2009-07-30 15:02:47 UTC
.bmp's:
I have been using the image preparation tools to convert my .bmps to .rgb and then to .yuv and vice-versa to view the locally-decoded file, per the README file. I know this works because I convert the resulting .yuv back to .rgb and back to .bmp so that I can see the .yuv that is being given to the dirac_encoder.
-num_L1:
I tried setting "-num_L1 0", this did set the dirac encoder to "intra". Thankyou.
But the problem I have still persisted.
DAUB9_7:
I tried changing the iwlt transform from DAUB9_7 to one of the others, I chose "DD9_7".
This did give an improvement. My "lossless" coding looks OK now - it did not before.
But there still are visible problems.
I have a few images that I am trying currently. Some photo's, graphics and test patterns.
Eg One frame of colour bars, top half white-to-black, bottom half in reverse, black-to-white.
I will continue and try some other wavelet transforms to see if any of those suit my test material.
There is certainly a bug with DAUB9_7.
