Analog Color TV Wrap-Up--Some extra info

You can support this channel on Patreon! Link below
In this installment on analog color television, we take a look at some odds-and-ends from the last video. Sit back, relax, and enjoy an eclectic jaunt into the realm of shadow masks, dubious inventions, weird mathematics, and more.

Matt Parker's video on 29.97 FPS:

Links about the Col-R-Tel:
The second link has many great pictures and details about this system. It's also a great site on its own for information about television systems, but it's a tad clunky to use.

Two-color TV:
This link explains the NTSC standard's two color transmission scheme, as well as Gonzalez Camarena's work dates 1963.
This is one link dating Gonzalez Camarena's two color system to 1962.
Link about the Telechrome and Baird's other works:

Lastly some links about color restoration:

You can support this channel on Patreon! Contributions from Patron have allowed me to pursue this channel as a potential career, and I've been able to dedicate more time to it because of viewers like you. If you're interested in helping this channel grow, please visit my Patreon page. Thank you for your consideration!

Thank you to all my current patrons!

[phirenz]

RE: chroma dots @6:27.
Your speculation is correct. The chroma signal is encoded by "wiggling" (modulating) the Y signal within that range.

I find it helps to understand this by looking at the signal from the perspective of the TV doing the decoding.

A black and white TV has two radio demodulators, one that locks onto the sound sub-carrier to demodulate the sound (which is FM encoded) and one to lock onto the video sub-carrier and demodulate the video signal (aka, Luminance or Y), which is actually AM encoded. These two radio demodulators are essentially independent, apart from the fact they are tuned to two signals that are right next to each other on the dial.
Early TVs are actually really simple, electronically. All they do is take the output of the video demodulator (which is a nice voltage between 0v and 1v), detect the horizontal and vertical sync pulses (which are used to lock the frequency/phase of the two flyback transformers controlling the CRTs vertical and horizontal scanning) and then send the remaining signal directly to the CRT's electron gun to control the brightness of the electron beam at that position. 0.33v Represents black, 1v represents white, while 0v represents a sync pulse.

The most logical way of adding color to such a system would be to add a 3rd radio demodulator, one which picked up 3rd sub-carrier and decoded a chroma signal. But this would make the TV signal take up more bandwidth, and the FCC had already allocated 6Mhz channels. Additionally, you would have to replace all the black and white video equipment in the recording studios and transmitters to carry and transmit this extra signal.
So instead, the two demodulators are left untouched and the chroma signal is actually modulated on top of the Y (aka Luminance, Black&White) signal to create a combined Luminance/Chrominance signal (which replaces the Y signal and decodes fine as a Y signal on old Black and White TVs). Color TVs actually have to demodulate the demodulated the combined Y/C (Luminance/Chrominance) signal a second time to extract the chroma I and Q signals.

The chroma signal is encoded in the high frequency of this combined Y/C signal. A black line on the TV would have a Y/C with a constant 0.33v across the entire line. A white line would have a Y/C signal of a constant 1v across the entire line. For a solid colored line (say bright-green) the Y/C signal will fluctuate between 0.93v and 1.07v at a rate of 3.58 mhz. The phase difference between those fluctuations and the 3.58 Mhz color burst signal at the start of the line encode the hue of the color, with 225° representing green (178° for Yellow, 100° for Red, 0° for blue). The height of the fluctuations represent the saturation of the color (±0.07v represent 100% saturation and ±0v would be 0% saturation, or grayscale). The average height of the Y/C signal of course represents the luminance.
For complex lines with multiple colors, the fluctuations in Y/C won't be a constant 3.58 mhz, as it will speed up and slow down rapidly to change into the correct phase for the color at each location on the screen.

