Thanks everyone who replied to the original survey! It turns out that I had
made a mistake in the way I generated some images. The results of the
survey were useful, but I was using an inferior technique to generate the
thumbnails. It would be awesome if you can answer the newer, better version
of the survey: https://www.surveymonkey.com/s/F6CGPDJ
This time there are 12 images to rate for sharpness and quality. Each
unique image appears 3 times, but it's generated using a different
technique in each case.
I've switched away from side-by-side comparison because in practice people
wouldn't experience those images side-by-side. And comparing 3-ways might
require too much effort for a quick survey, given that the images are very
similar to each other.
Thanks in advance for your time,
On Fri, May 2, 2014 at 3:43 PM, Gilles Dubuc <gilles(a)wikimedia.org> wrote:
If you have a little spare time, it would be very helpful if you can
answer this survey that takes less than a minute to complete:
It will help determine if a specific flavor of a performance optimization
proposal currently discussed  on the Multimedia mailing list  can be
implemented without impacting image quality in a noticeable way.