Audio and speech

Dialogue, TTS, voice, cleanup, separation, translation and localization

• articleWhat do the WER and speaker-similarity (SIM) numbers produced by the Seed-TTS-eval scripts actually compute, and what do they not tell me about a generated voice?
  At commit 752f4297 (2024-06-14) seed-tts-eval defines WER as the mean per-utterance word error rate of a Whisper-large-v3 (English) or Paraformer-zh (Mandarin, character-level) transcript against the target text after punctuation stripping, and SIM as the mean cosine similarity between speaker embeddings of the synthesized file and the prompt (reference) audio from a WavLM-large speaker-verification model. Both are proxies: WER measures ASR intelligibility of the text, SIM measures embedding closeness to the prompt clip; neither measures naturalness, prosody, emotion or timing, and the authors themselves say a lower WER can mean more standardized, less characteristic speech. The scripts report only means (no confidence interval) and skip unscorable files.
• articleOf the free sound-effects libraries I might use (Freesound, Sonniss GDC bundle, BBC Sound Effects, Pixabay sound effects), which say anything about AI or machine-learning training, in which document, and does the short summary page say the same as the legal text?
  A comparison at dated reads. Freesound: training allowed if each sound's CC licence is respected, plus uploader preferences since July 2026. Sonniss: AI training expressly prohibited in the licence and repeated on the bundle page. BBC: AI development or training needs BBC permission. Pixabay: the licence summary page says nothing about machine learning, while the Terms of Service prohibit scraping and data mining "for machine learning purposes" and describe a contributor opt-out for Pixabay's own AI training; so a reader of the summary alone would miss the ML clause. Mixkit's sound-effects licence text is script-rendered and was not obtained.
• articleWhat does the Sonniss #GameAudioGDC bundle licence (version 2.0, effective 27 August 2026) permit and forbid for commercial use, redistribution, attribution and AI training, how did it change, and where does the bundle page summary differ from the legal text?
  At the snapshot of 2026-10-10 the licence version 2.0 allows use and modification in personal and commercial projects without attribution, bars selling the sounds as they come and bars supplying them to others as sound effects (also in libraries or packs, modified or not), and expressly prohibits using them to train, develop or enhance AI technologies. The history page says the AI section was added on 28 January 2024, that earlier text allowed free distribution, and that the version published on the download day governs. The bundle page summarises the sounds as licensed for media production "only" and the AI ban as "training", both narrower or different in wording than the legal text.
• articleCan I use sound generated by MMAudio, HunyuanVideo-Foley or ThinkSound in a commercial delivery, according to what their licences and cards actually say?
  The texts differ per component. MMAudio: code MIT; checkpoints CC BY-NC 4.0 (non-commercial) with a README warning that training-dataset licences apply and commercial suitability is not guaranteed. HunyuanVideo-Foley: Tencent Hunyuan Community License, which says Tencent claims no rights in Outputs but excludes EU/UK/South Korea, bars using Outputs to improve other AI models, restricts use outside the Territory and adds a 100M monthly-active-user condition. ThinkSound: HF card says Apache-2.0 while the GitHub README says commercial use is not permitted (see -05). No text read grants clearance; this record does not conclude legal permission.
• articleWhat licences and documented limits apply to aeneas and Gentle for text-to-audio synchronisation, and what is not stated about Gentle's models?
  aeneas is AGPL-3.0 code (README 1.7.3) that aligns text to speech by MFCC/DTW against synthesised speech; it expects spoken audio matching the text, is unsuited to song, and its maintainer says it was not designed for word-level alignment. Gentle's source is MIT, but its Kaldi acoustic models are downloaded at install time from a separate host, and no model licence appears in the repository files read. Neither README gives an accuracy figure.
• articleWhat do the LRS2 and LRS3 pages and licence file say about permitted use of the data, and do they address model training?
  LRS2 is offered for non-commercial academic research only after signing a BBC R&D Data Sharing agreement (agreement not read); LRS3 is described as available for research purposes under CC BY 4.0 while its licence file says the underlying TED video excerpts must respect TED terms and CC BY-NC-ND 4.0, copyright stays with the video owners, and downloads are no longer offered on the site. None of these texts mention model training.
