Why is a filter on the lens ok, but a filter on the sensor not?

[peter_white2]

Leica wants us to put IR filters on our lenses. But they don't want to put the

same IR filter in front of the sensor. This doesn't make sense to me. Why can't

the same filter be put in front of the sensor? Leica says that it will degrade

image quality. But won't it have the same effect of degrading IQ when it's in

front of the lens? And wouldn't the problem with the corners of wide angle

images going cyan be eliminated or at least reduced if the filter were at the

sensor?

[Rob F.]

There is an IR filter over the sensor, but they kept it thin to avoid the optical porblems of a thicker one. It would have degraded image quality. A filter in front of the lens degrades the image less than one placed between the lens and the image plane. I think that a thicker filter over the sensor would have introduced more problems at the edges and corners, rather than solving problems.

 

Just my understanding, based on reading the discussions so far.

[johnny massey]

I wish I'd had the money and sense to buy shares in B&W before all this kicked off . . . .

[mark_tupper]

Or you could buy a Nikon D2x - a digital camera that works properly in the 1st place.. add on filters to make it work ?? - what a joke camera !

[john_layton]

Seems like some pretty unbendable rules of optical physics are in play here - and that

perhaps this "solution" of a thin sensor cover and the (optional) use of IR filters is actually

the most responsible path for Leica to follow. A previous poster may have had it right by

suggesting that had Leica opted for a thicker CCD cover, which may have eliminated the

need for an IR filter on the lens - there might be even more complaining than there is now,

due to a possibly noticeable degradation of image quality, which would be "irreparable."

My only concern would be that the IR filter would allow for maxumim light transmission

(no filter factor) and that the benefits of using the filters (where recommended) would

outweigh the potential downside of a bit of extra flare.

[vinay_patel]

I wouldn't beat myself up John. All the IR filters that are being or will be bought as a result of the M8 aren't likely to affect B+W's bottom line to any degree. The major Leica dealers are livid, as they had expected a windfall holiday season which has quickly fizzled. Depending on who you speak with, between 50% and 80% of the first-run M8 pre-orders have been cancelled or refused and this continues to happen even though Leica has announced "the fix". The biggest reason given to the dealers is not the cost of filters or the fact they won't be available until February, but the overwhelming belief that lack of acceptance of "the fix" will force Leica to roll out an updated "M8-mark 2" in very short order. It will be interesting to see how this plays out over the next six to twelve months.

[franka t.l.]

The issue on hand is not a thin or thick filter, but what type of filter. Leica had choosen to use an absorption type IR filter instead of tha sharp cut / hot mirror type. Fact is they could have made a hot mirror type just as thin ( although it will be pricier ). Still as of now, I simply do not see why Leica choose such a specification for the IR filter.

 

From an engineering point of view, the best option, if Leica is fear of the physical limit of a plane of optical matter ( in this case the IR ansortion wafer ) just in front of the sensor ( causing shift of focal plane etc ... ). the best solution is perhaps what Olympus and Sigma do. Place a filter infront of the shutter which is optically transparnt and flat and would not made a difference to the image quality as Leica feared. Olympus used that for their SSWF for busting dust which is probably a good idea in the M body, and Sigma simply use it for dust proofing the mirror chamber and palce the hot mirror there too.

 

But in any case, the M8 as it is and the way Leica specified the Solution, to me, is absolutely a failed product. The extra IR exposure is recorded on the analog level which is not consistent and not uniform across the frame. That amount to a camera that will naturally leak light and fog the media. Would we accept a new M body purchase that leak and fog the film. Well certainly not, why should we accept one in digital form !!

 

Whatever it might become, a digital M should be able to properly record image as photographic need dictated when mated with proper M optics. No string need attached , certainly not another IR filter IMHO !!

[Brad_]

If you'd like an accurate answer, one that you can rely on, seek out the advice from an

engineer who has a working background on this subject and in optics.

 

Otherwise, you're just getting "information" picked up here and there on the internet,

translated/interpretated along the way. Who knows if it's accurate or not.

 

OTOH, if yu're looking for words that can be relayed on to the next internet contact/

forum,

anything you pick up will probably be good enough for general chat and banter.

 

BTW, if you're looking for advice and information on neurosurgery, talk to me - I've been

reading about that subject a lot on the net lately...

[pico_digoliardi]

<i>My only concern would be that the IR filter would allow for maximum [sic] light transmission (no filter factor) and that the benefits of using the filters (where recommended) would outweigh the potential downside of a bit of extra flare.</i><p>

The hot-filters have no filter factor. <p>

Regarding a filter - I'd be interested in knowing how many Leica folks use a "protective" filter. A lot, I'll bet.<p>

<i>The issue on hand is not a thin or thick filter, but what type of filter. Leica had chosen [sic] to use an absorption type IR filter instead of Thai [sic] sharp cut / hot mirror type. </i><p>

Is that correct? And would a hot-filter on the sensor not have issues with the extreme angle of incidence with wide lenses?<p>

I am not an optics engineer. Is anyone here?<p>

Brad - good points, and let us add that one who is an autodidact had best have a very good instructor. :)

[vinay_patel]

"If you'd like an accurate answer, one that you can rely on, seek out the advice from an engineer who has a working background on this subject and in optics."

 

Or do what the average buyer (one of the thousands Leica still needs to convince in order to make a profit on the M8)will do: look around at all the other options in digital cameras aimed at pro and serious-amateur photographers, see how many of them need added filters, and decide if you even care about an explanation of why the M8 needs filters.

[Brad_]

I think that's an excellent approach, Vinay.

[blakley]

Bravo, Brad - exactly the right answer.

