Real speech cleanup example
Source speech: LibriSpeech dev-clean, speaker/chapter 2277/149896.
Dataset source: https://www.openslr.org/12
License: Creative Commons Attribution 4.0, https://creativecommons.org/licenses/by/4.0/
Changes: an 8-second excerpt, with synthetic background noise added; the same excerpt was then processed by the site's actual cloud speech cleanup service.
No new paid generation was performed to publish this example.

Speech enhancement example: the same licensed speech dataset, shortened to 3 seconds and given controlled room echo. The enhanced output is from a real prior processing run, reused without another paid call.

Song separation example: an 8-second excerpt of America the Beautiful, Courtney Williams / United States Navy Band. Original recording: United States Navy. Source page: https://commons.wikimedia.org/wiki/File:America_the_Beautiful_(male_vocalist)_-_United_States_Navy_Band.opus
Source page identifies the performance recording, composition and lyrics as public domain (PD-USGov-Navy / Public Domain Mark). Modified by excerpting, encoding and separating vocals and accompaniment. The current separated tracks are the real output of the approved November 10 comparison run using the identical original excerpt. Accompaniment is the sum of the returned drums, bass and other instrument tracks; no manual cleanup or overdubbing. This short excerpt is vocal-dominant and is not a studio reference-stem benchmark.

Real video cleanup example
Source: NASA Johnson, Karen Nyberg Shows How You Wash Hair in Space (12 July 2013).
Source page: https://commons.wikimedia.org/wiki/File:Karen_Nyberg_Shows_How_You_Wash_Hair_in_Space.webm
Original publication: https://www.youtube.com/watch?v=uIjNfZbUYu8
Rights: NASA-produced recording, marked public domain in the United States on its source page.
NASA media guidelines: https://www.nasa.gov/nasa-brand-center/images-and-media/
Excerpt: 00:08–00:20, 12 seconds, resized to 960 × 540. Original synchronous speech and ambient sound retained, with no added noise, dubbed speech or stock audio.
The processed file is the actual output of this website's Basic noise reduction workflow. Both files retain the same video picture. The illustration frame is extracted from the same recording.
Used as an informational demonstration of audio processing. NASA and the person shown do not endorse or sponsor this site.