bcd_cli crashes at a specific resolution.

[ykozw]

Hi, I found bug.
you can reproduce with a very simple procedure on Windows 10.

1. create 1920x1147, white pixel(1.0f) image, with 3 samples.
   https://www.dropbox.com/s/ae5mp4d8ute8b7x/bcd.raw?dl=0
2. $> raw2bcd.exe bcd.raw tmp
3. $> bcd_cli --use-cuda 0 -i tmp.exr -h tmp_hist.exr -c tmp_cov.exr -o out.exr

Thanks.

[superboubek]

Hi, Thanks for reporting ! Your input is indeed not valid, and we should indeed check these corner cases. The reason is that with all samples identical, we have a null covariance. Will fix that as soon as we have time. Best, Tamy -- Prof. Dr. Tamy Boubekeur LTCI, Telecom ParisTech, Paris-Saclay University Room C13, IDS Department Telecom ParisTech, 46 rue Barrault, 75013 Paris France tamy.boubekeur@telecom-paristech.fr http://www.telecom-paristech.fr/~boubek Le sam. 18 août 2018 à 09:16, Yuki Ozawa <notifications@github.com> a écrit :

…

Hi, I found bug, It can be reproduced with a very simple procedure. 1. create 1920x1147, white pixel(1.0f) image, with 3 samples. https://www.dropbox.com/s/ae5mp4d8ute8b7x/bcd.raw?dl=0 2. $> raw2bcd.exe bcd.raw tmp 3. $> bcd_cli --use-cuda 0 -i tmp.exr -h tmp_hist.exr -c tmp_cov.exr -o out.exr Thanks. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#7>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AM5QE_WTnTD5LiomWMUb0zVBwuSZTfsGks5uR79agaJpZM4WCdzl> .