Videokarma.org TV - Video - Vintage Television & Radio Forums

Videokarma.org TV - Video - Vintage Television & Radio Forums (http://www.videokarma.org/index.php)
-   Television Broadcast Theory (http://www.videokarma.org/forumdisplay.php?f=182)
-   -   Test pattern? (http://www.videokarma.org/showthread.php?t=252977)

jr_tech 12-28-2011 12:56 PM

Quote:

Originally Posted by Kamakiri (Post 3022369)
This brings up an interesting point. What was the meaning of the numbers in the diagram? While they may be measurements of some kind, one can only imagine that it's actually a points score for a TV based bean bag toss :D

I believe that they define the resolution scale in various parts of the "wedges" from 200 to 450 lines in the center and 200 to 350 at the corners

jr

old_tv_nut 12-28-2011 02:28 PM

Quote:

Originally Posted by jr_tech (Post 3022410)
I believe that they define the resolution scale in various parts of the "wedges" from 200 to 450 lines in the center and 200 to 350 at the corners

jr

Yes, resolution in "TV Lines per picture height." 330 lines is the most you could get on a broadcast signal with a black and white set (or color set with comb filter). Maybe 270 lines horizontal with a color set with 3.58 MHz trap.

jr_tech 12-28-2011 04:43 PM

So one would measure Horizontal resolution in terms of "lines per picture (vertical) height"? seems strange! :scratch2:
I would think that normalizing the measurement to CRT diameter/diagonal or width would produce data that would be more consistent... "portholes" got a big break with the "lines per picture height" definition.

jr

W.B. 12-28-2011 05:37 PM

Quote:

Originally Posted by jr_tech (Post 3022410)
I believe that they define the resolution scale in various parts of the "wedges" from 200 to 450 lines in the center and 200 to 350 at the corners

NBC, oddly enough, had a layout of their famed "bullseye" pattern in which they explained what the significance of the wedge calibration dots as positioned on each side of their horizontal and vertical wedges, were:

https://live.staticflickr.com/65535/...c85c24db_o.jpg

Alas, I haven't heard, seen or read anything about this variation being seen on the air.

NOTE: 200 TV lines = 2.5 MHz (or "mc," as it was known as of 1947).

Kamakiri 12-28-2011 07:10 PM

Quote:

Originally Posted by jr_tech (Post 3022425)
So one would measure Horizontal resolution in terms of "lines per picture (vertical) height"? seems strange! :scratch2:
I would think that normalizing the measurement to CRT diameter/diagonal or width would produce data that would be more consistent... "portholes" got a big break with the "lines per picture height" definition.

jr

Well not only that, but imagine the technician with a magnifying glass trying to count each line to be sure of the perfect picture!

old_tv_nut 12-28-2011 08:03 PM

Quote:

Originally Posted by jr_tech (Post 3022425)
So one would measure Horizontal resolution in terms of "lines per picture (vertical) height"? seems strange! :scratch2:
I would think that normalizing the measurement to CRT diameter/diagonal or width would produce data that would be more consistent... "portholes" got a big break with the "lines per picture height" definition.

jr

Doesn't matter what you normalize to, just so you normalize to something. Probably because the scanning lines are arranged vertically one below another, the vertical dimension was chosen. "330 lines per picture height" of horizontal resolution gives 440 total "lines" of resolution per picture width since the width is 4/3 times the height.

bandersen 12-28-2011 09:42 PM

Aren't there always 525 lines (262 1/2 interlaced) vertical ? Minus a few for closed caption and other info that is.

old_tv_nut 12-29-2011 09:25 AM

Quote:

Originally Posted by bandersen (Post 3022455)
Aren't there always 525 lines (262 1/2 interlaced) vertical ? Minus a few for closed caption and other info that is.

