IR CUT filter over sensor...

[atufte]

Advertisement (gone after registration)

Sorry if this topic is somewhat " old news" or irrelevant...

 

But i come over this today, and i thought, how would

the M8 act if a filter like this was mounted in front of

the sensor, will it fit at all..? i heard there is a slimmer

one on the marked, which (maybe) will fit..?

 

http://www.astrosurf.com/buil/d70/ircut.htm

 

I know there's are alot of "astroheads" doing this kind

of mod with various of cameras, here in Norway the rep

guys at Canon service, even provide this kind of mod

from their main workshop... (not even expensive..)

 

And of course, this can be done the other way around,

by fitting one instead of removing it, but it is probably

not enough space, or is it...?

 

I know that i for one would be more than happy to pay for

this kind of mod (reversed of course) if possible, since this

probably will be both cheaper and much more versatile in the

end, it may even be better for the wideangle cyan shift, etc...

 

Do you guys even think it's possible...?

 

I know that probably Leica would have done this if possible, but maybe

not, as they say:

 

"During the development of the LEICA M8, we made important design choices to insure that the camera delivers the quality in images the Leica M System is known for"

 

When they desided to skip the filter we may have gained some more detail, contrast, etc, but in the end, when we use filters on the lenses we are all back to the same results

anyway, so this may only intrest those of us willing to keep the filters on all the time and not in need of the extra IR sensitivity (astro, etc)

 

Best Regards

 

Alexander Tufte

http://www.alexandertufte.com

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[ed_oneal]

There is no room for a filter between the lens and the sensor because Leica wanted to keep the size of the camera the same as previous M bodies.

 

Leica says their choice of an IR sensitive sensor was deliberate and the only way to compensate for that is a filter on the front end. It seems to me that this is a lame excuse because of all the problems the IR sensitivity has caused.

[marknorton]

It all comes back to the fact that M lenses - especially wide-angles - sit closer to the sensor than they do in an SLR. The problem is not one of lack of space for a filter at all - a filter is only 1-2mm thick, reduced to 0.5mm in the M8 - it's the fact that light then exits the lens at a greater angle of incidence than in an SLR. Light striking the filter glass/sensor causes problems which have been fixed by using the offset microlenses and a thinner filter.

 

Unfortuately, the filter used is then not fully effective in removing IR. Other items on the astrosurf site include some impressive spectral response curves showing how well the filters used in Nikon and Canon SLRs work.

 

With the geometry of M lenses, the only solution seems to be to use an external filter, with the potential in the future for new filter glasses with higher IR attenuation and/or sensors with innate rejection of IR. Sadly, the quantum mechanics processes at work - turning photons into free electrons - work across a spectrum of visible and invisible light and don't know anything about what we can see and what we can't.

[Vivek Iyer]

Folks, Leica did not decide on how to construct the sensor (IR cut filter, micro lenses, etc).

 

It is a Kodak made sensor that is available for anyone (to purchase) with the same 0.5mm thick Kyocera filter and the microlens array.

 

The 0.5mm thick filter for M body, i don't know where that speculation started but nowadays it is typed as a fact.

 

That (filter/micro lenses) is a Kodak design. How Leica implemented it in the M8 (hardware and software) is fully Leica's and whatever performance (or lack thereof) from the M8 is also Leica's responsibility.

[rvaubel]

 

Unfortuately, the filter used is then not fully effective in removing IR. Other items on the astrosurf site include some impressive spectral response curves showing how well the filters used in Nikon and Canon SLRs work.

 

With the geometry of M lenses, the only solution seems to be to use an external filter, with the potential in the future for new filter glasses with higher IR attenuation ........

 

The astro community has had success with removing the factory filters and replacing them with custom designs. That approach only works for sensors that have filters that can be removed. If the filter isn't bonded maybe the Leica community could come up with something a little more aggressive. A filter that was both absorbtive and dichromic might be a better (and more expensive) choice than Leica made.

