Olympus-OM
[Top] [All Lists]

[OM] Re: Questions, Re: When is enough enough? ...

Subject: [OM] Re: Questions, Re: When is enough enough? ...
From: AG Schnozz <agschnozz@xxxxxxxxx>
Date: Mon, 18 Oct 2004 08:56:57 -0700 (PDT)
> Do you know where I could get a copy of the chart that you are
> using ? I wanted to do some
> resolution testing myself. Maybe a scanned tiff file that I
> can print out ? How does the
> experiment look, i.e. distance from the lens, etc. ?

I've got the PDF on one of my computers.  When I find it I'll
post it up on the website.

For most of my lens tests I printed out five of them and placed
them in an "X" pattern so I can check center and corner
resolution.  Mount the lens to test and place the camera close
enough to get all five charts squared up in the viewfinder.  I
also have a color/gray scale chart in there too for metering
sanity. If at all possible, use flash to expose the charts.  A
studio flash with variable power output works very well. Just
make sure the placement is such that you don't get flash
reflection off the charts.
 
> If you translate this to Mpixels, on 35mm Film this gives
> roughly anywhere between
> 800 kPixels and 8 MPixels.When people compare film to digital
> they mostly compare film
> grain to chip resolution.I never understood this as the above
> numbers show that the lens
> should be the limiting factor on a digital camera with more
> than 5 MPixels
> (> 100 lns/mm is rare, and easily impacted by attached
> filters, dust on the lens, etc).
> I always thought I missed something as I have seen good larger
> prints of photos taken
> with lenses with a resolution of less than 50 lines/mm.
> Forgive me if I just misunderstand
> the experiment setup used to collect the above resolution
> information.

First of all, with the 5MP digicam I DO experience a loss of
resolution with filters.  The 2/3" format is so "on-the-edge" of
performance limits of the lenses that just about anything will
skew the results.

My resolution tests didn't prove much, other than an ideal
analog setup (with extremely fine-grained film and flash
exposure) outperformed the same test in 2/3" digital by 6-8
steps.  It was a benchmark test for my own sanity.

However, it did prove something else.  When enlarged to obscene
sizes, the analog outperformed the digital because of the
natural occuring artifacts (random by nature, and finer than
what the lens was able to resolve.  Digital, by comparison, is
as sharp up to a limit. As soon as you hit that limit, it
becomes unnatural and does not withstand further enlargement.

I'm looking at an 8x10 at my desk, right now, taken with the
digicam. It is, by far, the sharpest photograph you've ever
seen, with details in places that I've never experienced before.
 However, at 11x14 it croaks.  It's still sharp, but not
pleasing. Absolutely worthless. In comparison, there is an 11x14
B&W that isn't as sharp, but withstands enlargements of any
size.

1. Grain is grain, but noise is noise.

2. Random grain is like sand--it's natural, expected, and easily
ignored (or accepted) by the eye.

3. Digital noise almost always has a pattern--it's unnatural,
unexpected, and not easily ignored (nor accepted) by the eye.

AG



                
_______________________________
Do you Yahoo!?
Declare Yourself - Register online to vote today!
http://vote.yahoo.com

==============================================
List usage info:     http://www.zuikoholic.com
List nannies:        olympusadmin@xxxxxxxxxx
==============================================

<Prev in Thread] Current Thread [Next in Thread>
Sponsored by Tako
Impressum | Datenschutz