One voice. Every language. Directed by ear.

WHO IT'S FOR

Where does your voice need to go?

Five kinds of work. Pick yours to see the job, an example and how to start.

CBrands & agenciesOne campaign voice in every market.E.G. A 30-second spot and its cut-downs, in French, German and Dutch.See how →
LGamesTrailers, barks and character lines for every market, including Japan, Korea and China.E.G. A launch trailer and 200 combat barks in Japanese, Korean and Mandarin.See how →
RStudios & streamingPromos, smaller shows and the back catalogue that never got dubbed.E.G. A six-part documentary series from the archive, in Spanish, Portuguese and German.See how →
LsCompanies & healthcarePresentations, reports and patient information, given a voice.E.G. Pre-surgery instructions in one calm voice, in four languages.See how →
RsNewsletters & mediaEach issue of your newsletter as a short narrated video.E.G. Friday's issue as a 90-second video with voice and on-screen text.See how →

This control comes from our dubbing studio: five speakers around one listener. Drag the voice.The five sit like the speakers of a 5.1 mix. Each card shows its place.

BRANDS
GAMES
STUDIOS
HEALTH
MEDIA

DUBBING STUDIO · HONDA SPEC SESSION · ONE LANE PER LANGUAGE

01 · BRANDS & AGENCIES · AI VOICE-OVER AND TRANSCREATION · C

Your AI voice sounds like an AI voice. Let's fix that.

One campaign voice in every market.

Your logo is the same in every country. Your voice can be too.

The job.

You need one campaign voice in every market. Then come the cut-downs, the late script change and the extra language. The voice is cloned once, with the speaker's written consent. Every translated line is rewritten to fit its slot, generated, and directed by ear until it sits in the mix.

For example.

A 30-second spot plus 15- and 6-second cut-downs, in French, German and Dutch, in one voice.

What you get back.

A full mix and clean voice stems for every language.
Lines timed to your cut, not stretched to fit it.
Revisions included.

Working through an agency? Your client, your credit, our hours.

CLIENT WORK, SHOWN WITH PERMISSION

Book 30 minutes Email us a script

LINE IDENJAKOZHFIT

BRK_001Flank left!0.9 s0.88 s0.93 s0.86 s✓

BRK_002Reloading, cover me.1.4 s1.38 s1.44 s1.35 s✓

BRK_003They're behind the gate!1.3 s1.33 s1.26 s1.29 s✓

BRK_004Hold this position.1.1 s1.08 s1.13 s1.05 s✓

BRK_005Medic! I'm hit.1.0 s1.02 s0.97 s1.04 s✓

BRK_006Clear. Moving up.1.2 s1.17 s1.23 s1.15 s✓

EXAMPLE SHEET. YOUR LINES, YOUR IDS.

EXAMPLE LINE SHEET · EN → JA · KO · ZH

02 · GAMES · AI VOICE LOCALISATION FOR GAMES · L

Every market hears the same hero.

Trailers, barks and character lines for every market, including Japan, Korea and China.

The job.

Trailers, store videos, barks and character lines. In every language you ship, including Japanese, Korean and Mandarin. Each character keeps one voice across all of them. Every line is fitted to the length of the original. Japanese lines are counted in morae, the unit Japanese timing uses, so they fit the same slot.

For example.

A launch trailer and 200 combat barks in Japanese, Korean and Mandarin, one voice per character.

What you get back.

One file per line, named to match your line sheet.
Each line timed to its slot.
The same character voice in every language.

HEAR A JAPANESE VERSION

Book 30 minutes Email us a line sheet

DUBBING STUDIO · EVERY LANGUAGE ON ONE TIMELINE

03 · STUDIOS & STREAMING · AI DUBBING FOR PROMOS AND BACK CATALOGUE · R

Dub the back catalogue that's waiting.

Promos, smaller shows and the back catalogue that never got dubbed.

The job.

Your big titles keep their cast. Promos, trailers, smaller series and the archive often don't get dubbed at all, because casting every language doesn't pay off. Each character keeps one voice in every language. For your titles, voices are cloned only with the performer's written consent, or designed new for the role. Every line is directed, and delivered as stems that drop into your mix.

For example.

A six-part documentary series from your archive, in Spanish, Portuguese and German.

What you get back.

Stems per character and per language.
WAV stems, or an AAF that opens in your mixer's session.
Stereo as standard. 5.1 and a headphone version on request.

7 VERSIONS, ONE FILM

Book 30 minutes Email us a title

VOICE BRIEF · PATIENT INFORMATION

VOICE: CALM, UNHURRIED

Pace 0.92×

Pause after sentences 0.6 s

Please don't eat or drink for six hours before your appointment. Bring your list of medicines.

EXAMPLE BRIEF.

EXAMPLE VOICE BRIEF · PATIENT INFORMATION

04 · COMPANIES & HEALTHCARE · REPORTS, SLIDES, PATIENT INFORMATION · Ls

Reports, slides and patient leaflets, spoken.

Presentations, reports and patient information, given a voice.

The job.

