The Silver Sing-Along by Demostue Allst★rs
The Silver Sing-Along An occasional song for TRSAC 2026 - Silver Wedding edition A 21 kb intro for the 22nd year of the 23 TRSACs - 26 years after the first! by Demostue Allst*rs The team behind this productions is: Psycho: Code, scripting, shaders, lyrics Blueberry: Synth, scripting, lyrics Revival: Clipart shaders Slide: Music (Claude: Vocals) ftx: Choir conducting, portraits, lyrics It is a Danish custom at family events to bring and hand out a dedicated song to be sung by the guests. Usually using some easy, well-known melody, but we only had one to choose from. So clear your throats, and sing along with us! The compo version is without the voices, for the authentic experience. All the code for the choir tooling and runtime was generated by Claude Code without any manual changes. It was started at the Demostue meetup just one week before TRSAC, so a lot of things could be improved. The vocals are entirely generated by the "svc" (singing vocoder) Jingler external running in parallel with the synth playback and fed separate event streams consisting of phonemes for two choir sections. Each singer has slight randomization in onset, pitch, drift, formant shift and panning. A glottal oscillator drives an all-pole filter to simulate the mouth resonances. The choir is divided into two sections with 16 voices for the primary melody and 8 voices for the secondary. Each choir section is divided up into a number of soprano/alto/tenor/bass voices, and the melody lines are automatically mapped to the respective ranges. The primary melody has more deeper voices than the secondary. Lyrics editing is done via an HTML based tool where the background music is loaded and shown as a semitone analysis in the background. The melody is drawn on top of this and associated with the hyphenated lyrics in a Vocaloid/Melodyne like way. The lyrics editor can switch between languages via [$da], [$de], [$no], [$en]. In a window below it shows the translation into phonemes. It also lets the user enter phonemes directly in the natural language lyrics inside [...] blocks. Dashes in the lyrics are used to split words between notes, and vowels can be stretched over multiple notes by linking notes in the graphical note editor. The editor exports an .svc file for each choir section which the "svc" external hot-loads and sends to the VST3. The singing control data compresses to about 3.9 kb, while the player for this compresses to about 3.6 kb. ...at the end this leaves us with one question: What more is there to do when the robots now have stolen one of the last things that made us human and gather as a community to join in on a common song? Well, to that my friends there is only to say: Go to more demo parties, meet the people, sing some songs! And come to TRSAC for the one and only real DVD experience!
[ back to the prod ]