<div dir="ltr"><div>Hello,</div><div> </div><div>Just a quick message, I have improved the -l3 lower quality setting.I have increased the quantization to 0.8 and decrease a little the denoising, and so to save 1Ko in average, I have added a 8x8 block variance analysis.Blocks that have a low variance are averaged to save bits.The -l3 lower quality setting is a little better now.</div>
<div> </div><div>For the -h3 higher quality setting, yes for now I do not compress a small data packet (but same compression scheme as for the other residual datas), so file size will be smaller for -h3 setting.Residual coding on higher wavelet orders seems ok now for the higher quality settings.Do not hesitate to give me your opinion.</div>
<div> </div><div>The most difficult is still the lower quality settings and increasing compression (-l3 and below), -l2 could be ok? For these lower settings, I would like to increase the denoising ("edge-preserving" denoising) -to increase the compression- and keep a good sharpness.Would this approach sound acceptable? For now (and for minimal extra computational cost), I have the wavelet coefficients, and also data of the pre-processing function (segmentation kernel), to perform the denoising for the lower quality settings, but it is not that easy...</div>
<div> </div><div>Many thanks,</div><div>Raphael</div></div>