• articleWhat terms do MedleyDB and FMA state for their audio, annotations and metadata, and do they address model training?
  MedleyDB's pages offer the dataset for non-commercial research under Creative Commons (named both as Attribution-NonCommercial and as BY-NC-SA 4.0) and ask users not to republish without consent; the Zenodo audio record (v1.0) is access-restricted with no licence field and points to a GitHub repository for annotations whose LICENSE is MIT, while the description page says audio and annotations are released under BY-NC-SA. FMA separates MIT code, CC BY 4.0 metadata and per-artist audio licences, for research purposes. Neither text addresses model training.
• articleWhat consent, watermark and disclosure mechanisms does VibeVoice document, and are they still enforced in the repository?
  The VibeVoice-1.5B HF card (c00898d2, MIT front matter) documents consent-based out-of-scope uses, an audible disclaimer and an imperceptible watermark; the GitHub README (16fb2cb1) says the TTS code was removed on 2025-09-05 after misuse and marks the 1.5B TTS entry Disabled. Card and README describe different artefacts at different revisions.
• articleDawDreamer docs at e0daff3c state both that automation granularity is the block size and that automation data is interpreted at audio rate; the render loop calls automateParameters once per block with the buffer size, consistent with block-wise application (interpolation code not...
  DawDreamer docs at e0daff3c state both that automation granularity is the block size and that automation data is interpreted at audio rate; the render loop calls automateParameters once per block with the buffer size, consistent with block-wise application (interpolation code not read).
• articleFor the audio-capable world models and sync-focused sound tools found here, what do the repositories and model cards actually say about licence and weights per component, and where do they disagree?
  HelixWorld (Noiz AI) is the one audio-visual world model with a published checkpoint: code and the HF card say Apache-2.0, but the repository NOTICE says the LTX-2 code and model assets inside it are governed by the LTX-2 Community License (paid licence for entities with at least US$10M annual revenue; fine-tuned weights count as derivatives), the paper says HelixWorld builds on LTX-2.3, and the Gemma text encoder has its own terms. WorldSonus is CC BY-NC 4.0 with .pt weights (pickle-capable). V-AURA code is MIT but its checkpoint is a tar.gz with no stated weight terms. Kling-Foley publishes a paper and a CC-BY-4.0-labelled benchmark dataset but no code or weights were found.
• articleWhat do Freesound's per-sound Creative Commons licences, attribution guidance and AI-training answers say at the access date, and where does Freesound's own summary go beyond or differ from the Creative Commons legal texts?
  Freesound distributes each sound under the licence its uploader chose (CC0, CC BY or CC BY-NC 4.0, plus legacy Sampling+). Its FAQ now answers AI training as "yes as long as individual sounds' Creative Commons licenses are respected" and, as of July 2026, adds uploader-level Gen AI preferences (open-source-only, non-commercial open-source-only, no generative AI, and a voice opt-out) on top of the licence, exposed to API users for filtering and in Data Packs. Freesound labels its CC-to-training reading as its own interpretation (for example CC-BY requires disclosing the training set); the CC legal code text does not mention AI and the CC FAQ says the answer depends.
• articleThe BBC Sound Effects archive says it is free to download: what does the RemArc licence text actually permit (personal, research, education, commercial), what must I do, and what does it say about text and data mining or AI?
  The BBC Sound Effects site is a script-rendered app whose licence text is embedded in its JavaScript bundle (read as text only). That text (the BBC's Content Licence for RemArc) permits non-commercial, personal or research use and formal education use by students and staff, requires a credit, bars sharing such as uploading to social media, and says text and data mining, computer analysis, or developing or training AI need BBC permission and may cost a fee. Commercial use is routed to Pro Sound Effects; metadata is under the Open Government Licence v3.
• articleWhere do the licence layers of an open V2A model disagree or stack (ThinkSound's Apache vs non-commercial text, Stability's VAE licence, VGGSound/AudioSet/BBC training data)?
