tag:blogger.com,1999:blog-19092890.post2917582206398628977..comments2024-03-19T08:15:45.946+02:00Comments on Image Sensors World: Depth from Chromatic AberrationsVladimir Koifmanhttp://www.blogger.com/profile/01800020176563544699noreply@blogger.comBlogger4125tag:blogger.com,1999:blog-19092890.post-44424555881736919942021-05-13T15:58:59.275+03:002021-05-13T15:58:59.275+03:00If I remember well, DxO did interesting work in th...If I remember well, DxO did interesting work in this field where they used the chromatic aberration to extend the depth of focus.Albert Theuwissen - Harvest Imagingnoreply@blogger.comtag:blogger.com,1999:blog-19092890.post-13970568773256292972021-05-13T15:58:10.697+03:002021-05-13T15:58:10.697+03:00Hi Dave, very interesting comments. Especially th...Hi Dave, very interesting comments. Especially the last one, thanks for sharing. Albert.Albert Theuwissen - Harvest Imagingnoreply@blogger.comtag:blogger.com,1999:blog-19092890.post-61811374486239857582021-05-13T00:59:25.570+03:002021-05-13T00:59:25.570+03:00It depends. The original 3-tube prism assembly use...It depends. The original 3-tube prism assembly used in broadcast color cameras was specifically designed to have a total offset of ten microns in the focal planes for the three channels. The lenses lenses were designed to match. The offset provided some benefits in assembly and channel alignment.<br /><br />Sensors with stacked photodiodes (Foveon) could also use offests in the focal plane position but the total offset would only be about 3 microns. <br /><br />The natural CA offsets in old lenses provided another interesting result. It made it possible to extract some color information from old thick monochrome film emulsions.Dave Gilblomhttps://www.blogger.com/profile/00335345152573150910noreply@blogger.comtag:blogger.com,1999:blog-19092890.post-82654402123533050952021-05-12T19:02:15.428+03:002021-05-12T19:02:15.428+03:00I remember seeing proposals for this sort of thing...I remember seeing proposals for this sort of thing in the mid 2000's, back when EDOF was cool.<br /><br />Most lenses have CA minimize. The residual error is usually a weird function, not the simple single-lens CA in the diagram here. <br /><br />On the other hand, deliberately designing a lens to have large CA for the purposes of depth perception leads to predictably bad RGB images. <br /><br />I'm sure this works fine in the lab and simulations, but runs into a lot of problems outside.Unknownhttps://www.blogger.com/profile/06164074033917462617noreply@blogger.com