OK, this is a very old concept, but here goes:

There are 525 total lines, but only 483 active lines due to vertical blanking. Let's just call it 480 for rough calculation (which also matches a common computer image format).
Because the lines sample the picture vertically (just like the rows of pixels in a digital picture), they cannot provide a full 480 lines of resolution for ordinary pictures - they could only do that if the picture details happened to line up exactly with the line structure. The ratio of practical viewable resolution is called the "Kell factor" after Ray Kell, who published the original research in 1934. There is an additional factor due to interlace causing interline flicker - how strong this is depends on the observer and the amount of detail contrast in the image. The first NTSC (1941) adopted an over-all factor of 0.7 for the achievable resolution compared to the number of active lines. Along with this they specified a combination of baseband video signal bandwidth (4.2 MHz) and active scan lines (486) that would give equal resolution per picture height vertically and horizontally, and also fit in a 6 MHz channel using vestigial sideband modulation and including a sound carrier. 480x0.7 = ~330, the accepted usable resolution - your mileage may vary, and there have been a lot of useless arguments over the years as to whether the useful vertical resolution number is exactly correct.

A few of the active lines were later devoted to other things like test signals, data, and closed captions, leaving 483 containing picture.

jr_tech 12-29-2011 01:04 PM

Interesting info! Thanks!
Does this Kell/NTSC facto apply in about the same way to DTV or must another factor be used? Perhaps a different factor for 1080i than 720p?

jr

old_tv_nut 12-29-2011 08:32 PM

Quote:

Originally Posted by jr_tech (Post 3022498)
Interesting info! Thanks!
Does this Kell/NTSC facto apply in about the same way to DTV or must another factor be used? Perhaps a different factor for 1080i than 720p?

jr

HDTV makes a little difference, in that the line structure is smaller/finer compared to the overall picture size. Some of the effects of interlace can be reduced with progressive display, as on a plasma or LCD TV, but the conversion to progressive is never perfect. Thus, 1920x1080 interlace generally has noticeably better resolution on static scenes, but 1280x720 progressive generally looks better on sports.

The best for resolution and motion at the same time would be 1920x1080P at 60 frames per second, but there are no sources at that rate. Blu-Ray discs carry 1920x1080P movies at 24 frames per second, which is converted to 60 frames for display - this is close to the ultimate depending on how well the conversion is done.

By the way, digital images have the same sort of factor in the horizontal direction because they are made of discrete pixels in both directions.
Without interlace, the factor is perhaps 0.9 instead of 0.7. The SMPTE standard for digitizing NTSC or PAL component video has filters that start to roll off (70% response) at a factor of 0.85. I estimate that for a test pattern you can still see the resolution in the wedges when the amplitude is down to 10%, which is probably around a factor of 0.9. So, a system with 640x480 pixels scanned progressively and displayed progressively would have a useful resolution of about 430 TV lines per picture height (lph) both vertically and horizontally. The 1280x720 progressive standard gives about 648 lph both vertical and horizontal, and the 1920x1080 interlaced system gives about 972 lph horizontal resolution, and a variable amount of vertical depending on camera settings and motion, nominally 756 lph.

jr_tech 12-30-2011 12:51 PM

Wow! thanks again! I am always amazed by the depth of knowledge to be found on this board!

jr

bandersen 12-30-2011 01:54 PM

I thank you as well. I think I understand the concepts now.

Kamakiri 12-31-2011 07:49 AM

The information in this post I felt was so invaluable, that I created a forum specifically around the topic. Thanks to old tv nut for such an eloquent explanation!

4343 06-09-2012 06:59 PM

Quote:

Originally Posted by Kamakiri (Post 3022437)
Well not only that, but imagine the technician with a magnifying glass trying to count each line to be sure of the perfect picture!

In the 80's, that was me. We used a digital monoscope (minus the Indian head) that had 800 lines resolution wedges in the horizontal for checking tube cameras that were often specced out to 700 lines (for luma only).

I used that generator to align my 25" Sony Profeel Monitor to get the minimum overscan possible when I bought it in about '87. Last year I put a tape of the test pattern on it and could see no change in picture size...

W3XWT 07-25-2012 07:48 PM

On the color wedge test pattern, my favorite was WOLO-25, which had its city of license in the lower right-hand quad as "COLUMBIA, SOUTH, CAROLINA". It was also one of those stations that attempted to cool the klystrons with tap water. Hard tap water at that!


All times are GMT -5. The time now is 02:56 AM.

Powered by vBulletin® Version 3.8.4
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
©Copyright 2012 VideoKarma.org, All rights reserved.