Direct answer: AI music, sound effects and localised dubbing differ a lot depending on whether they are billed per minute or per clip. At ElevenLabs list prices: music USD 0.15/minute, sound effects 0.12, automatic dubbing 0.33 with a watermark and 0.50 without, Dubbing v2 2.20; Google Lyria 3 is USD 0.04 per 30-second clip (about 0.08/minute). For a 10-minute export product film, music alone is USD 1.50, music plus SFX and watermark-free dubbing is about 6.74, and Dubbing v2 pushes it to about 23.74 — dubbing costs an order of magnitude more than music.
This guide is for manufacturers and content teams producing product films, trade-show videos and multilingual marketing assets. Data cut-off: 1 November 2026; prices come from official ElevenLabs and Google Cloud pages, in USD, excluding tax, with no FX conversion. The examples are editorial models, not quotes; duration, number of languages and watermark requirements must be replaced with your own project.
Music is billed per minute, clips per item, dubbing per minute with a watermark surcharge; in a video budget dubbing usually costs more than music; decide watermark removal and language count before comparing unit prices.
How does each vendor bill?
ElevenLabs: music USD 0.15/minute, SFX 0.12, dubbing 0.33 (watermark) / 0.50 (no watermark) / 2.20 (v2). Google Lyria 3 is USD 0.04 per 30-second clip, Lyria 3 Pro 0.08 per clip, Lyria 2 0.06 per clip.
| Capability | Vendor | Billing unit | Official price |
|---|---|---|---|
| Music generation | ElevenLabs | Minute | USD 0.15/minute; finetune USD 1.50 |
| Music generation | Google Lyria 3 | Clip (30 seconds) | USD 0.04 per clip (about 0.08/minute) |
| Music generation | Google Lyria 3 Pro | Clip (full song) | USD 0.08 per clip; duration not stated on the pricing page |
| Music generation | Google Lyria 2 | Clip | USD 0.06 per clip; duration not stated |
| Sound effects | ElevenLabs | Minute | USD 0.12/minute |
| Automatic dubbing | ElevenLabs | Minute | USD 0.33 (watermark) / 0.50 (no watermark) |
| Dubbing v2 | ElevenLabs | Minute | USD 2.20/minute |
| Voice isolator | ElevenLabs | Minute | USD 0.12/minute |
Sources: ElevenLabs API pricing, Google Cloud Vertex AI pricing (Lyria rows) and the Lyria 3 documentation (accessed 1 November 2026). Public list prices, excluding tax and subscription; Lyria 3 Pro / Lyria 2 clip durations are not stated on the pricing page, so this article does not convert them to per-minute.
What does a 10-minute product film cost?
Music alone USD 1.50; with 2 minutes of SFX 1.74; with watermark-free v1 dubbing 6.74; with v2 dubbing 23.74; adding voice isolation 24.94; using Lyria 3 clips for music about 6.04.
| Combination | Music | SFX (2 min) | Dubbing (10 min) | Voice isolation | Total |
|---|---|---|---|---|---|
| Music only (ElevenLabs) | 1.50 | — | — | — | 1.50 |
| Music + SFX | 1.50 | 0.24 | — | — | 1.74 |
| + v1 dubbing (no watermark) | 1.50 | 0.24 | 5.00 | — | 6.74 |
| + v2 dubbing | 1.50 | 0.24 | 22.00 | — | 23.74 |
| + v2 dubbing + voice isolation | 1.50 | 0.24 | 22.00 | 1.20 | 24.94 |
| Music via Lyria 3 (20 clips) | 0.80 | 0.24 | 5.00 | — | 6.04 |
Editorial model: music 10 min × 0.15, SFX 2 min × 0.12, v1 no-watermark dubbing 10 min × 0.50, v2 dubbing 10 min × 2.20, voice isolation 10 min × 0.12; Lyria 3 at 30 seconds per clip needs 20 clips × 0.04. Excludes tax, subscription, editing and asset licensing.
Why is dubbing so much more expensive than music?
Because dubbing bills per source minute and charges extra for watermark removal: v1 without a watermark is USD 0.50/minute and v2 is 2.20 — 3.3× and 14.7× the music rate. Multilingual projects multiply by the number of languages.
| Item | Unit price | Relative to music | Note |
|---|---|---|---|
| Music | USD 0.15/min | 1× | Per output minute |
| Sound effects | USD 0.12/min | 0.8× | Per generated minute |
| v1 dubbing (watermark) | USD 0.33/min | 2.2× | Automatic speaker detection |
| v1 dubbing (no watermark) | USD 0.50/min | 3.3× | Usually required for public release |
| v2 dubbing | USD 2.20/min | 14.7× | Preserves voice and emotion, 92 languages |
Official list prices; excludes tax and subscription. Multilingual projects add dubbing minutes per language — three languages triples the dubbing fee.
What should you watch with per-clip pricing?
Lyria 3's clip is 30 seconds, so 10 minutes needs 20 clips; Lyria 3 Pro and Lyria 2 do not state clip duration on the pricing page, so they cannot be converted to per-minute. Confirm output length in the console before signing.
| Model | Billing | Known duration | Conversion risk |
|---|---|---|---|
| Lyria 3 | USD 0.04/clip | 30-second clip | About 0.08/minute, but requires stitching |
| Lyria 3 Pro | USD 0.08/clip | Full song; duration not stated | Cannot compare per minute |
| Lyria 2 | USD 0.06/clip | Duration not stated | Cannot compare per minute |
| ElevenLabs music | USD 0.15/minute | 5-minute limit per generation | Long films need segments |
Sources: official pricing pages; the 5-minute limit comes from ElevenLabs' pricing page. Stitching adds editing and loudness-matching work.
Which setup fits which scenario?
