst, 00:07:13.320 --> 00:07:16.750 is tucked in just before the line in the back porch. 00:07:16.750 --> 00:07:21.250 This colorburst is simply the carrier that I and Q are modulated on. 00:07:21.250 --> 00:07:26.300 The carrier is suppressed during transmission except for during the colorburst, which eliminates 00:07:26.300 --> 00:07:30.350 most of the interference between the chrominance and luminance signals, allowing a black and 00:07:30.350 --> 00:07:32.280 white television to display the image. 00:07:32.280 --> 00:07:37.090 A color TV, though, will use this colorburst to synchronize its own crystal oscillator. 00:07:37.090 --> 00:07:41.260 Once it’s locked on, the television set’s oscillator will recreate the carrier. 00:07:41.260 --> 00:07:45.370 Imagine the colorburst continuing on throughout the entire process. 00:07:45.370 --> 00:07:51.160 The TV will then use its own generated carrier to complete the task of demodulating the color. 00:07:51.160 --> 00:07:55.270 As the three electron beams fly along the face of the tube, the circuitry multiplies 00:07:55.270 --> 00:07:59.730 both the sine and the cosine of the carrier wave with the luminance value, which will 00:07:59.730 --> 00:08:02.270 now provide the I and Q values. 00:08:02.270 --> 00:08:06.639 Then the derived I and Q values, along with the original Y value are passed through a 00:08:06.639 --> 00:08:11.420 matrix circuit applying the ratios we learned earlier, which will then spit out the original 00:08:11.420 --> 00:08:14.220 RGB values encoded by the camera. 00:08:14.220 --> 00:08:19.130 With this recovered, we finally have the ability to mix the three electron guns separately 00:08:19.130 --> 00:08:21.080 into a coherent RGB output. 00:08:21.080 --> 00:08:25.310 Now, limited bandwidth meant that the resolution of the chrominance component is not as high 00:08:25.310 --> 00:08:26.930 as the luminance. 00:08:26.930 --> 00:08:31.310 In effect, the picture is still mostly monochrome but with gobs of loosely-applied coloring 00:08:31.310 --> 00:08:33.029 thrown on top. 00:08:33.029 --> 00:08:37.460 Think of it as the ratio of the three electron beams--thus the apparent color--being locked 00:08:37.460 --> 00:08:42.360 in by the color signal over large chunks of the line, but still with the ability to in 00:08:42.360 --> 00:08:46.890 unison change intensity, and thus brightness, along with the luminance signal. 00:08:46.890 --> 00:08:51.020 We’re much more sensitive to changes in brightness than in color, so the image would 00:08:51.020 --> 00:08:52.790 seem completely natural. 00:08:52.790 --> 00:08:57.380 In fact, many digital formats such as MPEG-2 exploit our eye’s poor color detection in 00:08:57.380 --> 00:09:02.570 a similar way, dedicating more data to brightness than color to allow for compression. 00:09:02.570 --> 00:09:07.171 Anyway, I’m really not going to elaborate further on QAM because oh my goodness there’s 00:09:07.171 --> 00:09:11.120 a lot of math involved but if you’re intrigued I cordially invite you to visit some links 00:09:11.120 --> 00:09:12.440 down below. 00:09:12.440 --> 00:09:16.600 But to finish our story we started with the last video, when the FCC had given the go-ahead 00:09:16.600 --> 00:09:22.610 for CBS’s color wheel system in 1950, RCA’s color wasn’t working too well. 00:09:22.610 --> 00:09:27.339 In fact at the beginning of the “competition”, they were still pursuing a three tube optical 00:09:27.339 --> 00:09:31.240 system, but they were investing heavily in their replacement system using a shadow mask 00:09:31.240 --> 00:09:32.240 CRT. 00:09:32.240 --> 00:09:37.180 Regardless, the image quality from the CBS system was far superior to what RCA and others 00:09:37.180 --> 00:09:40.550 could muster at the time with their early shadow mask offerings. 00:09:40.550 --> 00:09:45.290 RCA fought very hard to get approval for their system, but it didn’t come. 00:09:45.290 --> 00:09:48.580 But two things happened to make sure they would eventually succeed. 00:09:48.580 --> 00:09:50.950 First was the reformation of the NTSC. 00:09:50.950 --> 00:09:54.640 The National Television Systems Committee created the original black and white television 00:09:54.640 --> 00:09:59.660 specifications, and having seen RCA’s work on a compatible color system coming down the 00:09:59.660 --> 00:10:03.200 pipeline, they got back together and worked hard to try and get the FCC to approve this 00:10:03.200 --> 00:10:04.730 new system. 