Credits and sources

The photographs behind the site and the open datasets our benchmark uses, with who owns each one and what we did to it.

Photographs

Photographs: owner, licence and what we changed
Photograph	Owner	Licence	What we changed
Home page, beside the headline	Jimmy Art Devier	Unsplash License	Cropped, blurred, recoloured and grain added.
The orange strip	Marek Piwnicki	Unsplash License	Cropped, blurred, recoloured and grain added.
The closing section on each page	Sixteen Miles Out	Unsplash License	Cropped, blurred, recoloured and grain added.

The grain, the charts and the dot mark are our own.

Datasets: owner, licence and what we changed
Dataset	Owner	Licence	What we changed
Crowdsourced UK and Ireland English Dialect speech (OpenSLR 83)	Google, with Cardiff University for the Welsh recordings	CC BY-SA 4.0	Not rehosted. Our script rebuilds the clips we list as 8 kHz phone audio with street and kitchen noise, on your machine.
EdAcc: The Edinburgh International Accents of English Corpus	Centre for Speech Technology Research, University of Edinburgh	CC BY-SA 4.0	Not rehosted. Our script rebuilds the clips we list as 8 kHz phone audio with street and kitchen noise, on your machine.
Cekura Turn-Taking Benchmark, public set (v1)	Cekura	CC BY 4.0	Nothing. We link to it, with credit.

Read from each dataset’s own page, 26 October 2026. How we use them

If you own any of these and want the credit changed or the item taken down, write to greg@gme.solutions and we will fix it.