[FFmpeg-devel] libavcodec/exr : fix float to uint16 conversion for negative value

[Martin Vignali]

Hello,

In attach a patch who fix this sample :
https://we.tl/HWgF0SFEll

This sample contain negative value in the red channel.

Before this patch, negative red float are converted to max red
instead of zero.

The previous hack doesn't seems to work for float to uint
so i add an explicit sign check

but i'm not familiar with float binary manipulation,
so maybe there is a better way to fix the problem

Comments welcome

Martin
Jokyo Images
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0002-libavcodec-exr-fix-float-to-uint16-conversion-for-ne.patch
Type: application/octet-stream
Size: 1097 bytes
Desc: not available
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20170423/6d937d5d/attachment.obj>