  ThinkSound publishes four incompatible-looking statements: HF weights card Apache-2.0 (S15, S16), GitHub README Apache-2.0 headline but research-only/no commercial use (S17), a bundled Stability AI Community License for the fine-tuned VAE (S18; commercial use needs registration and a revenue condition), and a paper statement that the dataset is strictly academic/non-commercial (S14). Its training sources include VGGSound (page: CC BY 4.0, copyright remains with video owners), AudioSet (CC BY 4.0 annotations for YouTube-linked data), Freesound and BBC Sound Effects (site: personal/educational use or license). A permissive label on one file does not resolve the others.
• articleCan the Meta open speech-translation toolchain (SeamlessM4T v2, SeamlessStreaming, SeamlessExpressive, NLLB-200) be used in a commercial dubbing pipeline, and which licence applies to which component?
  Seamless code (and the W2v-BERT 2.0 encoder, UnitY2 aligner, toxicity tools and the text-only part of mExpresso) is MIT; SeamlessM4T v1/v2 and SeamlessStreaming weights are CC BY-NC 4.0 (the HF v2 card agrees); Seamless and SeamlessExpressive weights fall under a separate Seamless Licensing Agreement limited to noncommercial research that also bars commercial use of outputs, terminates on IP litigation against Meta and requires indemnity; NLLB-200 models are CC BY-NC 4.0 while fairseq code is MIT. Commercial dubbing with these weights is not permitted by these texts.
• articleUnder what terms may an agent use AudioSeal code and checkpoints, and how does the repo's open detector sit with its own paper's advice?
  At audioseal e63a8a0e the LICENSE is MIT (Meta Platforms) and a dated README news line (2024-04-02) says the MIT licence now includes the model weights and allows commercial use, while the README License section names only the code; the Hugging Face card front matter says license: mit. The AudioSeal paper (arXiv v2) advises that detector weights be kept confidential because a white-box attack raised detection error by around 80% (author-reported), which sits uneasily with publicly released detector checkpoints.
• articleWhat do the model cards of four open TTS releases say about voice consent, impersonation, provenance marking and licence, and where do card statements and licence metadata disagree?
  At the pinned revisions, the VibeVoice-1.5B card declares MIT in its front matter while also limiting the model to research use and listing voice impersonation without explicit, recorded consent as out of scope, and claims an audible disclaimer and watermark in outputs (vendor statements); the XTTS-v2 card (Coqui Public Model License) advertises cloning from a 6-second clip with no consent text; the F5-TTS card declares CC BY-NC 4.0 with no use-policy text.
• articleWhat does the W3C imsc-tests suite contain for bidi, ruby and vertical IMSC features, and does any public source give per-engine results for it?
  The imsc-tests suite publishes stimuli and reference PNGs but no engine results (producer reading); per-engine evidence exists only as implementer-reported processor results in the TTML2 report.
• articleAt the mix-to-delivery handoff, which stem, M&E and track-layout rules do Netflix v1.4, EBU R 123 and AS-11 state, and what does the receiver still not get?
  P2-04's statement that role labelling is positional (track number) holds for AS-11 UK DPP HD; for X-series files MCA labels carry roles (CA-B066 C2-03 reading).
• articleWhat does the public AS-11 UK DPP HD specification require for audio channels and track layout, and where does loudness appear?
  AMWA AS-11 UK DPP HD requires one mono channel per track, 24-bit/48 kHz PCM and exactly 4 or 16 channels, and its text contains no loudness requirement; AS-11 X9 instead describes audio layout through MCA labelling (Audio Layout Mode 0).
• articleWhat reading-speed, line-length and duration limits do Netflix, BBC and EBU-TT-D state, per language, and where do they disagree?
  Line and speed limits are scoped by document and platform: Netflix English (USA) sets 42 characters per line with 20/17 cps adult/children; BBC sets 37 characters for broadcast (Teletext) and a percentage-of-width rule online, with a 160-180 wpm live word rate. These must be presented per source, not merged into one caption limit.
• articleWhat do the D-Cinema audio channel symbols VIN and HI (and the other common channel symbols) mean in SMPTE ST 428-12, and is the document public?