Most information never gets a voice. A studio session feels too big for a quarterly report, a training deck or the leaflet a patient reads before surgery. We design a voice that suits you: calm and unhurried if it's for patients. Then that voice reads your material in the languages your people speak. You approve every translation before it's voiced.

For example.

Pre-surgery instructions as a two-minute video, in Dutch, English, Turkish and Arabic, in one calm voice.

What you get back.

A voice designed for your organisation.
Narrated video or audio files, per language.
Pace and pauses set for the listener.

WE ONLY NEED THE TEXT. NEVER PATIENT DATA.

Send one page. We'll tell you what it could sound like.

Book 30 minutes Email us one page

ISSUE

This week: three things about the harbour that nobody tells visitors.

1. The cranes are older than the bridge.
2. The ferry is free before 7.
3. The best view is from the fish market roof.

EXAMPLE ISSUE AND FILM.

ISSUE → STORYBOARD → FILM

05 · NEWSLETTERS & MEDIA · NEWSLETTER TO NARRATED VIDEO · Rs

Your newsletter, as a film.

Each issue of your newsletter as a short narrated video.

The job.

You already write the issue every week. We turn it into a short narrated video with on-screen text, in one or more languages. The voice can be yours, cloned with your consent, or one designed for you.

For example.

Every Friday issue as a 90-second video, vertical for phones and wide for your site, in English and Spanish.

What you get back.

One video per issue, vertical and wide.
Your voice, or one made for you.
A weekly rhythm you set.

Book 30 minutes Email us an issue

RAW AI vs AyP SOUND

Same picture. Two soundtracks.

One is what an AI video model made, sound and all. The other is the same picture after a session in the studio. Flip the switch while it plays.

NOW: RAW AI

RAW AI AyP SOUND

The model's own sound, only levelled to match.

Both sides play at the same loudness, so you compare sound, not volume.

Space flips · A / B jump to a side

Clip generated with LTX-2.5 (Lightricks), picture and sound in one pass.

WORK

Switch language mid-sentence.

Press a flag while a film plays. It carries on from the same moment, in the new language. Every version plays at the same loudness (−16 LUFS), so you compare voices, not volume.

Versions marked ORIGINAL are the original film. Every other version is an AI voice, directed by AyP Sound.

Honda · spec work. A 60-second spot with two voices: a race commentator and a poet. EN DE ES FR IT NL PT
Chanel · spec work. Chanel's 30-second film, voiced again in Spanish, French, Italian, Japanese and Dutch. EN ES FR IT JA NL
Apple iPhone Duo · spec work. A three-minute product film, carried by one narrator. AyP Sound versions in production.
Waterdrop · client work, via UNIT9. Waterdrop's 60-second film, voiced in English, German, French and Dutch. EN DE FR NL
Wilkinson Sword · client work, via Pablo London. Product tags: Quattro in English and French, Hydro 5 in German. EN FR DE

Tools give you a voice. AyP Sound gives you a performance.

STUDIO · BUILT IN-HOUSE

Every language on one timeline.

We built our own dubbing studio. The original sits on top, each language underneath, line under line. These are the real sessions behind our Honda and Apple iPhone Duo spec films.

SCREEN CAPTURES FROM THE AyP SOUND DUBBING STUDIO, SHOWING OUR HONDA SPEC FILM. THE CURSOR IS THE STUDIO'S OWN VIRTUAL CURSOR.

DAW · BUILT IN-HOUSE

A voice is never judged in silence.

Our DAW, the audio editor we built, puts picture, effects and music around every draft. So each line is heard the way your audience will hear it.

WE SELL THE VOICE. THE EFFECTS AND MUSIC AROUND IT ARE THERE SO EACH LINE IS HEARD IN PLACE.

SCREEN CAPTURES FROM THE AyP SOUND DAW.

HOW IT WORKS

From one script to every language.

VO · YOUR PROJECT

01 · Send The film and the script. A link is enough. No script? We transcribe it.
02 · Voice For your project, we clone the voice once, with the speaker's written consent. Or we design a new one with you.
03 · Direct Each line is rewritten to fit its moment in the picture, then generated and directed by ear until it lands. Names, titles and wordplay are worked out by hand, not left to the machine.
04 · Deliver A full mix and clean voice stems for every language. Revisions included.

You hear every language before it's final.

THE DAW

ABOUT

A sound engineer, listening.

AyP Sound (say it A-Y-P Sound) is Juan Pablo Thummler. He is a sound engineer with more than ten years in audio post and his own studio in the Netherlands.

The AI makes the voice. He chooses the take, sets the pauses and the stress, and mixes it to the picture. Every line goes through his ears before it goes to you.

He built the dubbing studio and the DAW on this page to do it.

STUDIO: NETHERLANDS
IN AUDIO POST: SINCE 2014
LANGUAGES ON THIS PAGE: DE EN ES FR IT JA NL PT
DIRECTED BY: JUAN, LINE BY LINE

NEXT STEP

Tell me about the project. I'll tell you what I hear.

A 30-minute video call with Juan. Bring a clip, a script or just the question. You'll get a straight answer on whether this is a fit.

Juan

Book a 30-minute call or write to [email protected]