Rca left everyone else at the alter
 

I know that during the developement of color television all the ideas and principles that Rca used for its compatible color system were'nt theirs alone. I wonder if any of you old timers can shed more light on the fact that after the fcc adopted the never the same color system, most manufacturers then decided to let Rca sell color to the public, not to mention the bad taste CBS had for having their line secuential system thrown in the trash. any thoughts?

Nobody else had the deep pockets (for color tv) Sarnoff had.

Those same deep (greedy) pockets led to Armstrong's suicide over the crap they gave him for his invention of FM. Truly despicable ego driven corporate behavior.

Anthony
PS The real tragedy and irony of what Sarnoff did to this man is that if he had only embraced the technology then RCA could've made a fortune for themselves by selling new compatible radios and transmitters, just like what we now endure whenever a new recording format appears (Cylinder records-Discs-LP-Wire recording-RTR-4 track-8 track-Cassette-CD-Mini Disc-MP3........1/2" video tape-Beta-VHS-8mm-Digital.....Laser Disc-DVD-who knows what next!)

---

Where I can find the tragic story of Armstrong?

There is a book called "Man of High Fidelity" that has Edwin Armstrong's story.

Telecolor,

A GOOGLE search for "Edwin Armstrong" will provide several sites including a biography.



-Steve D.

You're not suggesting we would have been better off with CBS's system are you?

actually the colour wars were dirty on both sides of course ..CBS wanted to delay television by promising "Color Now" while RCA wanted to get some money back on its investment in mono TV ....but the behaviour of Senate which was effectively bought off by Sarnoff to prevent CBS broadcasting in color in the late 40s set a new low.

---

There would have been a severe motion artifact with the CBS system. Moving objects would have color fringes that changed from field to field. Only today's digital processing would be able to fix this. Something similar to MPEG.

But on the other hand there would have been no phase error ..but then again extracting RBG would been fun ..and no saturation control....mmm...as for color fringing ...perhaps a quick check with the Early Television museum may clear that up ..as they have some CBS field sequential sets working????

Of course just as CBS did the blue banana trick RCA had swirling batons which very neatly showed this problem! ...aaaagh fun and joy

Yes but would there have been any electronic alternative besides the tri-color tube? I mean they are still selling them now, if something cheaper AND better than the color kinescope had come along wouldn't we have seen it by now after 51 years? So IMO even if CBS had gotten rid of the color drums or wheels, it would have just been another version of RCA's with it's own drawbacks, plus the non-compatability issue. I think the farther you get from 1954, the better RCA's system looked compared to CBS....Frenchy

Very early color film used a two color system...sometimes pretty effectively. Were there ever experiments with a similar arrangement for television?

Anthony

CBS color at ETF 2004
 

Perhaps the thrill of finally seeing CBS sequential color live clouded my objectivity. But I do not recall observing color fringing on the small-screen, operational 1951 CBS-color equipment at the 2004 ETF convention. Small screen is probably the operational word; field sequential color on today’s hi-acreage bottles would change the picture.

Photographing the historic event proved to be more difficult with a moving image, however. It may have had more to do with shutter-speed interacting with frame-rate than something else such as color fringing. But, when we paused the DVD player supplying the NTSC video to the transcoder, our results were more successful.

If you haven’t already seen it, here's a link to some of the best photos (save time, skip over the prattle before the first picture). Thanks again to Reeferman who served as cameraman.

http://home.att.net/~pldexnis/potpou...color2004.html

BTW, the x-hatch pattern conundrum noted in the text has been solved: nothing technical, just a distorted reflection of the museum's ceiling!

Though I have been assured that flicker was a real problem in 1950 when CBS demonstrated their system, I have yet to see it on the 10 inch Gray Research monitor in the musuem. Could the problem have been in the camera back then, and solved now by the use of Darryl Hock's standards converter?

http://www.earlytelevision.org/gray_monitor.html

Another serious problem with the CBS system is resolution. The picture is only 405 lines, compared to 525, but more importantly, with 4 mHz of video bandwidth it is only possible to get about 250 lines of horizontal resolution. On a 21 inch screen that would look pretty bad.

---

Another thing were the giant egos involved. Sarnoff apparently was a world-class horse's arse-one of his quotes was-"I don't get ulcers, I GIVE them." Paley at CBS had a pretty high opinion of himself, too & Major Armstrong apparently was unable to "let it go", & that ultimately destroyed his life. Helluva shame-The Major & Sarnoff initially had been friends-one wonders what could have been accomplished if they'd been able to put aside their differences & worked together.-Sandy G.

