ElevenLabs Sound Effects v2 vs TTS

Choose Sound Effects v2 for non-speech events, ambience, impacts, and loops. Choose TTS Multilingual v2 when finished words must become a stock-voice MP3. Both use fal and the same private audio path, but their inputs and pricing units are deliberately separate.

Non-speech sound vs spoken voice1-22 seconds vs 3-2,000 characters1-4 credits vs 1-17 credits
Open Sound EffectsOpen TTS

ElevenLabs Sound Effects v2

Creates one private MP3 sound effect from a non-speech description through fal's v2 endpoint.

Best fit
Impacts, interfaces, transitions, foley, ambience, textures, and seamless background loops.
Output
One privately stored MP3 at 44.1 kHz and 128 kbps
Cost
About $0.002 per output second; 1-4 credits
Open Sound Effects
Sound Effects v2 model page

ElevenLabs TTS Multilingual v2

Creates one private MP3 voice track from finished text with a stock voice and speed.

Best fit
Narration, product reads, accessibility audio, and spoken drafts.
Output
One privately stored MP3 voice track
Cost
About $0.10 per 1,000 characters; 1-17 credits
Open TTS
TTS Multilingual v2 model page

Compare the same decision criteria

The decision starts from whether the deliverable should contain spoken words. Duration and character count are different supplier meters and never share one quote.

Live ElevenLabs audio contracts on Kanvora, verified August 22, 2026.
CriterionElevenLabs Sound Effects v2ElevenLabs TTS Multilingual v2Decision rule
Primary jobGenerate non-speech soundGenerate spoken voiceChoose from the audible result, not the shared vendor mark.
InputSound description and output settingsFinished script, stock voice, and speedDo not send dialogue text to the sound-effect route.
DurationExplicit whole seconds from 1 through 22Determined by text, voice, and speedSound Effects prices the requested output duration.
LoopOne shot or seamless loopNo loop controlUse Sound Effects for repeatable ambience, then inspect the seam.
Credits1-4 from output seconds1-17 from input charactersReview the correct unit before comparing totals.
Endpointfal-ai/elevenlabs/sound-effects/v2fal-ai/elevenlabs/tts/multilingual-v2Deprecated Sound Effects v1 is not registered.

On smaller screens, this matrix scrolls inside its own frame. The page itself does not scroll sideways.

Reserved audio proof without cross-model substitution

Both slots remain reserved until Kanvora approves source-tracked output from the exact endpoint. One model's audio is never presented as evidence for the other.

Text only - official sound proof reserved

Sound Effects v2 direction

Heavy wooden door closing in a large stone hall, sharp impact, short low-frequency tail, no speech and no music.

Five seconds, balanced influence, and one-shot output are attached. The exact hold is 1 credit.

Open this direction
Text only - official voice proof reserved

TTS Multilingual v2 rehearsal

Welcome to Kanvora. This voice rehearsal checks pronunciation, pacing, names, numbers, and punctuation before the finished track is published.

Rachel at natural speed is attached. Character count sets the exact hold.

Open this direction
What this evidence does not prove

Prepared directions do not prove event accuracy, loop quality, pronunciation, naturalness, loudness, or a universal winner.

Choose from what the audience must hear

These scenarios keep sound design and speech generation separate without bundling Noise Remover, Music, eleven-v3, dialogue, or cloning.

  1. A game button needs a short confirmation tone.

    Signal: No words should be audible.

    Choose Sound Effects v2.

    A concrete sound description, short duration, and one-shot ending match the endpoint.

    Create the sound direction

  2. A product script needs a calm spoken draft.

    Signal: The exact words must be voiced.

    Choose TTS Multilingual v2.

    TTS accepts finished text plus a stock voice and speed.

    Create the voice rehearsal

  3. Rain ambience must repeat under a long scene.

    Signal: The beginning and ending must meet cleanly.

    Choose Sound Effects with Seamless loop.

    The v2 route exposes loop behavior, but the stored seam still needs review.

    Open Sound Effects settings

  4. An existing noisy voice recording needs cleanup.

    Signal: The speech already exists in an MP3 or WAV.

    Choose the separate Noise Remover tool.

    Audio isolation accepts the private recording and prices the request from server-verified input duration.

    Open Noise Remover

Switching audio jobs creates a separate paid request

Wait for active or uncertain jobs to settle. Changing between sound design and speech changes endpoint, input schema, supplier meter, and credit hold.

Switching and handoff costs

  • Sound Effects bills requested output seconds at about $0.002 per second.
  • TTS bills submitted input characters at about $0.10 per 1,000 characters.
  • Kanvora rounds each supplier total up to a positive whole-credit hold under the same $0.012 coverage formula.
  • Neither request automatically triggers Noise Remover, Music, eleven-v3, dialogue, cloning, or the deprecated v1 endpoint.

What the comparison cannot prove

  • A matching prompt does not guarantee a specific recording, acoustic space, or loop seam.
  • TTS does not guarantee every name, number, acronym, language, or emotional delivery.
  • Reserved proof slots are not quality benchmarks.
  • The shared ElevenLabs mark does not make the two endpoints interchangeable.

Follow the decision into a focused model or workflow page instead of expanding this pilot into a generic directory.

Primary sources

Last verified: August 22, 2026. Product settings, costs, and provider paths can change after this date.

Design non-speech sound with Sound Effects. Voice finished words with TTS.

Both CTAs carry one legal live operation into the authenticated workbench. Safety checks, exact holds, and private storage stay at the real request boundary.

Open Sound Effects Open TTS