No edit summary |
|||
| Line 49: | Line 49: | ||
* [https://pjreddie.com/darknet/yolo/ YOLO: Real-Time Object Detection] <span style="color:#008000>(open source)</span> | * [https://pjreddie.com/darknet/yolo/ YOLO: Real-Time Object Detection] <span style="color:#008000>(open source)</span> | ||
* [https://github.com/deepfakes/faceswap Deepfakes Faceswap] <span style="color:#008000>(open source)</span> | * [https://github.com/deepfakes/faceswap Deepfakes Faceswap] <span style="color:#008000>(open source)</span> | ||
* [https://github.com/NVlabs/few-shot-vid2vid NVlabs/few-shot-vid2vid] <span style="color:#008000>(open source)</span> | * [https://github.com/NVlabs/few-shot-vid2vid NVlabs/few-shot-vid2vid] <span style="color:#008000>(open source)</span> | ||
Revision as of 23:37, 30 October 2021
Werk/Fachmodul
Lecturer: Max Neupert
Credits: 6 ECTS, 4 SWS
Date: Monday, 9:15 until 12:30h
Venue: Marienstraße 7b, Room 204
First meeting: Oct. 18
Moodle: 36605
Description
Micro-editing is a technique of rearranging tiny fragments of media to form a new work. In the context of music, microhouse is a subgenre of house which employs this technique. Akufen's Deck the House from 2002 may serve as an example. In the context of experimental film, Martin Arnold compiled his 1989 montage “Pièce Touchée” entirely from found-footage by copying frames in a specific order with an optical printer, emphasizing and amplifying gestures from the original movie. Steina Vasulka, Granular Synthesis and many other artists followed in exploring an aesthetic of deconstruction and reassembly of the timeline in moving images. In pop culture this “audiovisual cut-up” was used to expand the visual language of music clips and to have the audiences of live performances spellbound. Micro-edits are used in different contexts ranging from media art, experimental film-making to music clips and advertising. Digital video has become an almost infinite source of to-be-found-footage which is accessible to anyone, anytime through platforms like YouTube, which are essentially databases for moving images of almost any kind. They enabled pop culture phenomenons like supercuts: compilations of short shots of the same action, or YouTube Poop mashups of videos with a comical and at times immature humour: Today, meta information, close captions, machine learning analysis and music information retrieval can provide the means to generate automated edits. Real-time reassembly of media fragments based on databases, feature extraction or meta-information has become entirely feasible. In the class Computer's Cut — Generative Video Editing we will learn to let algorithms cut and edit.
Syllabus
Assignments
Final work
Files
- Cuttlefish Python scripts and tools
- Audiovideo Pure Data Workshop
Links
- Audiovision Linkliste
- Clint Enns: A Brief History of Algorithmic Editing
- http://thisunruly.com
- http://reorder.tv
- https://www.reclaimingremix.com
- A history of subversive remix video before YouTube: Thirty political video mashups made between World War II and 2005
- wikipedia:Edit decision list EDL
- wikipedia:Regular expression Online regex tester: RegEx101
- David Claerbout - The pure necessity https://www.instagram.com/p/B3nHj-MgEyj/ & https://davidclaerbout.com/The-pure-necessity-2016
Software
- Microsoft Visual Studio Code Editor (open source)
- VSCodium Visual Studio Code without the Microsoft telemetry (open source)
- VLC Media Player (helps you downloading YouTube videos: copy YouTube URL, open in VLC as network stream, then go to Tools -> Media Information, copy the Location and paste back into the Browser from where you can save it) (open source)
- downsub.com Web service to get the subtitles from a YouTube video (web-service)
- Aegisub Subtitle Editor (open source)
- Praxis Live (open source)
- Python (great video tutorial for beginners) (open source)
- PySceneDetect can autonatically detect scene changes in video and segment footage into clips (open source)
- Jupyter Notebook (open source)
- YOLO: Real-Time Object Detection (open source)
- Deepfakes Faceswap (open source)
- NVlabs/few-shot-vid2vid (open source)