I think I remember reading a story somewhere on the web where in 1951, one guy when he was a kid used to tweak the oscillators the family TV set to receive the CBS 405 line color system tests in New York every night, in black and white of course. It was kind of interesting. I guess if the FCC took the CBS color system instead, they could always tweak the B&W sets out there to receive it.

I think I've read somewhere that Philo T. Farnsworth suggested the use of a 625 line picture standard in 1941 when the FCC was settling on the TV standard we use today.

A side note, IIRC, when the World Trade Center was hit on 9-11, WNBC-TV lost their antenna and theyh ad to erect a quick backup antenna on, you guessed it, on Edwin Howard Armstrongs old FM tower at Alpine, New Jersey. Bet Sarnoff was spinning in his grave on that one. :)

This is only my conjecture and impossible to prove, but I suspect that Sarnof, when he asked Armstrong to develop noise free radio reception, wanted noise free *AM* radios that could receive existing RCA NBC owned AM stations. Armstrong delivered FM, which was low noise but (and this was likely a BIG BUT) required new broadcast stations and transmitters and a new band. Which would take time and be expensive. Would also require trying to sell FM radios in markets with only a few stations. Also (and this is probably the really big issue) FM would distract and suck off consumers away from Sarnoff's next big thing: Television. Maybe if Armstrong had suggested using FM sound for the soundtrack for television images to Sarnoff, Sarnoff might have seen this combo as a "killer application" product. A one-two punch. "Great sound and you also get to watch the action too".... Then the Armstrong Sarnoff pjssing contest might not have happened....

As for the shadowmask color CRT, Heard that Philco was trying to develop a color CRT based on other methods. "Apple" code name, saw it somewheres on the web. There's the stripes of colored phosper on the screen and a single electron beam idea, the beam switched RGBRGB in time step as it sweeps
across the phosper stripes. Issues on keeping track of which colored stripe the beam is now hitting area major problem for this scheme. Especially when the image has large black areas, you lose track. Would have gotten rid of convergence problems though. But so did the color wheel, but that had motion artifacts and wasn't compatable with B&W NTSC.

[QUOTE=Pete Deksnis]Perhaps the thrill of finally seeing CBS sequential color live clouded my objectivity. But I do not recall observing color fringing on the small-screen, operational 1951 CBS-color equipment at the 2004 ETF convention. Small screen is probably the operational word; field sequential color on today’s hi-acreage bottles would change the picture. >>

Yeah those Wisard of Oz screen photos are spooky looking, because it's the same 'pure' scanning line of a black and white screen, except with color...no dots, no misconvergence whatsoever.

Frenchy

My impressions of the field-sequential demo at ETF:

Viewing close-up on fast motion I could see some color fringing - but as expected, the fringing was mainly on the part of the image you are not fixed on. If you are tracking a moving object, it is fringe-free, but the background has fringes; if your eyes are stationary, the moving object has fringes.

Viewing from the back of the room (pretty useless for the small screen), you could see color "flashing" as your eyes moved (really, gross fringing of the whole image).

One thing I noticed is that at the brightness levels used, the flicker was not really bad. I always wondered what that would look like. I could see some. Flicker is a very strong function of brightness and scan rate, and a fucntion of individual threshold to a lesser extent. I think personal differences would easily account for Steve not seeing it.

Flicker was much worse on the Colortel converter which was using a color wheel on NTSC scan rates, and I would be surprised if anyone could not see it.

My former boss at Zenith saw the original CBS demos, and always said that the CBS system color reproduction was the best of the systems at that time. But of course, the recent demos all start with an NTSC source, so the original results cannot be judged until we have a color wheel camera operating.

If the demos last year were converted from NTSC composite they would have the reduced color bandwidth, but that may not have been noticeable on the small screen either. I am wondering now if the converter took in NTSC composite, or YPbPr for better color bandwidth.

could have been a sychronization problem
 

back then they didnt have servo drive circuits with micro's controling syncronous motors, my guess is that from time to time the scan rate would have a timing error between the camera and the color receiver. i worked on many of the newer pll motor drive circuits, and i find it amazing that they got as good a picture as they did back then

"If the demos last year were converted from NTSC composite they would have the reduced color bandwidth, but that may not have been noticeable on the small screen either. I am wondering now if the converter took in NTSC composite, or YPbPr for better color bandwidth."

