Newsgroups: sci.image.processing
From: paul@pcserv.demon.co.uk (Paul Carpenter)
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!zombie.ncsc.mil!paladin.american.edu!howland.reston.ans.net!pipex!peernews.demon.co.uk!pcserv.demon.co.uk!paul
Subject: RE: TV Resolution?
References: <3lh6iv$f9q@sue.cc.uregina.ca>
Organization: PC Services
Reply-To: paul@pcserv.demon.co.uk
Followup-To: sci.image.processing
X-Newsreader: Demon Internet Simple News v1.29
Lines: 92
X-Posting-Host: pcserv.demon.co.uk
Date: Fri, 31 Mar 1995 21:10:00 +0000
Message-ID: <796684200snz@pcserv.demon.co.uk>
Sender: usenet@demon.co.uk

genowpa@leroy.cc.uregina.ca "RCMP Forensic Lab Regina" writes:
-My apologies for not including the original posting -- I lost it   
-somehow.   
-Cedric,  
- 
-I think that you will get a variety of answers to your questions   
....
 
-You gave an example of Pulnix camera specs as "Pixel 768(H) x   
-494(V) EIA Scanning 525 lines  TV resolution 570(h) x 350(v)   
-lines".  In my experience these specs are not unusual for   
-scientific CCD cameras. The 768x494 does describe the pixel   
-number and pattern and 525 lines corresponds to the standard NTSC   
-video (TV) signal. 

Or RS170

 One thing to remember is that most cameras,   
-both CCD and tube, produce an analog voltage signal (TV signal).   
- The exceptions are non-standard or direct input digital cameras   
-like the Kodak MegaPlus.

The output is digital as the A/D is in the camera. Roll on Digital Cameras with
framestores. Non-standard can include line scan analog/digital output, but the 
process is analog at sensor it is just when it becomes digital.

  My understanding is that the "nominal"   
-RS-170 horizontal scan is 525 lines per frame.  However, this can   
-vary around that figure quite a bit and generally exceeds it   
-for CCD cameras; in the above case it's given as 570.

If it exceeds the 525 it is not RS170 but yet another standard as the vertical 
retrace and/or horizontal frequencies would have to alter. The Camera may have
570 lines but only 525 will be output for RS170/NTSC or corresponding number
for other standards (CCIR/PAL/SECAM).

  From a TV   
-viewer's point of view, the useable portion of that signal is   
-actually much smaller than 525, only 480 lines or less.  Any   
-excess is lost in retracing.

The excess IS the retracing 2.5 lines Vertical sync.....

  On the other hand, a good   
-image-processing framegrabber is capable of capturing ALL of   
-those lines and using them in its processing.   

All image processing framegrabbers use all lines (525/625) for timing amongst 
other things but will only store the lines (or subset of them) containing 
picture information. The rest is a waste of time unless you are only analysing 
the retrace lines or encoded data like Teletext.

-Now,as for the 350 vertical figure I'm a little less sure of the   
-underlying relationship but I believe that it relates to the TV   
-bandwidth limitation and it's effect on the number of brightness   
-values along each of the horizontal scan lines.  Given an approx   
-4-Mhz bandwidth limitation for all the information in a TV   
-signal, the number of separate voltage values distinguishable on   
-each TV scan line is around 330.

I personally think you will find that it is 340, as in 480 lines and a PICTURE
aspect ratio of 4:3 (H:V) giving a 680 resolution for square pixels when 
digitised for image processing.

TV Video bandwidths

	NTSC	4.2MHz
	RS-170	5.5MHz	(need to check that one)
	PAL	4.2 to 6.0 MHz depending on subset used for country and
		transmission method VHF/UHF etc..
	SECAM	5.5 to 6.5MHz
	CCIR	6MHz

Often the excess is blanked by timing or optics (electrically or optically 
black).

  This effectively limits the   
-resolution to approximately that number across each line but, as   
-with horizontal resolution, CCD's produce more values as a   
-general rule.  And a frame-grabber can capture all of those   
-values.  

If it is set up to grab narrow pixels (vertical rectangles) and the actual 
camera characteristics are known. Not much use from *standard* signals where 
relationships are determined more by the timings than the number of elements in 
a sensor. A 1024x1024 sensor could produce a standard TV output but sampling it 
at higher than 640x480 still won't change the fact of pixels being square or
necessarily usefulness, as large part of the sensor probably would not be used
in the picture information.

-- 
Paul - "Any people you should meet are the products of a deranged imagination"