To decode this complex signal, first Color TVs have to split the Y/C signal by frequency. Early Color TVs would have used a Notch Filter, with a range of frequencies around 3.58Mhz (say 2.8Mhz to 4.1Mhz) being extracted as the chromance, while the rest (say 0 to 2.8Mhz and then 4.1Mhz to ~5.5Mhz) being interpreted as Luminance.
Frequency in a luminance signal is the rate of change of brightness across the line. A solid color across the line would have a frequency of 0hz. An image with vertical b&w bars across the screen, each about 1/10th of the screen wide, would have a frequency of 0.2 Mhz. With vertical bars 1/100th of the screen wide the frequency would be 1.9 Mhz. If you had vertical bars which were were about 1/188th of the screen wide, a color TV would actually interpret it as color infomation and show a solid color. (A number of 8bit computers like the Apple II actually took advantage of this to create color). But as the width of of the vertical bars got thinner and the frequency increased over 4mhz, they would become visible as black and white bars again.


So what are chroma dots? They are simply the actual chroma infomation which has been modulated right in the middle of the luminance signal. B&W TVs that were manufactured after NTSC was standardized are meant to the same notch filter that color TVs use, and simply discard the infomation with that frequency.

[thomashalo]

the amount of work that went into giving us color tv while keeping B&W compatibility is mind blowing and I never gave it a second thought.

[itisonlyadream]

I enjoy your presentations, the fact that I already know the material gives me a greater appreciation of how well you are able to simplify the subject matter and cram so much into a short video. I also know when you make a mistake, but I didn't detect any in this video. I was eleven years old in 1953 and the introduction of compatible color was a big deal to me, even though my family couldn't afford a color TV set. From 1953 to 1956 I saw many of the first compatible color telecasts by going down to local department stores, which usually had at least one color TV set on display. I would sit, or stand there for hours, just to see TV in color. The salesmen stopped trying to run me off when they discovered that I could answer technical questions for customers and I kept the hue and saturation controls adjusted as soon as faces started to drift to green or purple. Nobody could adjust those early color sets for natural looking flesh tones better than I could. I'm glad you talked about chroma dots, because I always knew when a show was being broadcast in color (even when I was at home), because I could see the chroma dot crawl on our old B&W TV screen as plain as day. I quickly learned to recognize the color of items on a B&W screen by the dot pattern, and I got so good at it that I could almost imagine I was seeing the picture in color on a B&W screen. I could do this, because in the early days of live color broadcasting, people on TV commented on the color of things in the scene, so people at home with B&W sets would know what they were missing. It may have been a marketing ploy, intended to spur the sale of color sets, but people on TV were always commenting on the color of clothing and items on the set during the early days of live color TV, especially on variety shows. Thanks to this feedback, it didn't take long for me to associate the primary colors with specific dot patterns. The color bars TV stations used to broadcast along with the test pattern in the early days also helped me associate colors with chroma dot patterns. The dot crawl was also unique to each color, so it was helpful in guessing color on a B&W TV. It became a game for me to announce the color of an object on a B&W screen out loud before anyone on the TV show said what the color was. My friends and family could never figure out how I did it, and I won a few bets with the trick before people became convinced that I could really do it. I remember one night when a bright red convertible sports car drove on to the set during a Perry Como color broadcast, and I knew instantly that the car was red from the chroma dots, before Frank Gallop (the announcer) mentioned that it was red. Red was the easiest color to spot on a B&W TV screen, because the chroma dot pattern really stood out and it had a crawl that almost seemed to flicker. I could spot red from across the room, but for other colors I had to be close to the screen. During the first color broadcast of the Original Amateur Hour, Ted Mack mentioned that his fountain pen leaked just before the show began and there wasn't time to change his jacket. Then, he added, that this was his first show in color, so the lucky people with a color set could see that the ink stain was blue, while the rest of us only saw a black spot. They did a close-up of the ink spot while he was talking, so it was easy to see. By the time Ted Mack mentioned that the stain was blue, I already knew it was blue from the distinctive chroma dot pattern. Anyway, keep up the good work, and I hope this ramble down chroma dot memory-lane wasn't a bore. I never thought I'd be talking about this for the first time after more than 60 years.

[TechnologyConnections]

Greetings!
I've been looking through past comments (I do read them all, you know) and I've seen a common suggestion for more info graphics and less talking head. I really do appreciate this sort of feedback and I'm doing my best to address it. However, for this video, there's not a lot to show since it is really more of a string of factoids.
I decided not to go into SECAM for this video--I just delved a little into the PAL vs. NTSC fight we seem to still be going on about even though analog television broadcasts aren't happening anymore...
Thanks for watching, everyone!