The converter had inputs for NTSC composite and S-video. All the demos in 2004 were with a composite input; an S-video cable was not available.

As I recall, Darrell said there were too many issues involved for component video.

An historical event..
 

I also witnessed the demo. at ETF last year. It was spectacular and the highlight of the convention. No matter the video source it was a treat to view, for the first time in over 50 years, the field sequential system in action. The anticipation was palpable. You could hear a pin drop as the color wheels were powered up to speed and finally produced their full color pictures. The applause broke the silence. I had a sense of what the CBS engineers felt during their trials.


-Steve D.

Hello everyone. I guess I need to check this board more often :yes:

Quite a few questions have been raised about the CBS format and my converter. First the question of color fringing:

Color fringing is visible on the operating CBS sets using the converter, just to a lesser extent than would have been apparent originally. In the CBS system, a complete color image is made up of 6 fields repeated as RBGRBG. (Since it still used interlaced scanning, it required two sets of RBG sequences to make up the entire color frame since each color had to appear in both fields.) In the original system, if an object moved in any direction horizontally or vertically, it's component colors would separate. The amount of separation is proportional to the speed the object moves relative to the scanning speed in that direction across the 6 fields. In this system, each of the 6 fields could be temporally different from the previous one, and since each field represented a different color, color fringing would occur.

Since we are converting from NTSC which contains 2 fields for each color frame, only 2 groups of 3 fields in the converter output can be temporally different per frame. While this tends to decrease the number of fields that appear different, the time difference between the 2 fields in the NTSC input is greater than the difference between two fields in the CBS system making this worse.

To sum it up, because of the nature of the NTSC input, and the frame rate conversion, there is less color fringing in the output of the converter than what would have been observed originally, the field sequential source being the main difference.
As a side note, the converter performs a trick that senses if the original NTSC video came from a progressive source like a telecined movie, in which case it reconstructs the original 24fps progressive movie, which exactly matches the CBS frame rate, so no color fringing appears at all.


Of course none of this addresses the issue of darting your eyes around or moving an object quickly across the screen and seeing the color image break apart. This is inherent in any field sequential system, and is the same today using the converter as it would have been originally. I think this is the effect that most folks noticed when viewing these sets last year.


As for this being a viable option for color, keep in mind that while the frame rate was low on the CBS system to keep the bandwidth down, the field rate was at 144Hz. Compare this to most state-of-the-art single chip DLP projectors that use 4X (120Hz) or some that use 5X (150Hz) field rates. While the DLP projectors have the advantage of coming from non-field sequential sources, they will appear similar to what the CBS system would have looked like.


Someone mentioned the flicker difference between the CBS and ColorTel demonstrations last year. Remember that the CBS system while using a low 24Hz frame rate gave a complete color image at this rate. (actually your eye perceives a complete color image at 48Hz with the CBS system) while the ColorTel converter uses 6 NTSC fields to make a color frame at a 5 Hz rate. (similarly your eye will perceive a 10Hz rate with the ColorTel) This is why the CBS is basically flicker free while the ColorTel has objectionable flicker.


On the converter it self, only a composite and S-Video input were provided. As Steve McVoy mentioned, one of the downsides to the CBS system was it's relatively low bandwidth. The composite input is nice as it connects to just about anything, and the S-Video input is a component format, with the same Y bandwidth as a full YUV input, but with just somewhat less color bandwidth. Because the cost and size of the unit would have gone up with the addition of a YUV input, and the additional color bandwidth would not have been realized, it was decided to leave this off.

Hope this wasn't too long and boring of a post :)

Darryl

One thing I forgot to mention, in the section about how the converter handles telecined material, this would have been true back in 1950 as well. A film telecine chain would have been able to provide a pull down sequence in the CBS format that caused no color fringing. Color fringing would have only occured on material sourced from a camera.

Darryl

Most of that was over my head (just like those flying cows) but I found it most interesting and appreciate your having taken the time to write it!

"As a side note, the converter performs a trick that senses if the original NTSC video came from a progressive source like a telecined movie, in which case it reconstructs the original 24fps progressive movie, which exactly matches the CBS frame rate, so no color fringing appears at all."

Not quite right - The fringing disappearas with a live 144 field camera. There is still color fringing if the object motion (or stillness) and your eye motion (or stillness) don't match. So, even in the 24 frame film case, you see fringing.