00:10:04.730 --> 00:10:09.140 This very much annoyed the FCC because they had already given permission to CBS to proceed. 00:10:09.140 --> 00:10:12.810 However, the Korean War threw CBS’s plans under the bus. 00:10:12.810 --> 00:10:17.290 The National Production Authority declared that color television production must cease 00:10:17.290 --> 00:10:24.080 on November 20th, 1951--exactly one year after permission was granted to begin color broadcasts. 00:10:24.080 --> 00:10:28.860 And in the meantime, the NTSC and RCA remained hard at work improving their new compatible 00:10:28.860 --> 00:10:30.030 color system. 00:10:30.030 --> 00:10:35.850 By 1953, the shadow mask CRT was sufficiently developed and was producing excellent results, 00:10:35.850 --> 00:10:40.029 and funny enough it was one of CBS’s subsidiaries that made the first breakthrough. 00:10:40.029 --> 00:10:44.270 It became obvious that CBS’s sequential color wheel system didn’t make any sense 00:10:44.270 --> 00:10:48.820 whatsoever if compatible color was producing similar results. 00:10:48.820 --> 00:10:55.240 In March of 1953, CBS testified before congress that they were abandoning their color TV efforts, 00:10:55.240 --> 00:10:58.680 afterwhich time the ban on producing color televisions was lifted. 00:10:58.680 --> 00:11:04.779 The NTSC restarted their efforts for FCC approval, and it was granted to them on December 17th, 00:11:04.780 --> 00:11:05.779 1953. 00:11:05.800 --> 00:11:10.700 The first nationwide color transmission of the new compatible system occurred on January 00:11:10.700 --> 00:11:15.240 1, 1954, a broadcast of that year’s Tournament of Roses parade. 00:11:15.240 --> 00:11:19.490 In the US, color TV wouldn’t really take off for about 10 years. 00:11:19.490 --> 00:11:24.399 Early color sets were insanely expensive, costing about as much as a new car, and their 00:11:24.399 --> 00:11:27.870 vacuum tube electronics were very finicky. 00:11:27.870 --> 00:11:30.800 These color TVs required lots of patience. 00:11:30.800 --> 00:11:34.390 Though it’s amazing to me that this amount of complexity could even be achieved with 00:11:34.390 --> 00:11:36.330 1950s technology. 00:11:36.330 --> 00:11:40.000 But the fact was, most TV programming was still in black and white. 00:11:40.000 --> 00:11:45.200 It wasn’t until 1965 that the majority of TV broadcasts were in color. 00:11:45.200 --> 00:11:50.040 Until that point, owning a color TV was a luxury that only rarely saw its true value. 00:11:50.040 --> 00:11:56.920 It wouldn’t be until 1972 that color TVs outsold black and white units in the US. 00:11:56.920 --> 00:11:59.140 And now, some odds and ends. 00:11:59.140 --> 00:12:01.140 Actually, later, some odds and ends. 00:12:01.140 --> 00:12:05.040 I hate to do this, but I’ve found over 10 minutes of stuff that I want to talk about. 00:12:05.040 --> 00:12:09.190 I’m sorry, this was supposed to be two parts, but it’s getting longer still. 00:12:09.190 --> 00:12:14.330 In the next video, I’ll touch on the differences between NTSC and PAL, we’ll go over chroma 00:12:14.330 --> 00:12:20.680 dots and color restoration, Guillermo Gonzalez Camarena’s dubious invention, the 29.97 00:12:20.680 --> 00:12:24.529 frames per second nonsense we deal with here in the States, and more. 00:12:24.529 --> 00:12:28.300 Thanks for watching, and thanks for putting up with another cliffhanger. 00:12:28.300 --> 00:12:31.290 If you like this sort of video and are new to the channel, why not subscribe? 00:12:31.290 --> 00:12:35.430 I’d also like to thank all of my Patreon supports for keeping this channel possible, 00:12:35.430 --> 00:12:38.110 especial these folks who get their name in lights. 00:12:38.110 --> 00:12:41.700 You can support this channel and help keep these videos coming more frequently. 00:12:41.700 --> 00:12:45.600 Patreon Supporters have allowed me to focus on the channel to a much greater extent. 00:12:45.600 --> 00:12:49.490 If you’re interested in helping out, please check out my Patreon Page through the link 00:12:49.490 --> 00:12:51.610 on your screen or down below in the description. 00:12:51.610 --> 00:12:55.520 Thank you for your consideration, and I’ll see you next time!