  SMPTE ST 428-12:2013 Table 1 defines HI (UL byte 0Eh) as a dedicated hearing-impaired dialog-intelligibility channel and VIN (0Fh) as a dedicated visually-impaired narration channel; the standard labels channels and does not require a soundtrack to carry them (producer reading of scope; Table 1 rows reinspected).
• articleIs SMPTE ST 429-16 publicly readable at pub.smpte.org, and what does it define for audio channel labels such as VIN and HI?
  ST 429-16:2023-09 gives MainSoundConfiguration a recommended (should) ABNF syntax with 1-8 character tokens and delegates label meanings to ST 428-12 symbols; HI and VIN appear only in an informative Annex B example.
• articleFor DCP picture and sound essence, which checks do libdcp's verifier and ClairMeta actually run (read from the check bodies), where do their limits disagree, and what is not checked by either?
  RDD 52:2020 7.3 requires 48 kHz audio essence; libdcp (891facc8) flags any other sampling rate and any bit depth other than 24, while ClairMeta (b3b66a02) settings accept 48000 and 96000.
• articlepython-soxr (351f265b) defaults to quality 'HQ' in both its one-shot resample() and its streaming class (curator read), so 'soxr default' in Python differs from libsoxr's C one-shot default reported as LQ by the...
  python-soxr (351f265b) defaults to quality 'HQ' in both its one-shot resample() and its streaming class (curator read), so 'soxr default' in Python differs from libsoxr's C one-shot default reported as LQ by the producer.
• articlepython-soundfile 0.14.0 sf.read defaults to dtype 'float64' (torch pipelines typically expect float32); the package code is BSD-licensed by its author (licence header...
  python-soundfile 0.14.0 sf.read defaults to dtype 'float64' (torch pipelines typically expect float32); the package code is BSD-licensed by its author (licence header read).
• articleTorchCodec documents 0.14-0.17 as requiring torch >= 2.11 (README table), but the 0.17.0 PyPI metadata declares no torch dependency (only requires_python >= 3.10), so pip will not enforce the documented torch pin (curator...
  TorchCodec documents 0.14-0.17 as requiring torch >= 2.11 (README table), but the 0.17.0 PyPI metadata declares no torch dependency (only requires_python >= 3.10), so pip will not enforce the documented torch pin (curator addition).
• articleAt pytorch/audio main 1255104d, torchaudio exports only load/save (TorchCodec wrappers) and their torchcodec helpers plus datasets, transforms, functional, models, pipelines, compliance and utils (no info, backend selectors, io StreamReader/Writer or sox_effects); torchaudio 2.11.0 on PyPI...
  At pytorch/audio main 1255104d, torchaudio exports only load/save (TorchCodec wrappers) and their torchcodec helpers plus datasets, transforms, functional, models, pipelines, compliance and utils (no info, backend selectors, io StreamReader/Writer or sox_effects); torchaudio 2.11.0 on PyPI declares no dependencies, so torchcodec (and torch) must be installed separately.
• articlepedalboard v0.9.26 is GPLv3 and statically compiles libraries including JUCE 6 (dual commercial/GPLv3); its x86_64 wheels require AVX; the README states no determinism or bit-exact guarantee (producer search; curator read the licence/platform...
  pedalboard v0.9.26 is GPLv3 and statically compiles libraries including JUCE 6 (dual commercial/GPLv3); its x86_64 wheels require AVX; the README states no determinism or bit-exact guarantee (producer search; curator read the licence/platform lines).
• articleWhat should a dialogue editor check on AI-cleaned dialogue before delivery, given the failure modes and behaviours that vendor pages, repositories and papers actually document?
  No source read provides a delivery checklist for AI-cleaned dialogue, so this candidate assembles one from documented behaviours and labels each check as a proposal. The documented behaviours cover timing (STFT delay, chunk re-alignment, export drift, export parity), format (mono downmix, 44.1/48 kHz), level (normalisation, rescaling), over-processing (artifacts, silent waveforms, unprocessed syllables), breath leakage, word changes in generative enhancement, tool-version defects and where the audio is processed. Room tone and ADR conventions, and disclosure of AI use to a client, are covered by other batches.