[boblowes]

Another fascinating video. You're quite right that we started B&W PAL broadcasts in the UK in 1967. In fact, if you ever watch early episodes of a series like Doctor Who, which began as 405-line B&W, and then shifted to 625-line B&W in 1969 (that's roughly when BBC One shifted to 625-line) you can really see the difference in quality.

With regard to making telerecordings of shows made on video tape, one of the main reasons for doing this was that the BBC were fairly active in selling their programs on to foreign markets quite early on. This was usually done on a hierarchical basis - rather than go to the expense of making dozens of multiple film copies, the BBC would simply make a handful. These would then be sent to the first foreign broadcaster who had the right to show them, that broadcaster would send them into the next, and so on. And because of the myriad of broadcasting standards that existed at the time (some markets would be on 405-line, some on 625. Some on PAL, some on Secam (the French system) and others on NTSC - and then add colour into the mix by the early 70s), then the simplest and most broadcast-standard friendly method was to use B&W film copies.

Of course, there are three problems with this method. Firstly, you can't guarantee that the last broadcaster which had the film copies actually sent them onto the next in the chain. Indeed, the BBC have managed to recover programmes long thought lost for good by tracing telerecordings that were either never sent on to the next broadcaster, or in fact returned to their commercial sales department, BBC Worldwide (formerly BBC Commercial Enterprises). Before the BBC established a proper archive, the broadcast arm would often make telerecordings of videotaped programmes for Commercial Enterprises, then wipe and reuse the tapes to make something else. Commercial Enterprises would sell the telerecordings to overseas broadcasters, and after a few years of a programme being available for sale, assumed its commercial viability had come to an end. No further sales would be made, and they would burn the telerecordings, as they were under the impression that the broadcast wing did indeed have a proper archive. Many thousands of hours of vintage television was lost in this way. Often, the only reason there are copies of vintage programmes available, or we only have B&W film copies of programmes originally shot in colour is because of film telerecordings.

The second problem, is that censorship standards would differ from country to country. And what might be acceptable to UK audiences might not be for Australian or New Zealand censors, who would often remove offending material by simply cutting the offending sequence from the master telerecording, rather than go to the trouble of making their own copy. And when this was sent into the next broadcaster, that sequence was missing. Sometimes, these trimmed sequences have been recovered and are the only surviving example of the programme at all.

The third problem with the telerecording method, is that if that programme was original shot on video, the telerecording looses the smooth look of video; as instead of being made up of 50 fields per second, copying that programme onto celluloid blends those fields into 25 frames per second (if the camera was modified to run at that speed, or 24 frames if not). And that blending of fields into frames can lead to visible errors in vision mixing between different cameras from the original multi-camera shoot (almost all videotaped programmes in the UK were shot multi-camera, with a vision mixers switching shots between different cameras). Fortunately, this was rectified when one particular fan of vintage television noticed that a repeat broadcast of a programme that only existed at a telerecording, which he had taped on his domestic VCR, reverted to it's naturally smooth motion on screen as he fast-forwarded through it. Realising that the VCR was in effect "blending" the individual frames back into fields, he contacted some friends who worked in broadcasting, but who also specialised in restoration of old TV programmes. Armed with this information, they were able to devise a process where old telerecordings of programmes shot on video can be made to look like video again.

With regard to how long it took to establish PAL as standard due to the challenging topography of Europe, in fact the BBC was still broadcasting B&W 425 line television to some parts of the UK as recently as the 1980s, due to those places being unable to get a good UHF signal for 625-line PAL - but they could manage a VHF signal for 425 line. Since the introduction of satellite and digital terrestrial broadcasting, this is less of an issue. But (for example) at my house, I used to receive a fairly poor analogue terrestrial TV signal, as I don't have direct line of sight to a transmitter, and must rely on the signal bouncing back down to me from the ionosphere. When the switched to digital terrestrial broadcasting (using the DVB standard) the picture quality improved massively. Then, the strength of that signal was reduced, and I can no longer receive digital terrestrial signals. I can get digital satellite though, so all my television comes via that.

