Is it possible to correct the lens optics using software?
I am thinking of the following procedure using an HTPC:
Create 3 grids:
The Red grid (Red->255 and Green->0 and Blue->0)
The Green grid (Red->0 and Green->255 and Blue->0)
The Blue grid (Red->0 and Green->0 and Blue->255)
Each grid has lines with one pixel width.
For 1080p 192 vertical lines should be sufficient.
Send two times each grid the first time without anamorfic lens and the second one with anamorfic lens.
measure in the screen the displacement of the vertical lines with the use of the anamorfic len.
Then you make a plot displacement versus vertical pixel.
You make a fit and you have a calibration function .
Create a filter for a player the following filter:
for each frame and for each color (RGB)
all the vertical pixels at horizontal positions i (i=0,...,1919)
should be corrected using the inverse calibration function.
For test you send again the 3 grids with the anamorfic lens in order to see/measure the improvements
What do you think???
I am thinking of the following procedure using an HTPC:
Create 3 grids:
The Red grid (Red->255 and Green->0 and Blue->0)
The Green grid (Red->0 and Green->255 and Blue->0)
The Blue grid (Red->0 and Green->0 and Blue->255)
Each grid has lines with one pixel width.
For 1080p 192 vertical lines should be sufficient.
Send two times each grid the first time without anamorfic lens and the second one with anamorfic lens.
measure in the screen the displacement of the vertical lines with the use of the anamorfic len.
Then you make a plot displacement versus vertical pixel.
You make a fit and you have a calibration function .
Create a filter for a player the following filter:
for each frame and for each color (RGB)
all the vertical pixels at horizontal positions i (i=0,...,1919)
should be corrected using the inverse calibration function.
For test you send again the 3 grids with the anamorfic lens in order to see/measure the improvements
What do you think???
knc said:
This is one topic that has come up many times in the past.
The problem is CA is not like a panel misconvergence where the red green and blue are displaced arcoss the screen. CA is virtually non-existant in the centre, but gets progressively worse at the ends. The other point is that CA is symmetrical with correctly aligned prisms, so shifting Red to the left may fix CA on the right, but will in fact make it look worse on the left. The only way would be to have separate left and right sides. This then assumes the centre is always perfect...
Then of course, what happens when you remove the lens? You have massive convergence issues...
Mark
Mark thank you for your answer.
The correction is function of the horizontal pixel number. (1920 different numbers for each color {Red, Green Blue}
Of course , correction should be a smooth function of the horizontal pixel position
When we move the lens we deactivate the filter.
PS
a) I have a Phd in High energy Physics (CERN)
b) It needs a lot of work and it sound complicate but i am optimistic about the result...
The correction is function of the horizontal pixel number. (1920 different numbers for each color {Red, Green Blue}
Of course , correction should be a smooth function of the horizontal pixel position
When we move the lens we deactivate the filter.
PS
a) I have a Phd in High energy Physics (CERN)
b) It needs a lot of work and it sound complicate but i am optimistic about the result...
primary lens compatibility
I realize most people here are using commercial PJs, and those that have considered the 2-prism route with DIY PJs have been dissuaded due to the massive size of the aperture and short throw, but I STILL have a question!
Anyway, I understand the best way to guage compatibility with a PJ is with the paper test, but I have not yet purchased all my parts to do this, and the decision of which parts to buy hinges on the ability to do CIH with the longer throw lens.
I have two lenses: an 18"/5" lens (focal length/diameter) and a 22"/3.75" lens... the smaller aperture lens in my setup would have a throw ratio of 1.78, so it looks good from that perspective. Is there any reason to believe the sheer size of the lens (3.75" diameter) would make this not work if I use the 6x8" prisms?
I realize most people here are using commercial PJs, and those that have considered the 2-prism route with DIY PJs have been dissuaded due to the massive size of the aperture and short throw, but I STILL have a question!
Anyway, I understand the best way to guage compatibility with a PJ is with the paper test, but I have not yet purchased all my parts to do this, and the decision of which parts to buy hinges on the ability to do CIH with the longer throw lens.
I have two lenses: an 18"/5" lens (focal length/diameter) and a 22"/3.75" lens... the smaller aperture lens in my setup would have a throw ratio of 1.78, so it looks good from that perspective. Is there any reason to believe the sheer size of the lens (3.75" diameter) would make this not work if I use the 6x8" prisms?
Have you considered a VC? This thread started off as making a VC lens, but soon became a HE lens thread. If you wanted a longer throw and were prepared to leave the lens in place all the time (I do this with my HE lens) you can actually get better result from a VC.
Your TR is still going to be calculted the same way (native 16:9 image width divided into the distance from the screen to the projector's lens), but because the width of the native 16:9 light beam is the final width of the scope image, your projector needs to be further back to do this.
Some of the benifits to going with a VC are -
Pixel structure is reduced due to optical compression.
Longer throws allowing the projector to be placed at the back of the room instead of above the seating area.
Less visible CA. This has always been a sore point for the DIY guy and HE ananmorphics. CA actually runs horizontal for a VC, but because the light is compressed not expanded, it is less visible.
Less anstigmatism. The fact that the face of the prisms is now flat across the lens means that focus will not change across the image. You may see some top and bottom however - a trade off I guess...
Mark
Your TR is still going to be calculted the same way (native 16:9 image width divided into the distance from the screen to the projector's lens), but because the width of the native 16:9 light beam is the final width of the scope image, your projector needs to be further back to do this.
Some of the benifits to going with a VC are -
Pixel structure is reduced due to optical compression.
Longer throws allowing the projector to be placed at the back of the room instead of above the seating area.
Less visible CA. This has always been a sore point for the DIY guy and HE ananmorphics. CA actually runs horizontal for a VC, but because the light is compressed not expanded, it is less visible.
Less anstigmatism. The fact that the face of the prisms is now flat across the lens means that focus will not change across the image. You may see some top and bottom however - a trade off I guess...
Mark
You could, but you might also try simply changing the angle of the tilt of the lens itself. What I found was that the amount of compression could be changed, but I did not bother to measure the exact limits.
The same goes for HE lenses where by rotating the lens, the stretch can be expanded or reduced to get the exact 33% required...
Mark
The same goes for HE lenses where by rotating the lens, the stretch can be expanded or reduced to get the exact 33% required...
Mark
"you might also try simply changing the angle of the tilt of the lens itself."
I asked at AVS Forum and a Panamorph owner says it effects the geometric distortion but not the compression.
To be clear, by "You could" do you mean I could try it, or that understand the optics and know for a fact that it would change the compression ratio?
I asked at AVS Forum and a Panamorph owner says it effects the geometric distortion but not the compression.
To be clear, by "You could" do you mean I could try it, or that understand the optics and know for a fact that it would change the compression ratio?
Hello everyone, this is a great thread on Anamorphic Lens! 
Since I am a DIY person, I am interested in making a DIY Anamorphic lens. Looks like I found the right place to learn.
I found a supplier of optical crystal wedges, and was woundering if they would work.
website:
http://www.precisioncrystal.com/catalog/products/optical-crystal-wedge.html[/URL]
Let me know if they will, price is very reasonable.
Thanks
Mike
Since I am a DIY person, I am interested in making a DIY Anamorphic lens. Looks like I found the right place to learn.
I found a supplier of optical crystal wedges, and was woundering if they would work.
website:
http://www.precisioncrystal.com/catalog/products/optical-crystal-wedge.html[/URL]
Let me know if they will, price is very reasonable.
Thanks
Mike
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.