• articleWhen an AI dialogue-enhancement tool sounds clean, how can a dialogue editor tell whether it has changed the words, and what do recent papers say about objective metrics and listening tests missing such changes?
  Three recent arXiv papers say generative enhancers can substitute or omit phonemes while scoring well on reference-free metrics: the URGENT 2025 P.808 study reports a generative system with top DNSMOS/NISQA and high MOS but very poor phone fidelity and says ACR tests may also miss it; Yamauchi et al. say non-intrusive filtering often fails to detect such errors; Shetu et al. find limited hallucination by WER/LPS at moderate SNR but more at very low SNR. Their remedy is a content-based metric (WER, ESTOI, LPS) alongside MOS. No source tests DeepFilterNet or Resemble Enhance for this, so applying it to production dialogue tools is a proposal.
• articleFor the open models people name for dialogue cleanup (DeepFilterNet, Resemble Enhance, Demucs/HTDemucs), what inputs, outputs, controls and sync/level behaviour do their repositories document or imply, what do they actually claim about dialogue, and what are the per-component licences?
  At pinned commits: DeepFilterNet is a 48 kHz real-time noise suppressor whose Python CLI compensates its STFT delay by default while the Rust binary only does so with -D; Resemble Enhance averages to mono, works at 44.1 kHz, processes long audio in overlapping 30 s chunks with per-chunk peak normalisation and cross-correlation re-alignment, and is described as also restoring distortions and extending bandwidth; Demucs documents music stems only and its README (searched) never mentions speech or dialogue, so "Demucs for dialogue" is not a documented use. The Cocktail Fork paper is the film-relevant separation reference, but evaluates on synthetic single-channel mixes. Code licences are stated (MIT, or MIT/Apache-2.0); weights licences are stated only for Resemble (card says MIT), not DeepFilterNet or Demucs.
• articleWhich open, local ComfyUI workflows give explicit camera control, point-trajectory motion, audio-driven lip-sync and motion transfer, and what do they need?
  ComfyUI official tutorials (captured 2026-10-23), all Wan-family (Wan licences Apache-2.0 for 2.x per C3-B370-03; check each repo): (1) Wan2.2 Fun Camera Control (alibaba-pai/Wan2.2-Fun-A14B-Control-Camera): preset camera moves - pan up/down/left/right, zoom in/out and combinations - from a start image; measured by Comfy on an RTX 4090D 24 GB at 640x640, 81 frames: fp8_scaled ~536 s first run (84% VRAM), fp8_scaled + 4-step LoRA ~108 s first / ~71 s second (89% VRAM). (2) Wan-Move (Wan-Move-14B-480P): dense point trajectories propagate first-frame features for region-level motion control, 5 s at 480p, no extra motion modules. (3) Wan2.2-S2V-14B: image + audio to synchronized talking video, "minute-level generation" (fp8 or bf16 weights plus an audio encoder). (4) InfiniteTalk (on Wan2.1 I2V 14B 480p fp8): audio-driven full-body dubbing that preserves identity, background and camera movement while adding lip-sync and body motion. (5) Wan Animate 2: motion transfer from a driving vide...
• articleWhat do the open-weight licences of LTX-2.x, MiniMax H3, Apple SHARP, Meta SAM and Fish Audio S2 actually allow for commercial use and where?
  Licence texts read 2026-10-23 (not legal advice): LTX-2.x Community License (2026-08-11) grants broad rights subject to use restrictions (Attachment A), but entities with annual revenues of at least $10,000,000 must obtain a paid commercial licence for any non-non-commercial use; unlicensed commercial use by such entities is prohibited and fees may be charged retroactively. MiniMax H3 Community License (2026-08-02) is territorially limited: the "Applicable Territory" is worldwide EXCLUDING the European Union, the United Kingdom, the Republic of Korea and the United States, so the open-weights grant does not cover users there (MiniMax invites licence requests); also prior written authorization if commercial products/services exceed US$20M yearly revenue, and "MiniMax H3" must be displayed in commercial UIs. Apple SHARP model licence: Research Purposes only (non-commercial scientific research), derivatives likewise; excludes product development or commercial services. Meta SAM License...
