Simple tool to characterise blur

Tools
#1

I have in my head an idea for a simple tool to characterise blur. All it requires is to the take the astigmatism “sunrise” tool (the one with all the spokes radiating out) and put in small calibrated gaps. (Could also think of it as straightening out some Landolt C’s and using them to construct the sunrise.)

The idea is that blur causes the gaps to be filled in. Isotropic (myopic) blur should fill in all the gaps uniformly. Anisotropic (directional / astigmatic) blur would fill in the gaps only in on the lines which lie parallel direction of the blur.

(I actually conceived it to try to prevent ghost images from masking blur - same sort of principle - the ghosting happens in one direction, and so you can look for the blur happening on the line that’s perpendicular to that.)

Does this make any sense ?

2 Likes
Edge of Blur is an illusion
#2

what do you mean by characterise? I mean what metric are you trying to determine? Distance to blur?

#3

Characterise was as in determine whether it was astigmatic or not. But yes, I had in mind to use it for measuring cm, because I am finding ghost images complicate things, and I am hoping this might help with that.

#4

huh. maybe not for everyone, but for me it is simple to tell which is which just by looking at lines along my axis or 90 degrees to my axis. The blur starts at fewer cm and appears to one side of lines on the weaker axis

#5

I’ve put a quick hack of the sort of thing I had in mind at
https://wiki.endmyopia.org/wiki/User:Divenal/sandbox

While I have ghost images, haven’t quite decided if it’s astigmatism or not.

3 Likes
#6

Great idea Dave. I think I understand your logic and agree completely

However, in practice what I see isn’t what I first expected with your gapped radial line image. The angle I see most boldly, or clearly, at blur on other similar solid line images, is the same angle that I see the line fussing together on your image.

So the line I see as clear on other charts is actually the angle of most directional blur? Well I guess that makes sence then. I hadn’t really put much effort into understating the cylinder correction axis. Mine changes often so I dropped it and forgot it. Bad attitude I guess.

Thanks Dave for the nudge to think about the topic.

#7

Yes, I think that’s what I expect. The reason a line appears darker is that the blur/ghosting is happening along the line. ie the blur from one pixel coincides with the next pixel along the line, reinforcing the image. (You can’t distinguish the real pixels from the blurred ones.) Whereas in the orthogonal direction, the line appears wider and lighter because it is being extended out sideways from the line.

(I do see the darker line growing outwards at the end, when the blur doesn’t have anywhere to hide.)

#8

Everyday binocular vision for me, horizontal lines are clear, and vertical lines I see some ghosting. My previous cylinder axis was 90.

Checking the image, the darker line is about 30 degrees from horizontal. There aren’t many objects at the angle of my directional blur but I guess 30 degree is close enough. I’ll pray more attention to this.

#9

Outside my office window is a tree with twigs at all sorts of interesting angles. Makes my ghosting blindingly obvious when I look out wearing diffs.

#10

Either mine is much less severe or I’m not as perceptive as you are.

I can generally only identify directional blur as the cause of my blur on longer sections like poles or power lines. If AF is good I might see it on shorter things like branches.