 

Rex

[rvaubel]

Folks, Leica did not decide on how to construct the sensor (IR cut filter, micro lenses, etc).

 

That (filter/micro lenses) is a Kodak design. How Leica implemented it in the M8 (hardware and software) is fully Leica's and whatever performance (or lack thereof) from the M8 is also Leica's responsibility.

 

If Kodak is responsible, why did they pick a filter that attenuates IR to such a small degree? I assumed that Leica's requirement because of the short backfocus of wide angle lenses, dictated a modest IR filter so that they did not get radial overcorrection.

 

In other words, why would Kodak design a IR filter that didn't cut until 780nm or so if they were intending to sell to a telecentric customer base?

 

Rex

[marknorton]

Advertisement (gone after registration)

I feel certain there was close collaboration with Kodak in the development of the sensor. What we don't know is whether the sensor actually used is the standard sensor or one with a custom coverglass. Whichever it is, it doesn't do a good enough job of filtering IR which is why Leica is in this mess. Doesn't reflect especially well on Kodak either...

[Riley]

there is no speculation over that Mark, you are quite right

 

KODAK Image Sensor Solutions - New Kodak Image Sensor Powers LEICA M8 Digital Camera

 

ROCHESTER, NY, September 15, 2006 - Eastman Kodak Company is extending its partnership with Leica Camera AG to provide image sensors for the newly announced LEICA M8. The new KODAK KAF-10500 image sensor was developed specifically for use with the M8, and marks Kodak’s second collaboration with Leica, building on the success of Kodak’s interaction with Leica on the LEICA DIGITAL-MODULE-R. The first digital camera to use the Leica viewfinder system, the LEICA M8 extends the Leica M series into the digital world. By working closely with Leica, Kodak designed and optimized the new image sensor to meet the demanding needs of Leica photographers.

 

end quote

 

I identified this relationship way back when all this began, that certain enthusiasts of IR were removing glass filters from infront of sensors to enhance IR effects. Clearly then the process can be worked the other way, but the issue of reflections from the outermost glass surface is most commonly suggested as the reason this wasnt attempted. Perhaps the right approach then, is an examination of reflected and refractive qualities of IR glass filters and to add, if the light rays were parrallel in the first place, this cause and effect would be much reduced.

 

Riley

[rvaubel]

I feel certain there was close collaboration with Kodak in the development of the sensor. What we don't know is whether the sensor actually used is the standard sensor or one with a custom coverglass. Whichever it is, it doesn't do a good enough job of filtering IR which is why Leica is in this mess. Doesn't reflect especially well on Kodak either...

 

Right, this clearly had to be a collaboration with Kodak because of the non teleocentric requirements of a rangefinder camera. Well at least they got the microlensing right. It really is pretty amazing how much the vignetting of wide angle lenses is becoming a non-issue.

 

I have a feeling that with what they have learned about the magenta issue that the M9 will have a more effecitive implimentation of the IR filter.

 

Rex

[jaapv]

When they desided to skip the filter we may have gained some more detail, contrast, etc, but in the end, when we use filters on the lenses we are all back to the same results

anyway

 

I think this is incorrect. There is a considerable difference between a filter in front of the lens and one in front of the sensor and that is the incidence angle of the light. IR filters are sensitive to that. With a IR filter in front of the sensor there would have been cyan vignetting to contend with for all focal lengths except Visoflex lenses. That vignetting would have varied with focussing distance as well, making it virtually impossible to correct in camera or as batch in post-processing.

[marknorton]

I have a feeling that with what they have learned about the magenta issue that the M9 will have a more effecitive implimentation of the IR filter.

 

It would be interesting to know ehat the view within Leica/Kodak about the IR issue is - whether this is the way it's going to be or whehter there are new sensor/filter technologies which will fix it.