• articleCan I compare reported DeSync and IB-score numbers across Kling-Foley, HunyuanVideo-Foley, MMAudio, V-AURA and world-model audio papers, and what do real recordings score?
  Only partly. The same system gets different numbers in different papers: V-AURA is 155 ms in its own paper (2.56 s clips), 0.654 s in MMAudio (8 s, two 4.8 s crops), 0.72 s and 0.86 s in HunyuanVideo-Foley's tables, and 1.287 s in WorldSonus (5 s clips). Kling-Foley reuses MMAudio's baseline rows but reports an MMAudio row (IB 29.26, DeSync 0.45) that matches none of MMAudio's four published rows, and its abstract claims best distribution matching while its own Table 2 shows MMAudio ahead on FD and KL. HunyuanVideo-Foley's own VGGSound-Test table has MMAudio at lower DeSync (0.47) than itself (0.53). The only ground-truth reference found is Foley-Omni's V2ST-Bench, where real recorded audio scores DeSync 0.14.
• articleWhen a paper says its audio is "in sync", what do DeSync, IB-score, AV-Align and the human scores actually measure, and how is each computed?
  Five different things are called "synchronization" in the sources. DeSync (MMAudio and later papers) is a learned estimator's predicted audio-video offset in seconds, computed by the Synchformer model on 4.8 s windows. V-AURA's earlier "Sync" is the same idea reported as mean absolute offset in milliseconds over 2.56 s clips. AV-Align (TempoTokens) is a signal-processing check that audio energy peaks and video motion peaks fall within three frames of each other. MMAudio also reports a model-free onset accuracy on Greatest Hits. Human preference and MOS studies (Veo, Ovi, HunyuanVideo-Foley) ask people. IB-score is not a timing metric: it is semantic similarity.
• articleDo interactive world models make sound today, and which audio-capable ones are only a paper, a demo, or something I can actually download?
  The Genie 3 blog, model page and prompt guide describe visuals, navigation and text events but no sound (the words audio, sound, speech and music occur only in site navigation in the text read). Veo 3 and 3.1 do generate audio natively, but that is a video generator, not an interactive world model. Audio-capable interactive world models exist only as 2025-2026 preprints; of those inspected, only HelixWorld (Preview v1) ships a checkpoint that runs the full audio-visual model; WorldSonus ships weights for an add-on sound generator under a non-commercial licence. All performance numbers are the authors' own.
• articleWhich TTS voices and music models lead blind preference tests in Oct 2026, and at what API price?
  Artificial Analysis Speech Arena, provider-voice board (rendered 2026-10-23): Eleven v4 Turbo 1335 (+/-17; Sept 2026; $40/1M chars), Eleven v4 1329 ($80), Qwen-Audio-3.1-TTS-Plus 1301 ($19.3), Sonic 3.6 1289 (Aug 2026; $49), Gemini 3.8 Flash TTS 1278 ($16.5), Realtime TTS-2 1267 ($20.8), Qwen-Audio-3.0-TTS-Plus 1265, Gemini 3.8 Flash-Lite TTS 1251 ($11), Fish Audio Drama 3 Preview 1248 (Oct 2026; $15), Simba 3.2 1245 ($6.6), Voxtral 3 TTS 1218 ($13); ~2,000-3,400 samples each, 8 voices per model. AA-Music-Instrumental v1.1 (anchor MiniMax Music 3.0 = 1000): Suno v6 1148 (+/-22), Suno v6-mini 1114, Mureka V9.5 1084, Mureka V9 1083, Lyria 3 Pro 1080, Lyria 3.5 1078, Eleven Music 1021, Stable Audio 3 Large 1014, Eleven Music v2 1011, MiniMax Music 2.5+ 1007, ..., Lyria 2 926, MusicGen 868 (~2,100-2,300 samples each; separate vocal board exists, not captured). Use: price/quality trade-offs are large (e.g. Gemini 3.8 Flash TTS within ~57 Elo of Eleven v4 at about a fifth of the price); F...
• articleWhat do XAG 106 and GAG require for screen-reader and screen-narration support, and how do they differ?
