Hi brains trust,

In the past I would use the bounding box expression ($BBY*2-1) * sin a lot in a point sop. what's the equivalent of the expression in vex , for use in vexpressions.

Rob

you can use relbbox() function
for example:
float bby = relbbox(0, @P).y;
@P.x += sin(bby*10);

Thanks, I will have to have a play with that, just to make sure I understand it ! , While I am at it if I wanted $NX = $TX2 - $TX what should I look at in vex. I must say its not easy switching over after relying for so long on hscript expressions @!

Rob

You could do like this:
@N=point(@OpInput2, “P”, @ptnum)-@P;
Assuming you're in an Attribute or Point Wrangle node.

-b

@P = (relbbox(1,@P)*2-1) * sin(@Frame);
or for a more long winded version:

vector pos2 = @opinput1_P; // fetch second input position to variable pos2
@P = (relbbox(1,pos2)*2-1) * sin(@Frame);

Modern method to reference second input is documented in the last paragraph for the help on the Attribute Wrangle VOP.

@opinput0_P fetches first input or just us @P
@opinput1_P fetches second input P
…
There should be no need to use the point() function to fetch second, third and fourth input attributes.

relbbox(0,@P) fetches bbox position from first input
relbbox(1,@P) fetches bbox position from second input
…
All the vex functions that take a file as an argument, you can use the direct integer index to the input of the current SOP.

I find point(@OpInput2, “P”, @ptnum) to be a useful shortcut if the other objects have more points than the first one. Then it's easy to offset @ptnum to access points outside the range of the number of points in the first object.

-b

jeff
@P = (relbbox(1,@P)*2-1) * sin(@Frame);
or for a more long winded version:

vector pos2 = @opinput1_P; // fetch second input position to variable pos2
@P = (relbbox(1,pos2)*2-1) * sin(@Frame);

Modern method to reference second input is documented in the last paragraph for the help on the Attribute Wrangle VOP.

@opinput0_P fetches first input or just us @P
@opinput1_P fetches second input P
…
There should be no need to use the point() function to fetch second, third and fourth input attributes.

relbbox(0,@P) fetches bbox position from first input
relbbox(1,@P) fetches bbox position from second input
…
All the vex functions that take a file as an argument, you can use the direct integer index to the input of the current SOP.

I was expecting the opinput_attrib binding to be much faster than using point/prim functions but unfortunately it's the other way around . Using point/prim functions to get attributes from other inputs is faster. So I wouldn't use the opinput syntax if the performance is critical. Otherwise they would be nicer to use.

pusat
…
I was expecting the opinput_attrib binding to be much faster than using point/prim functions but unfortunately it's the other way around . Using point/prim functions to get attributes from other inputs is faster. So I wouldn't use the opinput syntax if the performance is critical. Otherwise they would be nicer to use.

- in my tests @opinput_ binding is the same or maybe a bit faster than point() for matching by ptnum
- but be careful since if you have id attribute present on your geo, it's binding by id by default, which you can control in Bindings Tab/ Attribute To Match in which case it's speedwise comparable to findattribvalue() workflow

tamte

pusat
…
I was expecting the opinput_attrib binding to be much faster than using point/prim functions but unfortunately it's the other way around . Using point/prim functions to get attributes from other inputs is faster. So I wouldn't use the opinput syntax if the performance is critical. Otherwise they would be nicer to use.

- in my tests @opinput_ binding is the same or maybe a bit faster than point() for matching by ptnum
- but be careful since if you have id attribute present on your geo, it's binding by id by default, which you can control in Bindings Tab/ Attribute To Match in which case it's speedwise comparable to findattribvalue() workflow

Thanks I will check the file I sent to SESI. They also saw the same behaviour. Otherwise I was anxious to update all of my VEX code to use opinput.

Though the recent addition of auto type recognition for the known attributes also prevents me from doing this as I don't want to type prefixes for known types as we don't use the latest Houdini 15 build here.

For H16 hopefully all will be well.