[stunsworth]

I can understand why the IR filter can't be fitted to the sensor, but why not apply the relevant coating to one of the lens elements, either the front or one of the internal ones? That way they: a, solve the problem; b, encourage people to buy new lenses :-)

[Riley]

while thats entirely possible, that might interfere with using the lens on other cameras

and to use all new lenses would adversley affect the suitability to existing customer base

 

a design solution not sought, and now impossible to retrofit

would be to have used a fixed mirror, situated at 45 degrees within the lightpath

this would mean that the sensor would be fitted perpendicular to the usual film plane

 

the mirror would attempt to keep the lightpath parallel

and therefore oblique light fall off would be greatly reduced

and fitting an IR coated glass filter above the sensor would not present the same fears it does now

 

Riley

[ed_oneal]

It looks to me like Leica/Kodak failed to understand what problems would result from IR sensitivity. How many people really want IR sensitivity?

 

Ulitmately, the problem stems from trying to make a digital camera that would use most, if not all, existing Leica/Leitz lenses, especially very wide angles, and keep it in the same size body. A very difficult engineering problem indeed requiring many compromises.

[Riley]

to the contrary ed

in fact this has been well understood for some time

and is the reason for the predominace of dSLR design

they both have known all along that fitting a filter within was too bold an option, the associated refraction issues to boot

 

i have seen documentation from as early as 2002 that specifically states that leica can not build the full frame M because of the vignetting issue, and the IR issue

 

 

Riley

[scott kirkpatrick]

I think this is incorrect. There is a considerable difference between a filter in front of the lens and one in front of the sensor and that is the incidence angle of the light. IR filters are sensitive to that. With a IR filter in front of the sensor there would have been cyan vignetting to contend with for all focal lengths except Visoflex lenses. That vignetting would have varied with focussing distance as well, making it virtually impossible to correct in camera or as batch in post-processing.

 

I think you have got it backwards. In the simplest picture of red vignetting due to the angle at which light passes thru a dichroic filter, the angle at the edge of the frame can be greater when the filter is in front than when it is in back of the lens. That's because lenses may sit further away from the CCD than their focal length suggests, but never closer. I agree that focusing on a close object moves the lens away and decreases any vignetting, so some compromise is needed when the filter is at the sensor. But these are not enormous effects.

 

scott

[ed_oneal]

Rob, I'm not saying that Leica didn't realize that there is an IR problem. I'm saying that they didn't realize that it would be found so unacceptable to so many people, especially since wide angle is so much in fashion. Low IR response and wide angle may well be mutually exclusive in the near term.

 

Ultimately, the problem will probably worked out, but not soon.

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[rvaubel]

Rob, I'm not saying that Leica didn't realize that there is an IR problem. I'm saying that they didn't realize that it would be found so unacceptable to so many people, especially since wide angle is so much in fashion. Low IR response and wide angle may well be mutually exclusive in the near term.

 

Ultimately, the problem will probably worked out, but not soon.

 

The problem is solvable, just like the vignetting issue. But it must eventially be done at the sensor level. I'm live with filters for the M8 but the M9 should be fixable at a level that doesn't require them. I will give Leica design a break as the technical solution is difficult.

 

Meanwhile we have the M8 which is, in most way, living up to my expectations. The color issues are solvable thru some rather inelegant kludges, but solvable nonetheless.

 

Rex

[Jean LeBlanc]

It might be interesting to know if the front element of wide angle lenses could be changed with one with the filter effected added. That would presumably eliminate the cyan shift since light reaching the corners of the image would have passed through the portion of the element at the edge, and thereby not have passed through at quite so oblique an angle as it would through a flat filter.

 

My understanding is that these filters have no effect on ordinary film, so a lens that has been modified that way should present no problem for film work, other than infra-red film.

[Vivek Iyer]

What will happen if someone uses the untreated lenses? Or if Leica changes the current sensor (see: http://www.leica-camera-user.com/digital-forum/10658-sensor-choice.html)?