Sources and licenses for the /wedding-video-audio/ samples
Prepared November 3, 2026. Each "original" is built from the sources below; each "restored" file is Diffio 4.6
Pro's restored speech from the Diffio API on November 3, 2026, with no background mixed in. Both files in a pair
have one fixed gain so they play at about the same loudness. No other processing. Details: manifest.json.
None of the people or organizations below are Diffio customers or endorse Diffio.

wedding-01 (wedding-01-original.mp3, wedding-01-restored.mp3)
Caption: Sample made by playing clean speech (LibriTTS, CC BY 4.0) through a real measured lecture-hall impulse
response (BUT Speech@FIT Reverb Database, CC BY 4.0) and mixing it with a real recording of about 100 guests at a
wedding reception ("Crowd at a British wedding reception venue" by Pandos, Freesound, CC0). Restored with Diffio 4.6
Pro. Both clips are loudness-matched for playback.

1. Speech: LibriTTS test-clean, speaker 1188, utterances 1188_133604_000024_000000 to _000001
   LibriTTS (Zen et al., 2019), from a public domain LibriVox reading. https://www.openslr.org/60/
   License: CC BY 4.0, https://creativecommons.org/licenses/by/4.0/
   Changes: sentences joined with a short silence, convolved with the impulse response, mixed with the crowd.
2. Room impulse response: BUT Speech@FIT Reverb Database, room E112 (lecture room, about 900 cubic meters)
   Brno University of Technology, Speech@FIT (Szöke et al., 2019).
   https://speech.fit.vut.cz/software/but-speech-fit-reverb-database
   License: CC BY 4.0, https://creativecommons.org/licenses/by/4.0/
   Changes: truncated to 1 s with a 50 ms fade-out and aligned to the direct-path peak.
3. Crowd: "Crowd at a British wedding reception venue" by Pandos, Freesound.
   https://freesound.org/people/Pandos/sounds/362353/
   License: CC0 1.0, https://creativecommons.org/publicdomain/zero/1.0/ (credited as a courtesy)

wedding-02 (wedding-02-original.mp3, wedding-02-restored.mp3)
Caption: Sample made by mixing clean speech (LibriTTS, CC BY 4.0; softened above 5 kHz) with real recordings of an
outdoor crowd ("Expectant Outdoor Crowd" by CHallSmith, Freesound, CC0) and a windy beach
("VeryWindyRocksBeach.wav" by shelbyshark, Freesound, CC0). Restored with Diffio 4.6 Pro. Both clips are
loudness-matched for playback.
The automated listening check confirmed the crowd but not the wind, so the page claims crowd noise only.

1. Speech: LibriTTS test-clean, speaker 8555, utterances 8555_284449_000028_000000 to _000001
   LibriTTS (Zen et al., 2019), from a public domain LibriVox reading. https://www.openslr.org/60/
   License: CC BY 4.0, https://creativecommons.org/licenses/by/4.0/
   Changes: sentences joined with a short silence, low-passed at 5 kHz, mixed with the crowd and wind.
2. Crowd: "Expectant Outdoor Crowd" by CHallSmith, Freesound.
   https://freesound.org/people/CHallSmith/sounds/870619/
   License: CC0 1.0, https://creativecommons.org/publicdomain/zero/1.0/ (credited as a courtesy)
3. Wind: "VeryWindyRocksBeach.wav" by shelbyshark, Freesound.
   https://freesound.org/people/shelbyshark/sounds/476504/
   License: CC0 1.0, https://creativecommons.org/publicdomain/zero/1.0/ (credited as a courtesy)