zpelgrims / lentil Goto Github PK

Could be confusing bfl/front focal length..

This would remove the C drawing library, use c++ one instead

Something to avoid certain flickering or something..

SPHERICAL

• Baush & Lomb 3 element lens (1948): Three element objective lens

• giant lens index relating lenses to patents

Zeiss Protar:

• The photographic lens leapt forward in 1890 with the Zeiss Protar (Germany). Paul Rudolph's Protar was the first successful anastigmat (highly corrected [for the era] for all aberrations, including properly for astigmatism) lens. It was scalable from f/4.5 portrait to f/18 super wide angle. The Protar was originally called the Anastigmat, but that descriptive term quickly became generic and the lens was given a fanciful name in 1900.
• The Protar is considered the first "modern" lens, because it had an asymmetric formula allowed by the new design freedom opened up by newly available barium oxide, crown optical glasses. These glasses were invented by Ernst Abbe, a physicist, and Otto Schott, a chemist, (both Germany) in 1884, working for Carl Zeiss' Jena Glass Works. Schott glasses have higher refractive index than soda-lime crown glass without higher dispersion. The Protar's front achromat used older glass, but the rear achromat used high index glass. Virtually all good quality photographic lenses since circa 1930 are anastigmat corrected. (The primary exceptions are deliberately "soft-focus" portrait lenses.)

ANAMORPHICS

• CARL ZEISS (1959): Anamorphotisches Aufnahme- und Projektionsobjektiv mit Einrichtung zur Fokussierungauf beliebigem Objekt- bzw. Bildabstand

• ARRI (1962): Anamorphotic lens system

There's several modern designs by Cooke, they are all pretty similar:

• https://worldwide.espacenet.com/publicationDetails/originalDocument?CC=US&NR=2015212301A1&KC=A1&FT=D&ND=3&date=20150730&DB=&locale=en_EP

• https://worldwide.espacenet.com/publicationDetails/originalDocument?CC=US&NR=2016103302A1&KC=A1&FT=D&ND=3&date=20160414&DB=&locale=en_EP

• https://worldwide.espacenet.com/publicationDetails/originalDocument?CC=US&NR=2016170186A1&KC=A1&FT=D&ND=3&date=20160616&DB=&locale=en_EP

• unknown but looks good (2014): Anamorphotic objective

• Kodak (1960): Anamorphosing lens systems for use in convergent light

• ISCO (1961): Anamorphotic attachment

• ISCO (1965): Four component anamorphotic attachment lens

• ISCO (1970): Anamorphotic objective

• ISCO (1970): ANAMORPHOTIC ATTACHMENT

• Schneider (2001): Anastigmatic anamorphic lens

• Schneider (2013): Anamorphic lens attachment for use in anamorphic lens for cinemascope-cinema technology for recording and projection of wide screen movies, has end lens axially displaceable relative to front lens and to cemented elements

• H. Chretien (1955): Objectif anamorphoseur pour prises de vues cinématographiques

• Bausch & Lomb (1960): Anamorphosing lens system

VARIOUS

• HCDNA patent: https://worldwide.espacenet.com/publicationDetails/originalDocument?CC=GB&NR=2532407A&KC=A&FT=D&ND=3&date=20160525&DB=&locale=en_EP
• Variable diopter patent, was cited by HCDNA (1956): https://worldwide.espacenet.com/publicationDetails/originalDocument?CC=GB&NR=748910A&KC=A&FT=D&ND=4&date=19560516&DB=&locale=en_EP

This will interfere with anamorphic setups, but was designed this way to support the 360degree field of view for fisheye lenses (sphere/sphere parameterisation).

Would be crucial to layout departments.

E.g exit pupil of bin/view 16 100

This would allow me to write to all buckets multiple times, replacing what I'm doing with tinyEXR.

Speak to Declan

I also need to think about what happens to all of the other AOVs. They need to reflect the bokeh changes as well.

libc++abi.dylib: terminating with uncaught exception of type nlohmann::detail::type_error: [json.exception.type_error.302] type must be number, but is discarded
Abort trap: 6

Looks like something is wrong with the .json database, figure out the diff between working lenses (1, 4) and the others.

Something must be null/discarded, but queried as a float/int value

Bug: should start it as "5":{} instead of "5":[]

Can pretty easily do this from within bash, spawning multiple processes.

This way I can get rid of lens.anamorphic and lens.aspheric booleans

Both for unidirectional and bidirectional

Fully Homomorphic Encryption seems to be a way. But it is slow. https://github.com/shaih/HElib

Either adapt ZOIC, or use the raytracing code from lentil.. Probably best to use lentil.

Need to create a user database which holds the lens purchases.

Then based on that database, we can compile custom builds to enable/disable specific lenses. The easiest way would be to pass the necessary lensIDs to the compiler and use ifdef's.

This database will also drive the website content, showing the purchases, etc etc

2D vector drawing is doing my head in, too hard to debug non-spherical elements

Is this even legal? Would be cool to know the amount of daily users, etc

id's: 3, 4, 8, 11

Rulers, grid, use millimeters

E.g. initial raypos/raydir:

sideview:
[-40, 0, 0]
[0, 0, -99999]

topview:
[0, -40, 0]
[0, 0, -99999]

Would only make sense if they are flipped. This is crucial to get right for the anamorphic lenses. Not sure if I am wrong though.

They don't change when switching viewing axis, figure out

• Compare results
• Compare speed

"Adaptive sparse polynomial regression for camera lens simulation"
"NeuroLens: Data-Driven Camera Lens Simulation Using Neural Networks"

Or is it okay to be eyeballing this? It only affects the ray vignetting.