[x265] x265: uncommon behavior by changing the 8-point DCT matrix

[Paulo André Oliveira]

Dear x265 development team,

I am trying to conduct the following experiment: assess the change in the
compressed video's quality by changing only the 8-point DCT matrix, which I
suppose is the constant 'g_t8' at source/Lib/TLibCommon/TComRom.cpp

However, the video's quality, which I am monitoring by the PSNR and SSIM
metrics, keeps the same with any random matrix that I define in 'g_t8'. I
am using the last version of x265 as of today.

Sincerely,

Paulo Oliveira
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20140527/92abaada/attachment.html>