It's been slightly more than month since I've modified my camera, and in that time, I've help three other persons modify their cameras, Nisa being the latest, the other two being colleagues from the US. The interesting thing is that each of our cameras were modified using different types of filters, so for those thinking about doing the mod, here is a simple finding:

Suitable filters:

Hoya R72 (Troublesome to cut but optically flawless. Hard and scratch resistant)

B+W 093 (Troublesome to cut but optically flawless. Hard and scratch resistant)

ELP (Very easy to cut, optically good. Hard but not very scratch resistant)


Unsuitable filters:

Cokin P007 (Very easy to cut, BUT optically FLAWED. Semi-hard and quite easily scratched)

Of the four different filters that have been used, the Cokin 007 is by far the worst. Unlike the ELP, which seemed more or less a uniform color when I cut it, the Cokin 007 seemed to be a clear piece of plastic, sandwiched between two layers of dark IRP material. What is worse is that even a brand new, unused piece that I used to cut for Nisa is flawed on both its surfaces. Unlike the ELP, which is again quite flat and uniformed on its surfaces, the 007 is dotted with tiny 'dimples'.

This flaw isn't just skin deep. When a 'dimple-less' piece was mounted infront of the sensor and shots were taken, it initially looked good. However, when the lens was stopped down below f/5.6, tiny dots began to appear. I initially thought that they were dust left on the filter or CCD, but several cleanings and 4 different pieces of cut filters later I realised that these dots were not dust but tiny air-bubbles trapped in the filter material. When mounted on the lens, these flaws become invisible, just like a speck of dust is invisible when on the lens but painfully visible when close to the sensor.

Here are two sample shots from two different cuts of a Cokin P007:

#1
(IMG:http://www.pbase.com/image/49403370/original.jpg)

#2
(IMG:http://www.pbase.com/image/49403371/original.jpg)

All the dots that you see in these photos are not dust bunnies on the filter or CCD, I made very sure of this by cleaning both and mounting the filters in a cleanroom. They do not exist on the surface of the filter and since they are located on different positions on different pieces of filters, that rules out deadpixels. The only logical conclusion that can be drawn is that they exist as air-bubbles in the filter.


Conclusion:
Stick to glass filters. They are harder, more resistant to scratches and optically superior to resin type filters. Their optical superiority far outweighs the inconvinence you'll face getting them cut. In fact, Life Pixels, a 'company' that does IR camera conversions, have also started switching to glass filters.

What about Gel types - haven't tried them as replacement filters but my gut feeling is that this should be avoided. Gel filters are even softer than resin ones and definitely not heat resistant. If placed in front of the CCD, they might very well warp or become deform.

Cheers,
Matt

errr what are u saying here?

Sorry

To get straight to the point, the Cokin 007 series of filters CANNOT be used for D70i conversions.

We had all assumed that Cokin, being such an established company, would have made an IR filter of good, uniform quality. However, the photos above clearly show that that isn't the case. Those dots you see in the two photos are not caused by dust/scratches on the filter. Remember that I told you I was going to use a completely new filter for you? Well, I did. There were no scratches on the filter and I cleaned it before mounting it in your camera. These dots are caused by air-bubbles inside the filter itself.

When I first completed the mod for you, I fired a couple of test shots at f/3.5 in the evening, so the dots did not show up. When I took some at f/8 in daylight however, they became very visisble.

So, it looks like I'll need to get you a Hoya R72 or B+W 092 or Edmund Long Pass (ELP) before I can complete the mod for you. Unless you can live with those dots

Matt

By the way Nisa, are you aware that the battery that came with the 2nd hand D70 you bought is not an original Nikon battery?

Matt,

thanks for posting this IR filter review, but why did you get thoose hot spots in this test pics???

Cheers,

Wolfgang

This post has been edited by NikkorAT on Sep 18 2005, 09:48 PM

Hi Wolfgang,

The hotspots? Hee! Hee! I was doing the test shots with a zoom lens instead of my AI Nikkor 28mm f/3.5.

You know how some of us believed that hotspots were caused by light bouncing between the sensor and rear element during long exposures? Well, that is actually not quite correct. I have experienced hotspots (when using wide angled zooms) even at shutter speeds of 1/1000 and they are just as big and bad as the hotspots created by long exposures of 1 sec or more.

The real reasons for hotspots is the diffraction of light as the light-rays hit the elements and the edge of the aperture. Simply put, the wider the lens and the more elements it has and the smaller the aperture we use, the worse the hotspot.

So it is like you discovered, primes with few elements make very good IR lenses.

Cheers,
Matt

Hi Matt,

I didn´t know that you have an Ai 28mm f3.5 nikkor
Is it the one I told befor, with the small rear element from the 4th series?
Just bought it...?