  XAG 106 asks games to voice all core UI text (menus, HUD, state changes, timed events) through the platform screen reader or a speech synthesizer; recorded audio files work but are 'not ideal'. Each control should be announced with role, state and position (e.g. 'Gamma, slider, 38%, 6 of 9'), non-decorative images need alt text, the game language should be exposed to screen readers, and focus order should follow meaning or visual flow. The Game Accessibility Guidelines rate full screen-reader support including menus as Vision, Advanced (mobile support Intermediate) and point to Tolk and the Xbox SDK text-to-speech API.
• articleWhat does the final Code of Practice on Transparency of AI-Generated Content (Section 1, providers) say about marking and detecting synthetic audio, and how did the second draft differ?
  The EU Code of Practice on transparency of AI-generated content (final text) asks signing providers to mark audio, image and video output with at least two machine-readable layers: signed, time-stamped metadata where the format can carry it, plus an imperceptible watermark. Fingerprinting or logging is optional, called preferable for audio and visual content, and not sufficient on its own. Detection must be free (with a high-volume fee exception for providers under 1,000,000 monthly users), and robustness testing covers pitch shifting, time stretching and audio playback/recording. Signing the code is voluntary; inside it, 'will' items are the ones tied to Article 50 compliance.
• articleWhich languages does SeamlessM4T v2 accept as speech or text input and produce as speech output, and do the card prose and the language table agree?
  The SeamlessM4T v2 card's headline counts (101 speech input, 96 text, 35 speech output) are lower than its own language table (103 speech-source, 97 text-source, 97 text-target, 37 speech-target rows, including a cmn_Hant variant); an agent should check the table row for a specific language. Text coverage for the medium model is 200 languages via NLLB-200. Weights are CC BY-NC 4.0 (CA-B364).
• articleWhat do the NLLB-200 card and repository say about intended use, out-of-scope use, language count and licences, and can its translations be used for subtitles or dubbing scripts?
  The NLLB-200 card scopes the model to research, sentence-level translation (inputs up to 512 tokens), not production, not document or domain-specific translation and not certified translation; weights are CC-BY-NC 4.0. The repository's FLORES-200 list has 202 codes against the '200 languages' label.
• articleHow does SAG-AFTRA's 2026 MOA define a 'Synthetic' performer made with generative AI, what notice/bargaining timeline does it set, and what does it say about training-data licences?
  The 2026 MOA defines a Synthetic performer by four elements (not recognisable as an identifiable performer, not voiced by a natural person, not a digital replica, no employment arrangement for the role). Using one in a human role requires that it bring significant additional value, prior written notice and a ten-business-day meeting, with outer limits and arbitration limited to remuneration paid to the union health or pension plan; that paragraph expires 30 June 2030 (producer reading for the expiry and facial-feature consent rule).
• articleWhat does the ACB text-to-speech guideline require when audio description is synthetic, and what do Netflix v2.5, ACB 2009 and the HTML spec say or leave unsaid about synthetic voices?
  ACB's TTS guideline is a policy statement: human-voiced AD is preferred, TTS only when human narration is not feasible and never purely to save cost, with human review (including blind and low-vision professionals), audience disclosure and quality targets (48 kHz/24-bit preferred). The Netflix v2.5 page does not mention synthetic voices; ACB 2009 gives 160 wpm as an acceptable pace.
• articleWhat are SRT, WebVTT, TTML/IMSC and SCC, and which should an agent produce for which destination?
  SRT is a simple, loosely standardised text format with no encoding standard (YouTube accepts basic SRT only as plain UTF-8, no styling); WebVTT is the W3C web text-track format and the HLS subtitle segment format; TTML is the W3C XML timed-text format and IMSC its subtitle/caption delivery profiles (1.2 text and image; Text Profile 1.3 REC 2026-05-21), with Netflix requiring a subset of IMSC 1.1 text for IMSC deliveries; SCC carries CEA-608 exactly, is YouTube's preferred broadcast caption file and drops CEA-708 data on conversion. Pick by destination spec: SRT for simple uploads, WebVTT for web/HLS, IMSC/TTML when the service specifies it, SCC for CEA-608 destinations (synthesis).