With regard to showing movies on PAL systems and then running at a slightly higher speed - yes, this is true. If fact, when James Cameron's Titanic was broadcast on BBC One some years ago, there were complaints from some members of the public who assumed that the BBC had cut out material, due to the slightly shorter running time. In fact, the film had been shown unedited - but due to the extreme length, the slightly higher speed of 25fps meant a noticeable difference (at least, for petty-minded fans of epic disaster movies). The slightly higher running speed of movies also means a slightly higher audio pitch, which most PAL broadcasters compensate for by lowering the pitch by the same amount.

With regard to the colour-recovery method you described - the software was actually written in a modern Windows-version of BBC Basic - a version of Basic that dates back to the Acorn Computer's BBC computer series of the early 1980s, introduced with the BBC's Computer Literacy Programme - a scheme they introduced to make computers available in schools and used to teach children. Almost all UK schools had at least one BBC Model B Microcomputer at the time, and the version of BASIC included on BBC Computers was a particularly good one.

There is another method of colour recovery that was used by that same team of vintage programme restoration specialists I mentioned easier. They were all fans of Doctor Who. Many of the early colour episodes of Doctor Who had been sold abroad as B&W telerecordings to those countries that didn't have colour in the early 70s; but sales to the United States and Canada tended to be colour, as that was obviously the preferred format for North America broadcasters. And then, in typical BBC fashion, the original PAL master tapes were often wiped. When this occurred, the BBC would (in later years) obtain back their NTSC-conversion masters, and reconvert those for PAL. However, one particular Doctor Who story only existed as relatively high-quality B&W telerecordings, and a domestic recording of a North American transmission on a home format, which wasn't a suitable for a proper VHS release, let alone a repeat broadcast on BBC TV. So, in order to make a good quality colour begin, the restoration team took the telerecordings for their picture quality, and matched that with the colour signal from the NTSC home recording - combining the two to get, in effect, a colour telerecording. This was the mid-nineties, so the differences in screen geometry were easily corrected by bending the overlaid colour signal at the edges of the screen to match the B&W telerecording. As this has gone by, the same team have built a rather sophisticated reverse standards conversion machine, which has vastly improved the quality of programmes that were shot on PAL, converted to NTSC, then converted back to PAL years later. These used to be pretty terrible, but now are almost indistinguishable from other PAL programmes of the time. And, by incorporating the frames to fields conversion, they have managed to make some very high quality restorations of vintage programmes that would otherwise be in a very poor state. Combine that with the standard repairs for hairs in the gate, scratches, speckle, etc etc, they Restoration Team (add they are known) often put as much work into a recovered programme as you'd get with a major Hollywood studio rescuing a movie shot on nitrate stock.

[xaverlustig]

A few assorted remarks ..

1. The terms PAL and NTSC do not describe frame rate or line count. They are colour systems only, and any of the two (or three, with SECAM) can theoretically be used with any frame rate and scan line count. For historical reasons, NTSC is mostly used with 525/25, and PAL is mostly used with 625/25, but that doesn't mean that is mandatory, or that the terms define resolution. There is one country that uses/used PAL with 525/30, namely is Brazil. PAL colour on top of a 525/30 signal was also used as a hack for bridging incompatibilities in multstandard VCRs, and apparently a reverse hack, i.e. NTSC colour on top of 625/25, also existed.

The 625/25 family of signals have been around in Europe since 1948 and were broadcast in black and white for nearly 20 years on. At that stage the term "PAL" did not exist, and colour television was far beyond the horizon. The fact that the introduction of PAL colour did not require a slight shift of the frame rate as in NTSC was just luck because the maths turned out differently, it was not planned for.