What happens on a moving object with 24 frame source: Your eye tracks the average motion, so on the R.B,G,R,B,G...etc fields, your eye is at equally spaced positions 1,2,3,4,5,6, for a first film frame and 7,8,9,10,11,12 for the following film frame. but the object is reproduced 6 times at position #1 for the first film frame, and then 6 times at position #7 for the following film frame, and so on - never at the other positions where your eye is centered at those times. Since the images are stationary for 6 fields while your eye is moving, the R,G,B fields fall on different locations of the retina, and voila! you have fringing. The high field rate helps reduce it, but cannot eliminate it.

Now, if you have a live camera, the object will be reproduced at all those positions at the right times, so if your eye is tracking the average speed, you see no fringing on the moving object. (However, the stationary background objects will show fringing -- oh well.)

The conversion from a different frame/field rate can create different (and maybe worse) fringing since the source images show the object in different positions compared to the 144 field or 24 frame cases.

Micromirror TVs break up the pulse of light for each frame into subframes in order to make a gray scale from a basically on/off procedure. Current sets with color wheels use carefully crafted algorithms so that the color fringing is reduced by the particular sequence of light pulses giving the gray scale, as well as by the high frame rate. Color fringing, which could be quite visible on early units, is mostly invisible these days, although you can still see some color breakup if you stand far away and move your eyes rapidly.

Hi old_tv_nut,
I think we're on the same page here. There are two effects that cause color seperation, movement in the source image using sequential scanning, and movement of the observers eyes, which is why I discussed these effects seperately. The source material issue can be controlled in a number of ways, one of which being using progressive material. The issue of movement of the observers eyes will always be a problem on field sequential displays like the CBS, DLP projectors, and many color industrial/medical displays currently in use. This is true when considering one effect at a time.

When considering both issues together, like the case of an object in motion with source seperation, and the observers eye following it, the background will have the expected seperation, but the object being followed will also show a color tail due to the persistance of a CRT. (assuming a CRT display type) This effect was observed during early testing using a computer generated, true 6 field sequential white square bouncing around the screen.

An interesting thing that came out of the demonstration last year was that the CBS color field rate seemed to be at the threshold of perception for some people. While many observers said they percieved no flicker or seperation, others said they could easily see the seperation. It would have been interesting to see how this effect would have been received 50 years later had the CBS system won out.

Darryl

Regarding flicker, I sometimes have trouble when viewing one of those automotive diagnostic scopes....you know the kind, the large machines at repair shops that display the spark plug firing lines.

I tend to see either a flicker or sometimes I don't see every line. Guess I don't have as much persistence of vision as others....although I more than make up for that with my persistence of annoyance!

Anthony

:thmbsp:

Yep, tubesrule - you have to consider the source and the display and what the eye is doing! :yes:

I went through some really dumb arguments with a person involved with the HDTV standards committees who insisted that progressive scan at 72 Hz would _eliminate_ judder on 24 fps film material. Well, it does not, although it is cleaner than the 12 Hz components you get with 3:2 pulldown and 60 Hz displays. I eventually wrote a little program to calculate the relative displacement over time oif the juddered images for any input and display rates.

I always hesitate pointing out the 3:2 judder to people who have not really paid attention to it, because once you notice it, it's hard to ignore.

By the way, film in theaters is shown with a double shutter so that the flicker rate goes up to 48 Hz. Some projectors have used triple shuttering to raise the rate to 72 Hz so the pictures could be brighter without visible flicker, but this usually results in poorer focus, as the film heats and expands between the three presentations of a single frame. Whether with double or triple shuttering, professional film makers are very careful to avoid bad pan rates or rates of motion across the screen, because the judder can look very bad. Micromirror pirojectors for digital cinema can eliminate much of this problem even when the source is only 24 fps, but good cinematographers still are careful.

For some of the tests of HDTV scan formats for the FCC ACATS, we used a ShowScan film, which is shot at 60 frames/second and therefore shows no judder when televised at 60 fps progressive. The ShowScan film is 70 mm, and at their facility they had a floor to ceiling screen that you viewed at about one picture height or less - much more detail than the 1 or 2 megapixel HDTV standards, and therefore an excellent, nearly transparent source material for HDTV tests.

I believe Baird experimented with a 2-color electronic system in the 40's - have to go look it up. And there is a mention of a color system invented in Mexico, in their 1964 New York World's Fair brochure, which may have been 2-color.

Sometimes some of the very early two-color cartoons will show up on TV, but I have never seen one of the two-color live-action films televised.

