evanrudefx Posted June 5, 2020 Share Posted June 5, 2020 (edited) hello, I am having a problem, but it almost seems more like a glitch. I have an attribute called "distance" on the curve. I am deleting points based on distance. The point's value that is circled is 0.087. I say to delete anything greater then 0.087, but for some reason that point gets deleted too. It's acting like I said equal to or greater than. I am also including a photo from my actual scene. If I set the value to 0.088 instead of 0.087 some how it works. Any help is welcomed (hip at bottom) curve.hiplc Edited June 5, 2020 by ejr32123 Quote Link to comment Share on other sites More sharing options...
Aizatulin Posted June 5, 2020 Share Posted June 5, 2020 Hi, If you turn on full precision in spreadsheet, you can see the distance is slightly greater than 0.087, but if you add a small value on your threshold it should work (something like 1e-6). Quote Link to comment Share on other sites More sharing options...
evanrudefx Posted June 5, 2020 Author Share Posted June 5, 2020 2 minutes ago, Aizatulin said: Hi, If you turn on full precision in spreadsheet, you can see the distance is slightly greater than 0.087, but if you add a small value on your threshold it should work (something like 1e-6). ooh, thanks, I didnt know I had to turn on full precision Quote Link to comment Share on other sites More sharing options...
evanrudefx Posted June 5, 2020 Author Share Posted June 5, 2020 (edited) @Aizatulin it still doesnt make sense because in my main scene the distance value is getting the cut distance from my attribute wrangle. Why would it be slightly higher is the channels are linked? Edited June 5, 2020 by ejr32123 Quote Link to comment Share on other sites More sharing options...
Aizatulin Posted June 5, 2020 Share Posted June 5, 2020 Usually these are effects of numerical inaccuracy. Even if your input value is 0.087, the calculated value can be slightly greater or less, and the floating representation for 0.087 is not exactly 0.087. In your case, the difference seems ok. It is around 1e-8, which is near the the float machine epsilon. Check my example, there is a similar effect. clip_0087.hipnc 1 Quote Link to comment Share on other sites More sharing options...
evanrudefx Posted June 5, 2020 Author Share Posted June 5, 2020 ahh, I see. Thanks Quote Link to comment Share on other sites More sharing options...
m-egger Posted June 6, 2020 Share Posted June 6, 2020 https://jurajtomori.wordpress.com/2018/11/24/precision-in-houdini/ In case you didn't run across this already with google, this is a super interesting read by a colleague of mine, @Juraj. He did a lot of testing in this area for his diploma project while working with fractals. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.