Home › About

About Fonti Studio

Fonti Studio is an AI-native subtitle and dubbing service for independent film distributors, sales agents, festivals, and broadcasters. We are a small team of post-production engineers based in Belgium, delivering worldwide.

The pipeline is automated; the accountability is not. Every delivery is reviewed by a named senior editor before the file leaves the building. Their initials go in the SRT header. We do not run a marketplace, and we do not hand files to anonymous freelancers.

What we hold to

• Broadcast spec, first pass. Output meets the Netflix Branded Delivery subtitle spec and the BBC Subtitle Guidelines. SRT is the default deliverable; TTML, IMSC 1.1, EBU-TT-D (Tech 3350) and CEA-608/708 conversions are available on request.
• Original cast only. For dubs we clone the original cast from the film you already licensed, never local dub actors, under cast-side consent, with an EU AI Act Article 50 disclosure shipped in every deliverable.
• Faithful translation. Register, idiom, and a locked glossary, verified by a cross-model review pass before delivery.
• Chain of custody. Encrypted at rest, EU-hosted, purged on written request. Your master never trains a model.

Reference delivery: The Last Kumite (2024)

Our reference run is a 101-minute English-language feature delivered in French and Dutch subtitles plus a French dub:

• 0 errors against the Netflix Branded Delivery subtitle spec, verified by automated conformance check before the file left the building.
• 1,212 cues per language, every cue within the ≤20 CPS reading-speed ceiling, ≤42 characters/line, ≤2 lines, ≥2-frame gaps, zero shot-change overlaps.
• 30 glossary entries (cast, place, and term names) locked and bit-identical across both languages.
• 0.68 speaker similarity (ECAPA-TDNN embedding cosine) between the French dub and the original actor, versus 0.55 for a generic preset voice, measured per delivery: the original cast cloned from the licensed master, with per-cue face-plus-voice speaker confirmation and emotion held across whispered, shouted, and broken delivery. On this scale even two recordings of the same speaker sit well below 1.0; the gap to preset voices is the point.

That is the floor every Fonti Studio delivery ships against, not the ceiling.