Heh old_tv_nut,

"...progressive scan at 72 Hz would _eliminate_ judder on 24 fps film material. Well, it does not, although it is cleaner than the 12 Hz components you get with 3:2 pulldown and 60 Hz displays"

This does seem counter intuative doesn't it :) Changing rates is just such a hard thing to do, but native CRT 24fps HD displays have so much flicker :sigh: That program you wrote to determine the judder sounds interesting.


"I always hesitate pointing out the 3:2 judder to people who have not really paid attention to it, because once you notice it, it's hard to ignore."

It also drives me nuts the way most broadcast TV is time compressed now with it's associated artifacts, but heh, if you can sell more commercial time per half hour, who cares :mad:


"By the way, film in theaters is shown with a double shutter so that the flicker rate goes up to 48 Hz."

That was such a cleaver method using a light cutter for reducing flicker in film, kind of like interlacing on television to reduce flicker, although interlacing has so many side effects.


"For some of the tests of HDTV scan formats for the FCC ACATS, we used a ShowScan film, which is shot at 60 frames/second and therefore shows no judder when televised at 60 fps progressive."

It's a shame that more film isn't overcranked, but few are willing to pay the price. Hopefully with the coming of HD more will be done in 720/60p. The sad part is the technology has progressed tremendously, but little can be said for the content :worried:


Hi Anthony,
Many of these industrial displays use light shutters or valves and do have a lot of visible flicker to them. One that comes to mind is a color Oscilloscope we have at work that is field sequential, and can be quite annoying as you look around at the front panel and see the three colors separate!

Darryl

Having no phase error was true IF the camera tube discharged completely on each field. So, image orthicons worked OK, but vidicon cameras using a color wheel would have contamination from one color into the next, requiring a complex apparatus to store frames if you wanted to fix it. (Or I suppose you could cheat and change the colors of the filters slightly.)

Also, the bit about no convergence errors is only true if the power supplies in the receiver are good. They must have good high voltage regulation so the raster doesn't change size between fields, even if the scene is mostly one color. My old boss who saw the original CBS tests talked about this problem in receiver design. Also, the power supply had to be low ripple, because the 60 Hz line was not in sync with the sweep. The monitors we saw at ETF last year seemed OK in this regard.

[QUOTE=old_tv_nut]I believe Baird experimented with a 2-color electronic system in the 40's - have to go look it up. And there is a mention of a color system invented in Mexico, in their 1964 New York World's Fair brochure, which may have been 2-color.

Not sure of the model or make, but I remeber an early 2 color projection television at a store I worked at in the late 70's. It was an all-in-one, front projection unit in an upright cabinet, and the two crts pointed forward onto a small, flip down mirror that projected the image back onto the front of the screen at the top. Not sure how well I'm describing it, but I think it might have been a Sony. Maybe someone else remebers it.

Darryl

[QUOTE=tubesrule
Not sure of the model or make, but I remeber an early 2 color projection television at a store I worked at in the late 70's. It was an all-in-one, front projection unit in an upright cabinet, and the two crts pointed forward onto a small, flip down mirror that projected the image back onto the front of the screen at the top. Not sure how well I'm describing it, but I think it might have been a Sony. Maybe someone else remebers it.

Darryl[/QUOTE]

I think I remember a design that was actually 3-color, but two of the colors were combined optically through one lens???? Could it have been something like that?

Weren't the 1st color TV shots from the Moon-Apollo 12-NOT 11-done in field sequential color? Seems like I remember some sort of blather about that-may have dreamed it. 1969 was a LOOOOOOONG time ago...-Sandy G.

It's possible as I did not have the opportunity to look inside. I do remember that the two colors where cyan and a magenta/orange which would make sense. I think I remember being able to see the targets inside the crt's, but it was so long ago. I hope someone else remebers this set.


And you are correct about the 60Hz power line on a 24Hz television. This could cause all sorts of misregistration between fields. There was also a discussion last year about how relatively quickly the color filters would have faded in use, and since there was no tint or gain control, nothing could be done at the user end, so changing your filters may have become a common service call.
I also didn't realize that the CBS sets did not have a way of locking to the color signal, and you had a one in three chance of it coming up correctly. This means that just changing the channel between two CBS color stations (not that this ever happened) meant you had a 33% chance of the colors coming up right! The Gray Research monitor added a "red field" pulse to phase lock it's color wheel automatically, and it was pretty cool seeing this in action on Steve's set. I assume CBS would have adopted this method had they won out.

Darryl