strange coll behavior
Aug 12 2009 | 3:09 pm
hello,
seem to be getting some strange coll behavior. i insert this list via the coll editor ui:
0, centre 40 lo 35.5 hi 44.7;
1, centre 50 lo 44.7 hi 56.2;
2, centre 63 lo 56.2 hi 70.8;
3, centre 80 lo 70.8 hi 89.1;
4, centre 100 lo 89.1 hi 112;
5, centre 125 lo 112 hi 141;
6, centre 160 lo 141 hi 178;
7, centre 200 lo 178 hi 224;
8, centre 250 lo 224 hi 282;
9, centre 315 lo 282 hi 355;
10, centre 400 lo 355 hi 447;
11, centre 500 lo 447 hi 562;
12, centre 630 lo 562 hi 708;
13, centre 800 lo 708 hi 891;
14, centre 1000 lo 891 hi 1122;
15, centre 1250 lo 1122 hi 1413;
16, centre 1600 lo 1413 hi 1778;
17, centre 2000 lo 1778 hi 2239;
once i close the ui, and then reopen i get these values:
0, centre 40 lo 35.5 hi 44.700001;
1, centre 50 lo 44.700001 hi 56.200001;
2, centre 63 lo 56.200001 hi 70.800003;
3, centre 80 lo 70.800003 hi 89.099998;
4, centre 100 lo 89.099998 hi 112;
5, centre 125 lo 112 hi 141;
6, centre 160 lo 141 hi 178;
7, centre 200 lo 178 hi 224;
8, centre 250 lo 224 hi 282;
9, centre 315 lo 282 hi 355;
10, centre 400 lo 355 hi 447;
11, centre 500 lo 447 hi 562;
12, centre 630 lo 562 hi 708;
13, centre 800 lo 708 hi 891;
14, centre 1000 lo 891 hi 1122;
15, centre 1250 lo 1122 hi 1413;
16, centre 1600 lo 1413 hi 1778;
17, centre 2000 lo 1778 hi 2239;
why is coll adding small decimal values to the input? seems to only occur with floats...
thanks. j
- Aug 12 2009 | 3:19 pma patch to illustrate the problem:
- Aug 13 2009 | 8:54 amIt's a matter of faith. You simply have to believe that 44.700001 and 44.7 are equal. If you don't believe it, have a look at this thread._ johan
- Aug 13 2009 | 9:04 am°44.700001 LITTLE GREEN MEN °°°°°°° °°°°°°°IN THE BACKYARD OF COLL? DONT WORRY ... [ ]YOU CAN KILL THEM !!!!WITH ELEPHANT POWDER.-110.
- Aug 13 2009 | 10:21 amthanks, so its float precision that's causing the issue... btw, weird this happens in max 5 and not max 4.j
- Aug 13 2009 | 11:53 amjustin wrote on Thu, 13 August 2009 12:21thanks, so its float precision that's causing the issue... btw, weird this happens in max 5 and not max 4.Oh, it happens in Max 4, all right. It's just you're not seeing the difference displayed.With floats, what you see is not always what you get. Just very very very very close.
- Aug 13 2009 | 12:07 pmPeter Castine wrote on Thu, 13 August 2009 12:53 Oh, it happens in Max 4, all right. It's just you're not seeing the difference displayed.With floats, what you see is not always what you get. Just very very very very close.that's what i assumed was happening, but why change things from max4 to max5? i dont really see the advantage of seeing the "inaccuracy"...maybe there is (or should be) an option to turn this on or off in max5.
- Aug 13 2009 | 4:56 pm"Float Display Correction" under preferences- default is "on".Eli
- Aug 13 2009 | 5:11 pmthanks!
- Aug 13 2009 | 11:05 pmjustin wrote on Thu, 13 August 2009 14:07 that's what i assumed was happening, but why change things from max4 to max5? i dont really see the advantage of seeing the "inaccuracy"...1. elephant powder.2. normally there is option in the maxmsp menu where you can choose wether you want to see the real values or not. it is just that for an unknown reason this sometimes stop to work. at least in my max4 i see this from time to time. usually to be solved by a patch restart.-110
