Hoya filters, image quality, resolution
 

Hi,

I found a web page that told that "fotoMagazin" did some filter resolution tests and published the results in the issue 09/2001.

Hoya Pro 1 filters: 160 lp/mm, "soft" contrast
B+W MRC slim: 300 lp/mm, "high" contrast
Heliopan SH-PMC: 400 lp/mm, "high" contrast

I do not know what these numbers mean in practice for a videographer.

If I put a Hoya Pro 1 filter (or two) in front of my camcorder
( Sony PDX10 with 1/4.7 inch one megapixel ccd x 3, f=3.6-43.2 mm ),
then will there be visible image degradation? Or are the higher resolutions and contrasts only usefull for still photography?

The numbers represent "line pairs per millimeter". I will assume those numbers were what was captured for this test. A line pair is a pair of black and white lines. So, in one millimeter, they were able to resolve up to 400 pairs of lines. Even the 160 is well beyond what video can do.

Hoyas are OK for 35 mm still film work.

You should know that anything you place in front of the lens causes degredation. That's why people who want the absolute best, buy the best. To get into line-pairs and modulation transfer function (which really tells you a lot more) is way overboard for most of us.

However, any of the mentioned filters will not noticably degrade a SD image.

"Even the 160 is well beyond what video can do."

I assume that you have worked with 1/3 inch chip camcorders. I have a miniature chip "wonder". Doesn't a small chip require optics with a higher resolution power? The image on the CCD will be a lot smaller...

I want to understand - at least to some extent...

---------------------------
It would also help to know how thick B+W filters and Heliopan filters are (filter thickness beeing an issue for me). I could very well take B+W or Heliopan just to make sure I never need to replace them (for digital still cameras or a HD cam after a couple of years...)

My camcorders chip is 1/4.7 inch = 0.21276 inch = 5.404 mm
Because the CCD is 4:3 it will be roughly 3.2 x 4.3 mm

Let's assume my camcorder has an infinitely long macro :)
Then the whole image comes straight, through a 3.2 x 4.3 mm area on the filter.

It's a one megapixel 4:3 chip (1070 000 actually). Thus it is very roughly 1154 x 866. This means 270 pixels per millimetre.

-----------------------------------------------------------------
Thus: Filter: 160 LP/mm. Required: 270 PIX/mm maximum
-----------------------------------------------------------------

Does a line pair correspond to one pixel, even roughly?
Is this a valid approach at all?

Would this not indicate that nearly 300 lp/mm would be required for long macro? The PDX has 12x (up to 492mm/588mm)...

Or is a line pair equivalent to 1.5 or 2 pixels?

Ralf, your chip is much smaller than you assume. The name of the CCD (1/4.7) is not the diagonal of the chip. I am not referring to active pixels either. The sizes are remnants of a 50+ year old method of measuring the image areas of tubes. Your chip is close the the size of a 1/5 inch chip, which measures 2.6mm X 2.2mm (L x W) X 3.2mm diagonal.

Since fotomagazin is not in English (or French or Latvian !), I'll have a tough time reading it. The filter models you mentioned are of the "slim" variety. Does any info exist on the B+W, Hoya
and Heliopan filter resolution for the regular models?

Thanks !

Gints

Some years ago I was told by a B + W rep at PMA that the glass was the same, only the metal filter ring was different. That, certainly could have changed.

Ralf,

A line pair is not equivalent to pixels in any way. It is a measure of resolving power.

What is done is a target of known size is set in front of the lens a known distance away. A photograph is taken of the target. You then view the image to see if the line pairs are visible.

To calculate all this, you can take a target image and set it in front of a camera one meter away. If you are using a 50mm lens, then the image of the target will be reduced by the ratio of 50mm to 1 meter or 1/20th the size onto the film or ccd.

Now, if your target has alternating white and dark stripes on it that are 1mm wide, ie, total width of a white and dark pair is 1mm, then each line pair projected onto the film or ccd will be 1/20mm wide. So 20 line pairs fit in one millimeter which equals 20lp/mm.

If you can view this image with your camera then you are getting that MINIMUM resolving ability.

Notice at no time did this calculation involve pixels and at no time did my hands leave the keyboard.

In optics "line pairs"is being used. In TV when talking about resolution "lines" are used. Two "lines" equals a "line pair" On the filter "resolution" I never heard about that fairy tale. There are many specs related to optical flats (like filters) and in best case one could talk about filter MTF. Though no commonly used for flats, it could reflect some impurities, flatness issues...resulting in contrast reduction, but this never sets the "resolving" power to limits like -10 or -20 db on the MTF representation. I would love find this article..

Thanks for the information. This is becoming technical...

All I need to know is whether I should

- get the Hoya S-HMC / Hoya Pro 1 series filters that are thin and multicoated and rated at 160 lp/mm and *low contrast* (whatever it means...)

or

- get multicoated B+W filters ( 300 lp/mm and *high contrast* )


I would use them with my current camcorder and maybe with a Canon Powershot Pro 1 or similar 8 MP digital camera.

It has been told that 160 lp/mm would be enough... at least for digital SD video. Still I feel a bit uncertain and undetermined, sorry. Well, what about still photography? What about contrast?

I would like to read about personal experiences with the Hoya filters to get an overall fell of them.

I would get the high contrast. This means that the general filter flare is low (high MTF levela)...but forget lp/mm for filters