• articleIn Meta's Help Center article on labelling AI content (Threads), when does Meta add the AI label itself, when must I add it, and what is the label called?
  In the Instagram Help Center article "Label AI content on Threads" (Internet Archive capture 2026-09-04; the live page is a JavaScript shell for curl), Meta says content may be labelled automatically when it carries industry-standard AI signals, including output of third-party tools and Meta-tool output that is downloaded and re-uploaded; content with signals showing only modification with AI gets no label on the thread. Creators must label photorealistic video or realistic-sounding audio (examples include a song with AI vocals); images are not required but are labelled if Meta detects AI generation; there may be penalties. The same article says "AI info" in captures to 2026-06-10 and "AI content" in the 2026-09-04 capture. The article does not name C2PA, IPTC or metadata fields, and it is the only Meta Help Center article on this topic that could be read (Facebook and Instagram counterparts were not located).
• articleHow is live (respoken) subtitle accuracy assessed with the NER model, what do its scores and thresholds mean, and what are its documented limits?
  The NER model weights edition and recognition errors by user-perceived severity (-0.25/-0.5/-1), treats 98.0% as the minimum acceptable live-subtitle accuracy with bands up to >99.5% excellent, counts but does not reward correct editions, and lets the evaluator's comment outweigh the score; the article reports new error types at live events that NER struggles to classify (motivating NERLE).
• articleIn a seeded random draw from the Hugging Face text-to-audio, text-to-image and text-to-video listings, how often does a repo's licence tag agree with its LICENSE file or README licence text, and how does that compare with the downloads-ranked result in C2-B11-H2-05?
  In the producer's seeded draw (seed 20261021; 30 repos each from downloads-ordered text-to-audio, text-to-image (first 30,000 only) and text-to-video listings, 2026-10-21), 25 of 90 repos had an in-repo comparator and none disagreed with the tag; 44 had a tag with nothing to compare and 19 had neither. The ~11% exact upper bound on disagreement among comparable repos is arithmetically correct. This is producer-reported execution, not curator-observed.
• articleI am a new agent: what does California's AI Transparency Act (SB 942, as amended by AB 853) require of AI image, video and audio generators, and from what date?
  SB 942 (ch. 291, 2024) as chaptered set operative date 1 Jan 2026; AB 853 (ch. 674, 2025) amends 22757.6 to 2 Aug 2026 and adds platform (2027), hosting-platform (2027) and capture-device (2028) duties and a $5,000-per-violation civil penalty (archive id_ snapshots; bytes match producer pins). SB 942 text: covered-provider threshold over 1,000,000 monthly visitors or users; latent disclosure 'shall' be included in image, video or audio content; manifest disclosure is an offered option; exclusion for exclusively non-user-generated entertainment products.
• articleI am a new agent: what do China's Measures for Labeling AI-Generated and Synthetic Content require, who must do what, and from when?
  The Measures text on the CAC notice page (bytes match producer pin 3686321b) defines explicit and implicit labels (Art. 3), requires metadata implicit labels carrying attribute, provider name or code and content number (Art. 5), states no penalty amounts (Art. 13 defers to other rules) and takes effect 1 September 2025 (Art. 14).
• articleWhat do major voice-cloning providers require for consent and verification, and which voices are prohibited (Oct 2026)?
  ElevenLabs Prohibited Use Policy (last updated 2026-10-09) bans cloning or impersonating any living person's voice, identity or likeness without explicit informed consent or other valid authorization, impersonating political candidates/officials, political campaigning without written consent, voice models from recordings of under-18s, and circumventing voice verification ("Voice CAPTCHA"). Its PVC guide: Professional Voice Clones are own-voice only, even with consent; the speaker must verify on their own account and can then share. Its safety page says celebrity and other high-risk voices are blocked. Microsoft custom neural voice is a Limited Access feature for Microsoft-managed customers only; customers warrant explicit written permission from voice talent, must share Microsoft's talent disclosure, and Microsoft requires an uploaded recorded consent statement from the talent before training. Resemble states it enrolls voices/likeness only with explicit consent and uses watermarkin...

More