Archived posting to the Leica Users Group, 2000/04/08

[Author Prev] [Author Next] [Thread Prev] [Thread Next] [Author Index] [Topic Index] [Home] [Search]

Subject: RE: [Leica] Analog vs. digital (long and barely OT)
From: Henry Ambrose <digphoto@nashville.net>
Date: Sat, 8 Apr 2000 23:28:39 +0100

Austin Franklin wrote as a reply to my statements:
>
>You don't "need" pixels.  Pixels are "picture elements" and is an 
>electronics term, it has nothing to do with printing, per se.  Halftoning 
>has been used for years, long before 'pixels' were ever thought of...  In 
>halftoning, the dots are referred to as, well, dots.  The dot is not the 
>same as the pixel, they are two entirely different concepts.

Yes, I need pixels and you do too if you're gonna use a computer to 
convert your photo to a halftone. I did not mix pixels and dots. You did. 
You are right, halftone dots have been around for quite awhile. That's 
how continuous tone images have been reproduced for a long time.
>
>> The usual resolution for average (133 to 150 line screen) magazine or
>> brochure printing is 300 pixels per inch. Thats 300 pixels per inch at
>> the final reproduction size.
>
>You are mixing about pixels and line screens (halftone term) here...and 
>they aren't related in this sense...

Yes they are related. What I stated is a generally accepted, empirically 
proven conversion for the resolution needed for offset printing. It is 
standard in the industry. It is not absolute and there's no math to prove 
it - only real world experience. How many print jobs have you managed 
over the past 10 years or so? You have some experience to back up what 
you say, don't you?
>
>> A 4 inch X 5 inch image will require
>> 4X300=1200 pixels by 5X300=1500 pixels. Measured in pixels you'll want a
>> 1200X1500 pixel image. For black & white you'll need 1,800,000 pixels or
>> as commonly expressed, 1.8 mb.
>
>Black and white has gray scale, and you still need some number of bits per 
>pixel (if you are talking about pixels), unless you want a monotone image.
>
Yes, and we need 8 bits for a 256 level grayscale to print by offset. I'm 
refering to output. You might (hopefully) start with something higher, 12 
or 16.
Anyway - it just 8 bits when you make film to make the plate. The 
resulting 256 levels of gray makes a convincing reproduction.
>
>> When digital sensors catch up to film I want a lens that will be
>> excellent. Just like I wanted better lenses to use with film, I want the
>> best lens to image digitally and on film.
>
>They are different requirements.  Physics governs the minimum size of both 
>a pixel and film 'resolution'.  The problem is a pixel is going to be much, 
>much larger than film resolution.  Why this is important is the sensor will 
>have to be larger to get the same 'resolution' (what same means is up for 
>debate, because film grain is not in a fixed xy pattern like an image 
>sensor is...so you would need to scan at some factor higher resolution than 
>film using a fixed xy pattern to actually faithfully duplicate the true 
>resolution of film...but that's a different discussion)...  Point is, the 
>same nice Leitz lense you have that renders your 35mm negative beautifully, 
>may not work as well on a digital sensor as you may think.
>
Then what lens will work as well? What's your experience with lens and 
digital cameras? Its still light going through glass.

"Physics" does not govern the minimum size of a pixel or film resolution. 
Our current level of technological development does. Density of sensors 
and resolution of film grows and has been growing since they were 
developed.

Henry