The UK is an exception in this regard, because as opposed to continental Europe they opted to retain their pre-WW2 405 line standard after 1945; they could have introduced colour television (with any of SECAM, PAL or NTSC) based on that standard, there was no necessity to switch to 625 for colour. (And that was seriously considered, they did tests with all three colour systems with 405). The UK switched to 625 eventually mainly in order to alleviate incompatibility issues with the rest of Europe. So with the 1962 introduction of 625 lines, the UK was a latecomer. So if you based your research mainly on UK sources, you might be getting a slightly distorted view of the events. In continental Europe, 625/25 in black & white had been the standard for nearly 15 years before then.

2. I think the major factor why Europe introduced colour tv nearly 15 years after the US was economics. Europe was still building up from the war and didn't have the resources to invest into such luxuries, while in the US consumerism was already in full swing.

3. I'm in Germany, in the 1970s and 80s in some areas you could receive terrestrial television in NTSC from AFN, the station serving the US military. You needed a multi standard set which were rare and expensive then, so not many people did. From my memory, the hue problems of NTSC were quite apparent and we had to use the tint dial quite a lot to correct those green and purple faces :).

[DodgeWatt]

Nice use of tint during the explanation!

[Edubarca]

You missed the SECAM French colour TV. Some explanation would be interesting because this system was used in the old USSR, Egypt, France and other French culture countries. When colour TV was going to be implemented in Colombia, we had representatives from the three standards, NTSC, PAL and SECAM. Each one gave very interesting demonstrations and all arguing that their system was the best. Finally our Government decided on NTSC which it had many detractors, it was the best system not only because all the studios and all TV sets were American standards and for a country so near from the US and also broadcasting many shows from the States made it the obvious choice. All this happened in 1978 and the first official colour TV broadcast was on December 1, 1979 at 6:30 pm, President Turbay started with a speech, then an American movie.

[moddymadeye]

The thing with Gonzales Camarena, and so many people mentioning him in the comments, it's that in Mexican elementary schools we are taught that HE invented color TV. Even one of the first Mexican tv channels, Channel 5, made his name part of the name code, XHGC.

[gabef]

I liked old tv tint control as a kid. Fun to decalibrate for funky colors.

[digitalmetadata]

A few points: the NTSC (National Television System Committee) in 1951 demonstrated PAF (Phase Alternate Field) and PAL (Phase Alternate Line). Unfortunately to facilitate the line averaging to cancel the hue error in PAL, a 64us glass delay line had yet to be invented. It was not available until about 1960 which led to the PAL standard.
The BBC was planning to adopt NTSC as late as 1966 but at the last moment switched to PAL.
The BBC launched 625 lines in 1964 which would eventually supersede the 405 line system dating back to 1936. The 8MHz channel adopted in the UK in 1964 as opposed to the North American 6MHz channel adopted in the US in July 1936 is the main reason for US television resolution appearing somewhat less. PAL because of the line averaging, reduces color vertical resolution and the alternating phase encoding didn't allow for as efficient chroma-luma interleaving which meant that fine detail in the PAL picture say was more prone to 'crosscolor' a flickering rainbow effects. This was seen on presenter's shirts and ties and was fun to watch in PAL. This and other problems with PAL led it to be referred to as "Problems Are Lurking".
Alas, if the BBC had pursued 625 line NTSC, it would have provided the best of both worlds: superior resolution and color.

[crffdarkdays]

Cannot believe how on point this guy's production quality and editing is, the way he emulates the NTSC colour shifting.

Even though pal is 10hz slower, I now am very thankful I grew up in a pal region.

[geogeo]

As a German i say, NTSC in the 50‘s was a great performans by american engineers! Heart of NTSC is of the two Color Signals, you need mulitplyers, free wheeling oscillators and so on. In this times no problem, but in the 50‘s there are only very bad and expensive vakuumtubes.

[SkyCharger]

about 25vs30: this has nothing to do with quality, but all to do with mains-hum.
Europe has 50Hz/230V mains and PAL has a field-rate of 50Hz to keep the mains-hum outside the visible portion of the signal.
America has 60Hz/110V mains and NTSC has a field-rate that's close enough to do the same thing.