The only problem I still don´t understand is why every lens is so unsharp wide open. Why
it´s absolutly nessesary to stop the lens down to f16 or even worse to f22 to get sharp results???
Befor I tought it´s about the lens construction .... but then I realised that also with the 16mm fisheye I need to stop down to f16 o get perfect sharp pics.

Then I tought its kind of IR photography law, but then I found out that there are some lens which you can use wide open with perfect sharp results at f4.0. So know I don´t know anymore what to believe.

What is your experience about sharpness in IR photography with nikkor lens?
This is for sure, a questions to all of you here...

Cheers,

Wolfgang

PS.: This weekend I will be with you already

This post has been edited by NikkorAT on Sep 19 2005, 01:44 PM

thanks matt for testing it out. guess I can't use cokin anymore, gotta use other filters


Nisa, u wanna share a ELP?


Wolfgang, an early welcome to Singapore. your invitation for a chitchat session still on?

Hi Wolfgang,

No, the Ai 28mm f/3.5 nikkor is from a slightly earlier series and I have owned it for more than 10 years, since when I was shooting film.

Regarding sharpness, there are two things we usually get confused about:

1. Sharpness is how clear and how much small details we can see in a photo at the point where we focus.

2. Depth of Field (DOF) is how much of a photo is acceptably 'clear', in front of and behind of the point where we focus.

Sharpness is determined by the lens quality alone, so the better the glass, the sharper the details in the photo.

DOF is determined by the size of the aperture. The larger we open the aperture, the less is clear and the smaller we stop down the lens, the more things in the photo becomes clear, this is a law of nature (or physics). This is the reason why many photographers will stop down their lens to capture more details in the front and background for landscape photos. But this is no good for IR photo because the diffraction of light through the very small aperture will cause big hotspots.

There is a way to avoid this problem that is used by experienced landscape photographers and that is to use a lens' hyperfocal point. The hyperfocal point of a lens is a point you focus at to get maximum DOF, depending on the size of the aperture. I posted a short note on hyperfocusing in another forum:

http://forums.clubsnap.org/showthread.php?t=149057

To test this out, try this with your 28mm Nikkor:

1. Set the aperture to f/8
2. Rotate the focusing ring until the focus distance is lined up at around 4.9m (Remember to use the infrared mark on the 28mm lens instead of the normal mark. It should be a red dot or line next to the normal center white line). No need to use the autofocus.
3. Compose your shot and shoot.

If done correctly, you will notice that everything that is more than 4.9 meters in front of you will be clear. Compare this to landscape photo that you take by focusing at infinity.

If you would like to try with a 16mm fisheye, open the aperture to f/4 and set the focus distance to about 3.2m (again using the infrared mark instead of the normal mark)

The photo below is an example I always use for hyperfocal sample:

(IMG:http://www.pbase.com/deadpixel/image/45698620/large.jpg)

The 12mm lens was wide open at f/4 but set to its hyperfocal point of 1.8m, but you can see that despite the fact that it was focused at 1.8m and the aperture wide open, all of the background is still clear.

Do give this a try.

Cheers,
Matt

Hi quekky,

sure, same I said befor.

Cheers,

Wolfgang

This post has been edited by NikkorAT on Sep 19 2005, 07:15 PM

Hi Matt,

I understand what you mean, but what I´m talking about is something different.
For me DOF is normaly limited by the lens you are using.

For example:
If I use a fisheye I don´t have to think about unsharp details in the background, there aren´t any

But if I want a very small DOF I would use a tele lens.

The problem about the sharpness is much more difficult because it is limited by the qualitiy from the lens,
and as we know there are not many lens in IR producing sharp images with an R72 Filter.

The question is also what is an sharp picture in your mind?
For me it must be perfect sharp befor working in photoshop.
No unsharp masking inside the camera, no Photoshop work and if have to be sharp, for everyone understand what I mean.

And yes, this is possible, even wide open if you are using the right lens from nikon.

Cheers,

Wolfgang

This post has been edited by NikkorAT on Sep 19 2005, 07:34 PM

Hi Wolfgang,

Ok, sorry for the long explanation but when you focus with the manual AI/AIS prime lenses, do you compensate for the shorter IR focal length? With prime manual Nikkor lenses, I will normally do the following:

1. Compose and focus, making sure the image is sharp

2. Compensate for IR focusing by turning the focus ring to the left to where the red dot/line for IR focusing is

3. Put on the IR filter (when I'm using a non-modified camera) and take the shot

This compensation becomes more important when the focal lenght of the lens gets longer because DOF becomes narrower.

For me, if I want a subject to be sharp, a photo must be sharp from the camera without camera sharpening or photoshop sharpening. This 'artificial' sharpening is unacceptable.

Many, but not all, of Nikkor's f/2.8, f/1.8, f/1.4 and f/1.2 lenses are very sharp when wide open but they are not always good for IR. With older manual lenses, the f/3.5 and f/4 lenses are also quite good.

Cheers,
Matt