calvinsilly (@ALVY_aaa) header
calvinsilly (@ALVY_aaa) profile picture

calvinsilly

@ALVY_aaa

Aslan

 Sweden
github.com/calvinsilly
Joined January 2015
  • Tweets1.9K
  • Following263
  • Followers502
312Media
Photo by calvinsilly (@ALVY_aaa)Photo by calvinsilly (@ALVY_aaa)Photo by calvinsilly (@ALVY_aaa)Photo by calvinsilly (@ALVY_aaa)Photo by calvinsilly (@ALVY_aaa)Photo by calvinsilly (@ALVY_aaa)Photo by calvinsilly (@ALVY_aaa)Photo by calvinsilly (@ALVY_aaa)Photo by calvinsilly (@ALVY_aaa)
About 3 months ago i uploaded an old project i worked on since 2022: youtu.be/ZA_ICDs9W9U?si=Fv8A… ive since then talked to a fellow editor about the project, and i felt inspired to share the rest here :3 A thread on some parts of the process:
Image from calvinsilly (@ALVY_aaa)
Image from calvinsilly (@ALVY_aaa)
5
3
38
2,233
more replies
Breakdowns of the main two scenes:
1
5
401
These were inreality proof of concepts that were meant to motivate me, but as i describe in the desc. of my video above i natrually lost intrest in the project; thinking after 4 years of (on and off) work that it was a waste of time. So in sharing this im hoping for closure :3
4
338
Reuleaux! A color model for film characterization github.com/hotgluebanjo/reul…
RGB to Reuleaux

ALT RGB to Reuleaux

One of the tools at play, using the new color model

ALT One of the tools at play, using the new color model

calvinsilly (@ALVY_aaa) profile picturecalvinsilly@ALVY_aaa8 Jul 2023
Replying to @ALVY_aaa
For the curious: Repo is currently being overhauled, one of which being the temporary name. In the meantime there’s a fork of mine up.
5
3
32
4,831
Oops! ALT text for the first image should instead be **Reuleaux to RGB** and not the other way around!
1
1,133
A neat overview of the project: github.com/hotgluebanjo/reul… And for people interested in the derivation: github.com/hotgluebanjo/reul…
A section of the overview article

ALT A section of the overview article

A section of the derivation

ALT A section of the derivation

1
1
4
1,364
Working with *hotgluebanjo* was delightful, a constant feedback loop of great things!
942
Replying to @ALVY_aaa
Really Nice work, I use a version of this on my tools some time ago. github.com/JuanPabloZambrano… Mine has an extra mode to normalize the saturation mode as well which is how I have been using it.
1
1
2
967
For the curious: Repo is currently being overhauled, one of which being the temporary name. In the meantime there’s a fork of mine up.
1
3
5,571
calvinsilly retweeted
Frame Found (@framefound) profile picture
Frame Found@framefound27 Apr 2023
Ghost in the Shell (1995)
Ghost in the Shell (1995)

ALT Ghost in the Shell (1995)

7
10
2,345
calvinsilly retweeted
Frame Found (@framefound) profile picture
Frame Found@framefound25 Mar 2023
Keanu Reeves and Laurence Fishburne in The Matrix (1999)
Keanu Reeves and Laurence Fishburne in The Matrix (1999)

ALT Keanu Reeves and Laurence Fishburne in The Matrix (1999)

7
29
2,942
Added invertibility. (2 year old problem lol)
calvinsilly (@ALVY_aaa) profile picturecalvinsilly@ALVY_aaa18 Feb 2021
Added "TetraAutomater" github.com/calvinsilly/Tetra…

GitHub - calvinsilly/Tetrahedral-Interpolation: Color transformations using tetrahedral math.

Color transformations using tetrahedral math. Contribute to calvinsilly/Tetrahedral-Interpolation development by creating an account on GitHub.

github.com
3
1,448
calvinsilly retweeted
Frame Found (@framefound) profile picture
Frame Found@framefound24 Jul 2022
Joseph Gordon-Levitt in Brick (2005)
Joseph Gordon-Levitt in Brick (2005)

ALT Joseph Gordon-Levitt in Brick (2005)

1
6
One cheap way of easily creating somewhat more realistic 2D camera shake is to fake parallax with a depth map.
5
80
more replies
Small details that may deviate from preference is to e.g. focus onto something else than the sky; or whether the z position for the camera is moving too. Adding on-top the same shake but *inverted* using a 2D transform can be helpful when the camera is focused on e.g. the sky.
3
1
1
12
One step further might be to for example in some attempt reconstruct unwanted visual *artifacts*. arxiv.org/abs/1909.09349
Deep 3D-Zoom Net: Unsupervised Learning of Photo-Realistic 3D-Zoom

Deep 3D-Zoom Net: Unsupervised Learning of Photo-Realistic 3D-Zoom

The 3D-zoom operation is the positive translation of the camera in the Z-axis, perpendicular to the image plane. In contrast, the optical zoom changes the focal length and the digital zoom is used...

arxiv.org
5
Load more