[therealhardrock]

3:13 Yes, all of you from PAL regions have actually been watching our movies in the wrong pitch. If you've searched for clips from your favorite movies and heard them with a lower pitch, you're actually hearing them in the correct pitch.

[CC-kenp]

And here is why the Delta Shadow Mask is not used for “normal living room” TVs:
While the Delta mask gives you a good resolution and the electron guns are nicely packed tight in the CRT neck, it also swallows way over 80% of the beams. This is why a mid-sized monochrome TV can work with an acceleration voltage of 8kV or less, a color TV the same size needs at least 21kV. You need much more energy inside the beam to have an adequate brightness of the picture.
A Trinitron will not only allow a maximum of resolution, it eats only about 40% of the beam energy (1/3 in theory), you get more than twice the brightness out of your beams. But there are two major problems:
1) the shadow mask is made out of thin wires which have to be really taunt which adds a tremendous force to the screen. The glass of the screen has to be a lot thicker to handle the forces. Just put a 14” trinitron and a 17” delta monitor on a scale, the smaller trinitron is a lot heavier due to all its extra glass.
2) The electron guns need to be sitting in a line next to each other making the neck of the tube a lot wider which in turn makes the deflection coils a lot larger and requiring a lot more power for deflection. Distance is your enemy when working with magnetic fields, the field weakens to the cube with distance! And since the gun beams have different distances to the deflection coils, you need to have an extra homogeneous field which in turn requires even larger coils to create a “Helmholz pair”
A compromise is the slotted mask. Like the trinitron, it eats up less beam power (somewhat more than 50% if I recall correctly) but doesn’t add forces to the screen like the trinitron.
So living room sized TVs are much cheaper and brighter using a slotted mask while small TVs still can use a delta mask for better colors. Computer monitors have to have a delta tube to be able to display an adequate resolution – or need to be trinitron.

[kzhd]

You do a remarkably good job of explaining a difficult subject. I was in the TV industry for 40 years and am still learning things about the NTSC (and competing) color system, including the difficult math. Luckily I only had to make it work, not derive the equations!
BTW, while I agree that the technical stuff would benefit from more illustrations, yours is one of the better "talking heads, " IMHO.

[regnaringversen]

In a NTSC or PAL signal (also called composite video) the color signal based on a subcarrier is modulated with the luminans signal. It is very difficult to separate the composite signal into Y (luminans) and C (color) without aliasing. Simpel filtering using passive electonic components like capacitors and coils will give the aliasing (the crawling dots) on a black-and-white TV. Broadcast signals were internally based on component video, thus no crawling dots in tape editing. Component is Y, R-Y and B-Y signals making the video processing much easier. Yes, it requires three cables and three identical amplifier circuits to transmit a component video signal. However, due to black and white TV the colors have to be transmitted as luminans and chrominans mixed together - the composite signal. The subcarrier would when be used to set the hue and decoding correct on colour TVs. On broadcasting the tape recoders or synchronizers used a comb filter to split the received NTSC/PAL into luminans and chrominans for later editing. The comb filter came in the early 90's and very expensive, and it did not make sense to implement a comb filter on mono chrome TVs. Most people will not know about the crawling dots, mainly on cyan colours, and will be accepted as a trade off for having backward compability with B/W TV when NTSC/PAL made is possible to transmit colour is a single and relative low-rate bandwidth.

[richjames]

A more informed analysis than many, well done. Two points. PAL did not get everything right, e.g. Hanover Bars which they fixed by repeating the previous line color info. PAL was also much more prone to Chroma noise on long distance high power transmissions. This together with the Hanover Bars fix were built into SECAM which was a much better system than PAL. Of course there is always a downside. Post production in Ntsc is easiest and Secam is hardest. As a result most Secam production was made in PAL and transcoded fir transmission.

