D7000 optical low pass filter

Vitaliy Kiselev

Vitaliy Kiselev

Carbonite Member
Look at this photos

sensor-2.jpg


sensor.jpg



It seems that D7000 is very good candidate for OLPF replacement.

If you insert strong OLPF you'll have no aliasing and moire in video.
Similar to AF100.http://kenrockwell.com/nikon/d7000/sensor-2-1200.jpg
 
More interesting by the day! :) I love the image. This would be an INCREDIBLE addition. If it were hacked....ooooohhhhhh baby! Watch out! :)
 
what is a low pass filter btw. i mean i know what it does, but is it a flat lens of some sort, just a ground glass or what?
 
Vitaliy Kiselev said:
If you insert strong OLPF you'll have no aliasing and moire in video.
Similar to AF100.
Click to expand...

The downside being that it would make the camera pretty much useless for stills. Interesting idea, though, if you don't need to take stills.

JoeJITSU said:
Vitaliy,
Does this mean that it will have great lowlight capabilities?
Click to expand...

Not at all... Low pass filters blur high frequency edges to eliminate aliasing and 'stair stepping.'
 
that's why I'm wondering what it is. simply blurring does reduce high frequencies but if you blur enough to fully remove the detail the filter will be too wide thus blurring the detail into neighboring pixels resulting in a loss of sharpness. an ideal filter turns an edge into half a sine period, while a blur creates more of an s shape.
 
mattsand said:
that's why I'm wondering what it is. simply blurring does reduce high frequencies but if you blur enough to fully remove the detail the filter will be too wide thus blurring the detail into neighboring pixels resulting in a loss of sharpness. an ideal filter turns an edge into half a sine period, while a blur creates more of an s shape.
Click to expand...

This is all sampling theory. Yes, there has to be a loss of sharpness. You need to blur detail across neighbouring pixels to eliminate aliasing. The Nyquist theorem states that aliasing can be avoided by sampling at twice the maximum frequency of the signal. In an image signal, high frequency components are edges and the sampling rate is essentially the distance between the pixels. In order to reduce the bandwidth of the image to match the pixels on the sensor, a low pass filter is used to slightly blur the image. Hard edges, which are high frequency, are attenuated and aliasing is removed.

Vitaliy Kiselev said:
For D7000 price I don't see any problem here.
Click to expand...

Not a problem, just something to be aware of.
 
yeah but a blur as opposed to a real low pass filter isn't an ideal filter, thus in order to cut off all high frequencies it needs to be quite wide, thus blurring pixels *in the downsampled image*, thus lowering the resolution. this isn't really necessary in theory with an ideal filter which only removes the "fake" detail or aliasing, not the real detail at lower than half nyquist. then when theory becomes engineering it's always a tradeoff so depending on how these filters work, i'm suspecting it's unlikely that you can find one that removes aliasing without reducing resolution. in fact i'd say it's likely that one that doesn't reduce resolution is too weak to remove much of the aliasing at all. i mean not even racking the image quite a bit out of focus solves the problem completely. but as i said, i know very little about the engineering side of this, maybe these filters use some amazing technology that's much closer to an ideal filter than i can imagine?
 
A blur is a real low pass filter but yes, that's right, there's no such thing as a practical ideal filter, it's impossible to cut off sharply at the exact frequency.

I think that the best way to get the best resolution for a given pixel size is to oversample and then use a good resampling algorithm which doesn't cause its own aliasing.
 
You must log in or register to reply here.
Back
Top