Plays full length & native resolution Rickroll with subtitles. This uses 170 costumes that plays for 3 minutes 30 seconds. Recorded at 8 frames a second (FPS) using 1698 frames with resolution 480x360, compressed to 33Mb with 3Mb audio. Press space to toggle subtitles Phone version: https://scratch.mit.edu/projects/941463385 17/12/2023: Added subtitles
170 costumes containing 10 frames each. Each frame only contains only updated pixels from the previous frame, using 255 colors + transparent.(8bit+Alpha). RickRoll is the worst to compress, as the camera pans and uses fade cuts very often. When either occurs it requires to redraw a whole keyframe rather than smaller changes from previous frame. Similar method as previous christmas video https://scratch.mit.edu/projects/938048623 It uses ffmpeg and ImageMagick, but this uses webp instead of png, webp is then embedded in the SVG costumes for a 40% saving in file size. Wont play more than 30% on phones due to phones limited memory. Good discussion about this from a 2015 rickroll project that uses 500Mb of images and needs 32Gb Ram https://scratch.mit.edu/discuss/topic/522770 Workflow below just needs couple of free tools ffmpeg and ImageMagick: 1. Images generated from mp4 using ffmpeg eg: ffmpeg -i 480.mp4 -r 8 %d.png 2. Generated partial images using ImageMagick eg: magick 1.png 2.png -fuzz 2% -layers OptimizeTransparency -delete 0 .\diffgif\2.png 3. Merge 10 frames into 1 costume eg: magick montage -mode concatenate -tile 1x %d.png[1-10] -background none .\montage\1.webp 4. Embeded webp in svg with powershell using script below Get-ChildItem .\webp\*.webp | ForEach-Object {'<svg xmlns="http://www.w3.org/2000/svg" width="480" height="3600" viewBox="0 0 480 3600" ><image href="data:image/png;base64,'+[convert]::ToBase64String((Get-Content -path ($_.Name) -Encoding byte)) +'"/></svg>' | out-file -encoding ASCII (".\svg\{0}.svg" -f $_.BaseName)} Commands used in Google Sheets file https://t.ly/mrydI Music Credits: Rick Astley - Never Gonna Give You Up (1987) https://scratch.mit.edu/discuss/youtube/dQw4w9WgXcQ Written by Mike Stock, Matt Aitken and Pete Waterman. Modified PTE based on @awesome-llama