The one error in your piece is where yu talk about tapes being wiped and b/w films being made. A telecine is a device which scans a frame if film and converts the image and the accompanying sound to a video and audio signal for transmission or recording on VT. Telecines do not record anything. The devices used to make these film copies (mainly used for export sales to developing markets) were called telerecorders and made telerecordings. These were basically film cameras with an optical sound head pointed at a tv monitor. So you should correct that. Telerecordings were very useful in the early days of television where there was no real standards conversion and with careful line up it was possible to use a telerecorder to overcome this. Subsequently analog then digital converters recomposed the pictures with vastly superior results. Oh one last thing, you confused color standard which CCIR broadcasting standard. The line field frequency is the result if the broadcast system not the color standard. So in the USA M stands for 525/60, the black level, the Peak white standard and the placement of synch data in fly back area. So Japan has NTSC J which had 525/60 but differed in other areas including frequencies. PAL is often 625/50 but in South America PAL M using PAL color with US line field rate and Argentina had PaL N which had a further mixture if the two.

So we all have one HD standard if 1080i/ 720p now right? Nope we have ATSC, DVBT DTMB and ISDB (2 flavors) as well as the line field rate the compression codecs the frequencies, mux etc etc. and 4K and 8K drift further apart lacking full standards yet. Look forward to you doing one on UHD. Thanks

[CC-kenp]

So much information in my head I want to tell you but so little time to do it…
The reason why PAL was not adopted everywhere (else) is that it isn’t the superior system and it is a lot more expensive!
The idea of PAL is that the color information is inverted every second line. Any error will e.g. show up positive in the odd lines and negative in the even lines. The average error should be zero – in theory!
By mixing the color information of two lines and taking the average, you get half the color resolution but eliminate the broadcast errors caused by interference while the signal is traveling to your areal.
In reality, the saturation of the colors degrade with the error. So with PAL, a phase shift will still give you correct colors but they are less intense. It just makes the problem less noticeable, not vanish.
Another problem is the costs. How can you mix the color information of a line with the information of a previous line? The previous line is a matter of the past, long gone!
Digital video signal storage was impossible in the early PAL colour TV era. And when it became possible, the circuit board in a studio broadcast machine which does it digitally was sold for over $4000 (mid 80s). Even today those boards are sold for more than $300 (used)!

So a PAL TV contained an ultrasonic delay line. An ultrasound transducer injects the color signal into a crystal which is then picked up by an ultrasound receiver 63, 942 µs later. This method was replaced in the late 1980s using sophisticated chips but still around until 1995. This crystal is a real jewel so while NTSC was nicknamed “Never the same color”, PAL was known as “Pay Additional Luxury”.
The problem is that if the video runs a bit faster or slower, the delayed color video line can’t match with the current one. Also there are manufacturing tolerances as well so it is impossible to get a good match. This also decreases the saturation of the result. But there is a simple remedy, crank up the saturation. So while the picture still looks good, the color resolution is decreased.
You loos 1/2 of the color information in the PAL system and then you loose more information due to tolerances. But this is still OK since the colour resolution of the human eye is only 1/3 of the brightness resolution. So PAL which is actually real bad for the colors is still OK for the human eye.
SECAM tries to fix this by just sending one of the 2 information on the carrier alternating. You still need the delay line which never really matches up but as a result, you get no losses in saturation and you only use 1/2 of the color information. Since the human eye can’t notice that the colors won’t quite match on the brightness pattern, it also works good for the human eye – just with more accurate colors than PAL.
Since the late 1980s, NTSC became superior to all the other formats! They just add a line of test patterns above the visible part of the screen. The micro-computerized Tvknows how this line has to look like and can detect all errors and then work simple filter circuits to compensate those errors. So modern NTSC TVs have much more color information and can eliminate errors just as good (or better) than classic PAL or SECAM.
PAL+ also does the same nowadays so NTSC and PAL can break even here. SECAM on the other hand misses half of the color resolution and there is no way to restore it. So the former superior SECAM is now the worst system while PAL and NTSC sharing the same higher quality.

Fun fact, the video encryption system “Nagravision“ also scrambles the PAL+ line. When PCs became powerful enough (>400Mhz), by identifying where the PAL+ line is, they could rule out 90% of all different ways the picture could be scrambled. By simple try and error they could figure out how the picture is scrambled and decode most European pay-TV channels in real time. It didn’t take long until all EU Pay-TV stations went all digital.