Google video enhancements add English-only voiceover: opt out or ship

Google Ads video enhancements use AI to make versions of your uploaded video: flipped or extended into 16:9, 1:1, and 9:16, shorter clips with Gemini picking key moments, and an added voiceover that Google AI builds from one of your text assets (Google Ads Help, read 2026-10-05). The voiceover is only available for accounts and assets in English, and you can opt out by unchecking the boxes at campaign creation or edit. If you want one exact cut running in the holiday window, opt out or upload the cuts yourself.

What it does by campaign type

The page says availability varies. Performance Max supports all three enhancements, App campaigns support orientation changes and shorter versions only, and Video views and Video reach campaigns support orientation changes only. No minimum video length is specified.

Decide per asset, not per account

Enhancements are most useful when you have one horizontal master and no time. They are less useful when the cut is the creative: a countdown that ends on a date, an offer line, or a voice that has to match the footage. In that case the cleanest control is to supply each orientation yourself and leave nothing for the system to flip.

On Sume you can do that without a second shoot. Gemini Omni Flash 1.1 makes 16:9 or 9:16 clips of 3 to 10 seconds with native synced audio, so one prompt gives a spoken, scored 9:16 and a 16:9 sibling. An 8-second 720p clip is $1.00 each (list $0.10 per second times 1.25, times 8), so the pair is $2.00.

A note on languages

Because voiceover is only available for English accounts and assets, a multi-market holiday plan has to treat other languages differently: the enhancement will not speak for you. Burn the local offer as captions, and keep the spoken line to what the clip already carries.

If you keep the voiceover on

Read the voiceover text asset aloud as a viewer would hear it. Google AI selects one of your text assets and merges it with the creative, so a headline that reads fine in a text ad may sound abrupt as speech. Write at least one description that reads as a sentence.

If a clip is silent, burn the offer as text yourself: POST /v1/video-captions takes cues with text, start, and end, at $0.20 per job. The viewer then sees the offer whether or not Google adds speech.