[old_tv_nut]

The big changes in color image quality were in several areas, some improving gradually and some in steps.

Color phase and intensity - tended to vary considerably with tube broadcast gear, becoming more stable as transistorized equipment came in.

Signal to noise: this was marginal with the original color cameras, taking a big leap when Plumbicons were introduced and another big leap with solid state pickups. Image orthicon noise, being independent of gray level, showed up strongly in the shadows due to being amplified there by gamma correction. Noise near black is also partly rectified in the picture tube, resulting in a raising of the picture average black level and hiding shadow detail. Besides the basic image orthicon noise, the early camera preamps were contributors, later improved by converting to transistor preamps. The noise level influenced how much gamma correction was practical, so the earlier cameras had much less contrast range, losing details in the shadows. This was compounded by the light-colored screens of the early color picture tubes, which reduced the contrast in the home unless you turned out the room lights. The lack of 100% DC restoration in receivers also contributed to contrast problems, dark scenes getting washed out towards gray and very bright scenes clipping the blacks. In a side by side comparison of pictures with late analog gear end to end vs. early gear, you would immediately be struck by the improved contrast and shadow detail.

As a result of all this, early cameras did not have full gamma correction, but used the same technique Kodachrome and Technicolor used of increasing the contrast to make up for the loss of contrast and color saturation due to all the effects that tended to wash them out. This works OK if the camera exposure is adjusted just right, but makes it more critical than it would otherwise be. This method of increasing saturation was necessary because the early cameras did not have color matrixing, which would have introduced additional noise. Plumbicon cameras included matrixing to make the color saturation match the CRT phosphors independently of gamma and contrast adjustments.

Geometry, color shading, yoke ringing (alternate light and dark bars near the left side of the picture), and RGB registration drift were all problems with the image orthicon cameras. RCA improved these things as time went on by building more precision orthicon magnetics. Another shortcoming of the early cameras is that the shading adjustments were added waveforms only, mainly affecting shadows, while Plumbicon and later cameras could have both additive and multiplicative shading adjustment. The high end studio cameras eventually had built in optical test patterns and control software that automatically adjusted all these things; plus, the lens characteristics were programmed in so that lens distortions were automatically compensated as the zoom setting was changed.

The problems with shadows, I think, were a main reason why the early color experimental broadcasts depended on a human subject (Marie McNamara, "Miss Color TV") for a final set-up check. Not only her skin tone, but especially her auburn (dark red) hair would have been affected by shading and RGB black level balance problems.

Analog video tape also contributed to noise and various beat patterns, again improved gradually over the years, and eventually replaced by digital recording.

If you want to see some of the early stuff, look for DVDs of the Bell Telephone Hour, the color episodes of the Howdy Doody show, and/or Peter Pan with Mary Martin. It is interesting to see what gross errors were hidden in the corners of the pictures in some cases, and were not visible at home due to the round picture tubes.

The early image orthicon cameras were very finicky to set up, and some studios' crews learned to do it better than others. I think it is generally acknowledged that NBC on the west coast got more consistently good results than NBC New York.

There were additional more minor improvements in going from Image orthicons to Plumbicons to solid state pickups. Image orthicons had problems with bright highlights (reflections from jewelry or brass instruments) producing a dark halo. In monochrome, this was deliberately used to give an approximate gamma correction, but was not good for color, where the halos would be in different colors. I remember seeing programs where someone wore a bright red dress or shirt that caused a cyan halo to spread around it and onto their face. Plumbicons had problems with extreme highlights causing "comet tails" trailing the moving bright spots. Plumbicons were also subject to differences in resolution of leftward diagonal lines and rightward ones, due to a phenomenon called "beam sharpening."

In a nutshell, things improved gradually or sometimes in steps over periods of years, and at a slow enough pace that most people could not accurately recall what the quality was like five or ten years earlier; but a comparison of the beginning and end should be instantly obvious to the most naive viewer.