From c07d0c2d2f6f7b0eb6e92cc6204bf05037957e82 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?J=C3=B6rg=20Frings-F=C3=BCrst?= Date: Mon, 1 Sep 2014 15:43:52 +0200 Subject: Imported Upstream version 1.6.3 --- target/ofps.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'target/ofps.c') diff --git a/target/ofps.c b/target/ofps.c index fdea8cb..e485035 100644 --- a/target/ofps.c +++ b/target/ofps.c @@ -32,7 +32,7 @@ Some profiles are too rough, and slow/stall vertex placement. Reducing the cache grid and/or smoothing the rspl values - ay mitigate this to some degree, and make this more robust ?? + may mitigate this to some degree, and make this more robust ?? */ @@ -45,7 +45,7 @@ We then initially add sampling points at the largest estimated error verticies of the veronoi natural neighbourhood. - This gives us an optimal distribution measuring in mestimated position + This gives us an optimal distribution measuring in estimated position error within a tollerance of 2:1 We then iteratively improve the distribution of point nodes by @@ -84,7 +84,7 @@ all possible device channel values. This also applies at higher dimensions (ie. the CMYK values exploring response to different K values doesn't spread the CMY values - evenly appart.) + evenly apart.) Stratification seems to be somewhat at odds with the primary goal of minimizing the maximum estimated error from any point in the -- cgit v1.2.3