MTF/CTF and image quality

[Herve RICHARD]

Dear all,

At the moment I'm specifying a projection lens for colored image
projection.
These images will be seen by human eyes at a min distance of 3m.

In order to specify the image quality and forgetting color aberration
issues I guess that I have to specify MTF values for the lens.

My question are:

-what is the most relevant spatial frequency values on which I have to
specify the MTF level (and also what is the most relevant MTF level:
3%, 10%, 50% ?)?

- I heared that a good quality criteria for image is the MTF area
concept: it's the area between the lens' MTF curve and the eye's CTF
(Contrast Threshold Function) curve. The area is directly related to
the amount of information which is conveyed to the viewer.
Is anybody aware of this concept?
Does anybody have the eye's CTF function?

Thank you very much for your help.

Best regards.

Herve RICHARD.

[Nelson Wallace]

Herve -

Here's my shot at your question. I suspect there will be other
opinions.

The average human eye (whatever that is) can resolve about an arc
minute. At 3 meters that's about 0.872665 mm. Assuming that's one bar
of a dark bar / light bar pair, that's 1.745 mm per cycle. Which is
0.57 line pairs per mm if I did the math right. That's the limiting
spatial frequency that you want to spec your projection system at.
And most people like a system MTF of about 10% or better at the limiting
frequency.
If people are closer, redo the math.

You may want to look at my web page:

http://home.earthlink.net/~nwwallace/MTF.htm

And remember, I'm talking system and not optics, so you need an MTF
budget that has an optics MTF that's higher than the system MTF.

Regards,
Nelson

[Steve Eckhardt]

Nelson's comments are very helpful. I would only add a couple things. First
is that MTF is usually specified at the slide, not the screen. This means that
you will have to specify the MTF not at 0.57 cy/mm, but at that frequency times
the magnification of the system. For an overhead projector, the magnification
is about 6X (1.5m wide screen, 250 mm wide stage), but for an 0.7" DMD on a 1.8
m diagonal screen, the magnification is 100X. So you'll need to be specifying
the MTF at somewhere between 3.4 cy/mm and 36 cy/mm. (Actually, some newer
LCOS imagers have 8um pixels, so they require good MTF at 60 cy/mm.)

The other question is what constitutes "good" MTF. I think Nelson is being
somewhat lenient with 10% MTF. I'd never let a supplier get away with less
than 20% MTF for the system, as manufactured. That's only 1.5:1 contrast! The
lens design should have a theoretical MTF of at least 35-40% at the limiting
spatial frequency. As manufactured, I'd accept 25-30% MTF for a projection
lens.
--
Best regards,
Steve Eckhardt (skeck...@mmm.com)

In article <3C14E857...@earthlink.net>, nwwa...@earthlink.net says...

[Herve RICHARD]

Hi Steve,

Thank you (thanks also Nelson) for your help on this matter.

At the moment I've got an aditionnal comment:

In your response, you explained that the quality criteria for good MTF
(and so for good image) is the MTF value (35-40%)at the human
eye limit resolution. But is it for you the only image quality
criteria? Don't you think that the difference between the lens MTF
curve and
the CTF eye function could be another good criteria. The eye CTF
function is the minimum image contrast needs to the eye to make image
distinguish vs spatial frequency (it's a J shape function: ie the
function reaches its min value for a given spatial frequency). My
concern is that I don't have the CTF value. Perhaps you have it?

Regards.

Herve.

skeck...@mmm.com.deletethis (Steve Eckhardt) wrote in message news:<9v2r0i$84m$1...@magnum.mmm.com>...