Seedance 2.5 Audio: Sound Generated in the Same Pass, at No Surcharge

What this tool does

• generate_audio defaults to true on Seedance 2.5, and VdoBloom sends it as true on every render in every tab — there is no audio switch to forget and none to find.
• Audio carries no surcharge on Seedance 2.5. The rate is 9.2 credits per second at 480p and 20.0 at 720p with sound or without: 5 seconds is 46 or 100 credits, 10 seconds 92 or 200, 30 seconds 276 or 600.
• Sound is produced by the video model in the same generation as the picture — one job, not a dubbing pass bolted on afterwards.
• Because Seedance 2.5 can render a single 30-second take, its audio arrives as one continuous 30-second pass rather than six clip-length takes that have to be crossfaded together.
• The prompt is the only audio control. In Text to Video and Image to Video you write it, up to 30,000 characters; on the Seedance Effects tab the effect supplies it. There is no dialogue field, no voice selector and no music picker in any of the three.
• Charging for audio is the norm elsewhere in the Seedance family's older tiers: Seedance 1.5 Pro costs 8 credits for a silent 5-second 720p clip and 15 credits for the same clip with audio.
• Other native-audio families on VdoBloom include the rest of the Seedance 2 line (Seedance 2, 2 Fast and 2 Mini), Wan 2.7 and Wan 2.6 Flash, MiniMax H3, SkyReels V4, PixVerse, Vidu Q3 and Q3 Turbo, and Lightricks LTX-2 Fast.
• Reference audio is supported for Seedance 2.5 in the main composer: the Seedance settings panel in Text to Video and Image to Video takes reference audio URLs. It cannot be combined with a first-frame or last-frame image in the same job, and the Seedance Effects flow does not send the field at all.
• The cheapest way to hear what Seedance 2.5 audio sounds like is a 4-second 480p render at 37 credits, which is more than the 10 free credits a new account starts with.

How it works

1. 1.Open Text to Video if you want to describe the sound yourself

   Go to /dashboard/video-creation/text-to-video/ and pick Seedance 2.5 from the Seedance group. This is the flow where the prompt is yours, which matters for audio because the prompt is the only thing steering it. Image to Video works the same way with a photo attached; the Seedance Effects tab replaces your prompt with the effect's.

2. 2.Write the sound into the prompt, not just the picture

   Say what the space sounds like, what the subject is doing with their voice, and what the ambience is under it. A prompt that only describes framing gives the audio pass nothing to work from. Actions with an obvious sound — footsteps, handling an object, speaking to camera — give the model more to score than a static portrait does.

3. 3.Or choose an effect for the sound it implies

   On the Seedance Effects tab the effect's stored prompt is what the model reads, so the effect is your audio direction. Pick on the strength of the action rather than only the visual style, and read what each live effect describes. Each one also states how many photos it needs from you, and the request is rejected with a specific count if you send the wrong number.

4. 4.Set duration and resolution — audio scales with them at no extra cost

   Every whole second from 4 to 30 is available, at 480p or 720p. The audio length simply follows the video length, and neither resolution changes what audio costs, because it costs nothing. That makes 480p the sensible place to audition sound: 10 seconds at 480p is 92 credits against 200 at 720p for the same audio pass.

5. 5.Audition short, then commit

   Run 4 to 6 seconds at 480p first, for 37 to 55 credits, and listen before you buy length. If the sound is wrong the fix is a rewritten prompt, a different effect or a different source photo — not a longer render. Credits are checked against your balance before submission and refunded automatically if the job fails.

Frequently asked questions

Related tools