Background music for short clips favours per-clip pricing; continuous long-form with commercial licensing favours per-minute; multilingual export films should budget dubbing as the main line; generate SFX only as needed.
| Scenario | Start with | Why | Do not ignore |
|---|---|---|---|
| Short-video background music | Per clip (Lyria 3) | 30 seconds is enough; low unit price | Stitching and licensing |
| 10-minute product film | Per minute (ElevenLabs) | Continuous, less stitching | 5-minute per-generation limit |
| Multilingual export film | v1 no-watermark dubbing | Controllable cost | Multiplies by language |
| Brand film / ad | v2 dubbing | Preserves voice and emotion | 4.4× the v1 price |
| Noisy field footage | Voice isolator | Improves usable rate | USD 0.12/minute |
Editorial guidance, not a vendor ranking; compare usable rate, language coverage and total cost on the same source clip.
What should procurement and acceptance check?
Billing unit, clip duration, watermark and commercial licence, language count, per-generation limits, and whether failures are billed — not just "price per minute".
| Acceptance item | Suggested definition | Evidence to keep |
|---|---|---|
| Billing unit | Minute / clip / character, clip duration | Price screenshots, invoice |
| Watermark and licence | No-watermark price, commercial scope | Terms, licence files |
| Languages and voices | Supported languages, voice preservation | Samples, listening notes |
| Per-generation limit | Maximum length, whether segmentation is needed | Docs, test |
| Failure billing | Failed generations, retries | Terms, invoice |
| Delivery format | Sample rate, bitrate, stems | Samples, spec |
| Human post | Stitching, mixing, lip-sync hours | Time logs |
| Compliance | Music copyright, likeness and voice rights | Licence chain, review records |
Editorial guidance, not legal advice; confirm music and dubbing licence scope before public release.
Who should read this — and who should not?
For teams producing product films, trade-show videos and multilingual marketing assets. Not for one-off subtitles, or projects without a defined release language and channel.
Read if: you are in marketing, export sales, content or procurement; you are comparing ElevenLabs and Google Lyria; you need a budget for a 10-minute film. Skip if: you only need text translation; you have no asset licensing; you want this article to make creative decisions. Supports: per-minute/per-clip modelling, dubbing version choice, multilingual budgeting and acceptance criteria.
Lingyi does not rank audio models or promise voice quality. To connect audio, video and multilingual assets to a content workflow, see LingHarness; to fold localisation cost into project budgets, talk to us.
Related reading: AI voice-over cost and licensing, AI video generation cost, and AI translation cost; this article only covers music, SFX and video dubbing localisation pricing.
FAQ
How is AI music billed?
Per minute or per clip. ElevenLabs music is USD 0.15/minute with a 5-minute per-generation limit; Google Lyria 3 is USD 0.04 per 30-second clip (about 0.08/minute); Lyria 3 Pro is 0.08 per clip and Lyria 2 0.06, with clip duration not stated on the pricing page.
What does a 10-minute product film cost?
Music alone USD 1.50; with 2 minutes of SFX 1.74; with watermark-free v1 dubbing 6.74; with v2 dubbing 23.74; adding voice isolation 24.94; music via Lyria 3 clips about 6.04. Excludes tax, subscription and labour.
Why is dubbing more expensive than music?
Dubbing bills per source minute and adds a watermark surcharge: v1 with watermark 0.33, without 0.50, v2 2.20 per minute — 2.2×, 3.3× and 14.7× the music rate. Multilingual work repeats it per language.
Is removing the watermark worth it?
Public release usually requires it; the price rises from 0.33 to 0.50 per minute (+52%), or USD 1.70 more for a 10-minute film. Compared with rework or rights risk it is usually worth it, but confirm the commercial licence scope.
How long is a Lyria clip?
Lyria 3 is explicitly a 30-second clip; Lyria 3 Pro and Lyria 2 list "per clip" without duration, so they cannot be converted to per-minute. Run one generation in the console to confirm output length and stitching cost.
How should multilingual dubbing be costed?
Multiply by language count: a 10-minute film in three languages is about USD 15 with v1 no-watermark or 66 with v2. Add per-language human review and lip-sync/subtitle adaptation.
How should we estimate our own project?
Four lines: music (minutes or clips), SFX (minutes), dubbing (minutes × languages × version), voice isolation (minutes). Test one sample clip for usable rate and post-production hours, then extrapolate — not unit price alone.
Method and limitations
This article uses only official pages accessible on 1 November 2026: ElevenLabs API pricing, Google Cloud Vertex AI pricing (Lyria rows) and the Lyria 3 documentation. All amounts are converted from public list prices and example assumptions; no enterprise agreements or reseller quotes are used. USD is not converted to RMB and no vendor ranking is implied. Lyria 3 Pro / Lyria 2 clip durations, commercial licences and failure-billing rules follow the official terms; the 10-minute film, 2 minutes of SFX and 20 clips are example assumptions, not benchmarks. The article does not assess sound quality, dubbing naturalness or language coverage and is not legal advice.
Sources and basis
Data cut-off: 1 November 2026. The following are official pages; prices, units and terms follow the original text and invoice.
- ElevenLabs, API Pricing: elevenlabs.io/pricing/api. Music USD 0.15/minute (5-minute limit), SFX 0.12, v1 dubbing 0.33/0.50, v2 2.20, voice isolation 0.12 per minute; music finetune USD 1.50.
- Google Cloud, Vertex AI generative AI pricing: cloud.google.com/vertex-ai/generative-ai/pricing. Lyria 3 USD 0.04 per 30-second clip, Lyria 3 Pro 0.08, Lyria 2 0.06 per clip.
- Google Cloud, Lyria 3 documentation: cloud.google.com/vertex-ai/generative-ai/docs/models/lyria/lyria-3. Model and generation details used to check clip and output specifications.