What Axeho does

Nine modules, in the order you actually work an episode: from transcribing the original version to a mix-ready file. All in one application, on your own machine.

An episode, end to end

1. Get the text

   Import an existing subtitle, a dialogue list, a spreadsheet, or have the original version transcribed straight from the video, in any of the 26 recognised languages.

2. Detect

   The application marks labials, semi-labials, rounded and open vowels in the source text. This runs locally: no account, no key, no connection.

3. Adapt

   You write the target version under mouth constraint, with reading rate watched throughout and viseme-filtered suggestions whenever you ask for them.

4. Align

   Multitrack rythmo band, one track per character, aligned to timecode over the waveform, broken into recording loops.

5. Record or generate

   Print role sheets for the floor, or generate voices per character with a timbre kept consistent from one episode to the next.

6. Deliver

   DetX, SRT, VTT, TXT and spreadsheet exports, separate stems, levelling and mastering to several broadcast targets.

The services

Each service has its own page, audience and answers.

Script : the text and its detection

• Offline phonetic detection over a 77,288-entry lexicon shipped inside the application: labials (B, P, M), semi-labials (F, V), rounded and open vowels. No network access, so it works in the booth, on the road, or on a programme that is not allowed to leave the studio.
• Three ways to trigger detection: on demand for one line, as automatic underlining while you type, or by double-clicking a word.
• Mouth constraint: suggested synonyms and turns of phrase are filtered by viseme: a suggestion that is correct but does not hold against the picture is never shown to you.
• Reading-rate control: characters per second flagged both ways. A line that is too tight is a problem; a line that is too loose is another one, and it usually goes unnoticed.
• Checking: spelling, grammar, spoken register, calques from the source language, overused conditionals, subject-verb inversions.

Rythmo : the band

• Multitrack, one track per character, with the waveform under the text.
• Frame-accurate timecode alignment, adjustable frame rate and reference point.
• Loop breakdown for the floor, roles and stage directions.
• Synchronised video playback, including from a local file with no network at all.

Series : the bible

A series is not handled one episode at a time. Axeho keeps character names, forms of address, recurring phrases, adaptation choices and voice casting from one episode to the next. Episode twenty therefore speaks like episode one.

Conform : the cut has changed

The most expensive situation in dubbing: the locked version arrives after the adaptation is written. This module carries the existing text onto the new cut by timecode alignment, and flags what moved, what disappeared and what is new, instead of starting again.

Transcription : 26 source languages

The original version can be transcribed straight from the video, with automatic language detection or an explicit choice. African languages are treated on the same footing as the rest, which is rarely the case elsewhere:

• Wolof
• Yoruba
• Lingala
• Igbo
• Hausa
• Swahili
• Xhosa
• Zulu
• Amharic
• Shona
• Somali
• Northern Sotho
• Ganda
• French
• English
• Spanish
• Portuguese
• German
• Italian
• Dutch
• Arabic
• Mandarin Chinese
• Cantonese
• Hindi
• Korean
• Japanese

Voices : casting and generation

• Casting per character, kept from one episode to the next: the timbre does not change mid-season.
• Line-by-line generation, with targeted retakes: you redo a sentence, not an episode.
• Per-character performance settings: stability, clarity, intensity, kept from one generation to the next.
• History of generated versions: you can go back to an earlier take.

Delivery : assembly and mastering

• Timecode assembly of generated or recorded lines.
• Levelling and mastering to several broadcast targets.
• Separate stems for the mix, with dialogue and reactions kept apart.
• Subtitle burn-in, in batches over a set of files.

Formats

Axeho slots into an existing chain: it reads what other tools produce and returns what they expect.

Client feedback

Retake requests arrive unstructured, in a spreadsheet or an email. Axeho imports them, finds the lines concerned, flags them, and leaves you a list of what is left to redo, instead of an annotated document to cross-reference by hand.

Working as a team

Projects can live in a shared folder: Drive, OneDrive, Dropbox, NAS. They are small: the heavy audio stays on each machine. A cooperative lock warns when two people open the same episode, without blocking the work, and a guard names the author before anything gets overwritten.

Worth knowing: voice generation runs on the workstation, not on a shared server. In a team, adaptation is shared and generation spreads across the machines.

Offline, online

Online services run on your own accounts, entered in the settings: you see your volumes, you choose your providers, and nothing goes through us.

See pricing Download