PocketMatrix

Hey,<br><br>Tutorial 5 is online. This time I explain a bit about fixed point math, using an example rotozoomer (wich I also explain).<br>If someone could proofread it, I will be happy to make some adjustments based on your comments.<br><br>- Jacco.

<br><br>Looks good. You may also want to mention that to add / subtract fixed point values they must be of the same precision, and that the result stays in that precision (unlike mult and div). Also, you may mention that multiplying / dividing a fixed point value by an integer (a "real" integer, not one that represents a fixed point value) doesn't result in any change in precision either.<br>Here are a couple formulas for determining the resulting precision of fixed point mult / div:<br><br>r1.p1 * r2.p2 = r1+r2-32.p1+p2<br>r1.p1 / r2.p2 = r1+32-r2.p1-p2<br><br>Where r=Range and p=Precision. Of course once you know either the range or precision of the result you can calculate the other by subtracting from 32.<br><br>Also, it may be good if you summarize with something like:<br>Multiplying two fixed point values has a result with less range but more precision. Dividing two fixed point values results in a value with less precision and more range.<br><br>I think your tutorial is going to bring back PQ induced fixed-point-conversion nightmares. <br><br>Dan East

I've added most of your comments. Tnx.

erm back to tutorial 1, I can't seem to get it to compile. I get about 10 errors when linking, something to do with gx methods. <br>I think I have it looking in the right place for all the gx libraries and headers, but not sure about the dll.<br>anyone any ideas? I can't get the full error messages right now as i'm on a different machine.<br>thanks.

Make sure you import gx.lib into the Workspace

cheers, i'm not doing that at the moment, i'll give it a try.