1951 CBS Color Broadcast Screen Shots
 

Most of you probably know already about Google's many hidden search engines and digitized content, such as Google Books. Many magazines are available on Google Books including Life, Popular Science, Popular Mechanics, and Billboard. All of them have a large number of articles on early tv.

One thing you might not know about is that Google scanned Life Magazine's photo archive. These are not just photos that made it to print, but all the ones that didn't from the same event. A few years ago I ran across Life's article on CBS's historic first commercial color broadcast on June 25, 1951. Here is the original article on Google Books: http://books.google.com/books?id=l04...20life&f=false

They published six screen shots that look pretty good in the magazine. But, go to the Life archive and you'll see all the photos taken from that event: http://images.google.com/hosted/life...be65815c8.html

If you click on the thumbnails on the right you can scroll through the collection of about 40-50 screen shots. On some the color is evident, on many they aren't. One shows a peacock, so I guess they beat NBC in more than one way! If you look at the photos printed in the magazine, some were reversed and much better color than the originals. Probably Life enhanced or retouched them, a common practice back then.

I haven't seen these anywhere else on the web. There are many pictures from behind the scenes of this broadcast, and the six printed photos have made their rounds. But, I thought the members here would like to see the rest and get a better idea of what was broadcast that night.

That Life photo collection, like Google's other "hidden" collections, is great but frustrating. There is no rhyme or reason to the search results, and they seem to change from day to day. That said, it's a gold mine, or a huge time waster. I'm never sure of which. I like the attached, which comes up under a search for "television set." At least it does today.

Here is a film taken off the screen of a CBS field sequential receiver in 1949:

http://www.criticalpast.com/video/65...n-Monitor-Tube

Cool! Thanks for the link, I've been looking for footage of Patty since I saw an article showing her and Ed Sullivan in a 1950 Popular Science collection of articles.

These pictures have been scanned, but no effort has been put into restoring the faded color. I am certain that the original color must have been better, and that justifies some restoration effort if someone is willing to tackle it. (I may give it a try when I get home).

Actually, these were almost certainly shot on Kodachrome color transparency film as that was the only color film magazines used then. And, Kodachrome film simply doesn't fade. Ektachrome was pretty new then and it might have been used because it was a faster film. But, Ektachrome shifts to the red over time and there is no evidence of that on these slides. The picture of the test pattern is still very balanced.

Because some of the pictures show decent color and some don't, my guess is it was either an issue with the receiver or the exposure setting for the film.

Ralph Morse is credited for these photos, and he was probably Life's best staff photographer at the time. Here is a wikipedia link about him: http://en.wikipedia.org/wiki/Ralph_Morse

I was lucky enough to see one of those sets in operation a few years back, and the picture was outstanding. I would NOT have believed it had I not seen it.

I respectfully disagree. The lowlights are quite strongly purplish red, as a check of the histograms will confirm. Of course, it is possible that the TV images looked like this, but I suspect not, as the CBS system was capable of excellent color, as was reported to me directly by an eye witness to CBS demos, and stated in the press at the time, and has been confirmed (at least on the receiver side) by Cliff Benham's experiments.

Here are some first attempts at color correction.

The Bisquick picture gives a shaded gray background as a reference. Based on this, I adjusted levels to get a decent grayscale. So, there is a reasonable basis for this adjustment.

In the second adjustment, I made a hue and saturation adjustment on top of the levels adjustment. This does not have any justification except my own eye.

Then, I show the same three stages for the test pattern, using the same adjustments as for the Bisquick.

Finally, a slide with people. Here it appears the hue adjustment is too strong (towards yellow) in the highlights, but the shadows on the man's face are close or a bit reddish.

The couple picture

Wayne,

Was there some significance to the middle, mostly green, image? Perhaps it's just an iteration on the way to the third, mostly blue, image? It piqued my interest because a less-than-perfect quadrature transformer alignment in a CTC2 can cause a blue/green problem.

The middle shot (file name ending in "levels") has the levels corrections the same as all the other shots with file name ending in "levels." So, the gray scale is corrected (mostly) the same as the Bisquick "levels" correction. All of these seem to reproduce the blues too greenish, as you can see on the test pattern. Is this due to the film's reaction to the CBS blue primary? I don't know.

There could be a hue shift due to carry over of successive images due to lag in the image orthicon. In fact, the CBS color sequence was red, blue, green, so camera lag would produce the direction of shift shown in the test pattern - reds become mixed with a little blue, blues with a little green, green becomes yellower, yellow shifts towards orange. From the eyewitness report I have heard, though, color was very good, so it is hard to believe that the shifts were as large as shown here - but just maybe it's possible.

We can't know what color the curtain behind the people was, can only guess that the correction of the test pattern is also right for the live scene.

Of course, there is no subcarrier and therefore no quadrature adjustment in the CBS system.

Thanks for the added info. Quick followup question: Is the 'gray' no-seam behind the Bisquick box an assumption, albeit a good one?

It's partly an assumption, but it's backed up by the fact that the black area of the photo outside the screen is similarly faded - harder to see by eye because it's still quite dark, but shows plainly in the histogram.

By the way, part of the quality loss in these pictures is due to the high compression settings of the jpeg. This means the color is pretty blocky, so that the correction may be too much / too little depending on the innacuracy of the coding. If you look at the test pattern 3rd version (with the hue/saturation change), you can see artifacts being amplified by the correction process. I wonder if there are higher quality files somewhere, or if Google went to this high compression without saving them.