<p>

<i>

look around at all the other options in digital cameras aimed at pro and serious-amateur photographers, see how many of them need added filters, and decide if you even care about an explanation of why the M8 needs filters.</i>

<p>

Phrasing this in a way more friendly to Leica, look around at all the other options in digital cameras aimed at pro and serious-amateur photographers, see how many of them need asymmetrical lenses which introduce distortion and chromatic aberration and lose sharpness because of the action of a swinging mirror, and decide if you even care why the M8 needs filters.

<p>

Every design requires tradeoffs. Put a symmetrical lens really, really close to the sensor and you end up with optical issues. Put an asymmetrical lens far away from the sensor and you end up with optical issues. You pays your money and you takes your choice.

[blakley]

(I was referring to your earlier post, Brad, not the one immediately above mine)

[Brad_]

Bob, I think the phrase is:<P>

 

You pays your money and you takes your <s>choice</s> <I>chances</i>.

[vinay_patel]

"Phrasing this in a way more friendly to Leica"

 

Their marketing department has been trying that for a while now with no effect on cancellations :-(

[mark_tupper]

To Quote Jimi Hendrix - Hey Joe Live - You're Livin in the Past !

[nels]

A new Jimi Hendrix remix of "Magenta Haze" is due out for christmas.

 

Anyway you slice it, this camera is headed for the museums.

[geoff_portas]

To answer the original question, the light rays entering a filter on the front of the lens are almost parallel and will pass through virtually unchanged, whereas the rays at the sensor are far from parallel. Hence it is more difficult to avoid errors particularly at the edge of the frame with short focal length lenses.

 

Geoff

[pico_digoliardi]

<i>Or do what the average buyer (one of the thousands Leica still needs to convince in order to make a profit on the M8)will do: look around at all the other options in digital cameras aimed at pro and serious-amateur photographers, see how many of them need added filters, and decide if you even care about an explanation of why the M8 needs filters.</i><p>

Yah, sure. (In a Fargo accent).<p>

Trust the uninformed masses who have not the discriminating eye and be one with them. Go with the flow. Think not. Consensus as Truth. It's all a Merrily, Merrily Life is but a Dream downstream trip.<p>

Next? Alien Visitors helped Nikon and Canon.

[harvey_edelstein1]

Since it seems we have some optics experts here, can I pick your brains for an answer to the following question? How much exposure compensation will be needed to counter the light absorbed in overall light spectrum and in the IR spectrum. Would hot filter alter this and how.

 

One more, if you want to use uv or other filters is there going to be an interaction with any type of other filter (polariser etc.) that you think will cause unforseen complication from the Leica fix.

[blakley]

Not sure what the first half of your first question means, Harvey, but the Hot Mirror filters (incl. B+W 486) have a filter factor of zero - that is, they require no exposure compensation in the visible spectrum.

[vinay_patel]

"Trust the uninformed masses who have not the discriminating eye"

 

Discriminating enough to know when there's wool pulled over it, or so the sales figures would seem to say :-)

[pico_digoliardi]

<i>Discriminating enough to know when there's wool pulled over it, or so the sales figures would seem to say :-)</i><p>

Sales figures have as little to do with facts as popular concensus opinion. I suppose if one is a collector concerned mainly with prestige-as-value he might be disappointed by concensus, but experts are not so concerned.<p>

Imagine if concensus defined science, and there you have it: a flat earth.

[vinay_patel]

"Sales figures have as little to do with facts as popular concensus opinion."

 

I must be misunderstanding your point because I'm certain you didn't mean to say that sales figures are not facts. Manufacturers are keen to get accurate sales figures on every item. That's basic to calculate their profits or losses, to decide when and if to discount or rebate an item, and when to stop production so as to not have a ridiculous surplus of inventory when a successor model is rolled out. You couldn't possibly be trying to deny all of those well-known precepts, so you must have meant something else. Perhaps you meant that the sales figures the public is told are often not factual, and with that I would agree, often times they are exaggerated to boost consumer interest and/or stockholder confidence. However, I know you would agree as would any intellgent person, that dealers would have no possible reason or benefit for UNDER-stating sales figures. So when they are citing 50%-80% order cancellations and report many customers have expressed they will not be interested in the M8 as long as they need to use IR filters, don't we thinking people really have only 2 choices: believe them or believe the situation is actually even worse than they're saying?

[john_shriver]

The dichroic filter in the "hot filter" works on an interference basis. The thickness of the layers is related to the wavelengths of light to be passed or reflected. When the light arrives at a 45 degree angle, the effective thickness of those layers is increased by 1.414, and the characteristics go way off. Note also that some lenses probably hit the sensor at steeper than 45 degrees in the corners...

 

So, using a reflective filter in the sensor would have worked much better in the center than in the edges. The failure mode at the corners might well be worse than just reflecting too little IR, it could well have reflected too much red, which would be even worse. Also, reflected visible light is the last thing you want bouncing around the shutter box!

 

So the sensor appears to use an absorbtive filter. See page 28 of the full specification of the KAF-10500:

 

http://www.kodak.com/ezpres/business/ccd/global/plugins/acrobat/en/datasheet/fullframe/KAF-10500LongSpec.pdf

 

Note that it's not very steep in it's cutoff.

 

So, a hot filter works much better in front of the lens, since it is a less problematic environment. Yes, I'm sure that the performance does go off a little on extreme wide angle lenses, but not as badly as it would at the sensor.

 

Also, the hot filter on the lens reflects it's light out where it's harmless, rather than in the shutter box. Remember the problems many digital cameras have had with reflections between the sensor and the lens.