returning absolute values

stefan · 24 Mar 2005

This checks if the value is negative and if it is we make it positive:

float val; val = -1.0; val = ((val <= 0.0) & -val) + ((val > 0.0) & val)

galaa · 24 Mar 2005

When I put this into the code component within SM the colouring stops halfway through. Does this mean that there is a problem with the syntax or will it still work?

stefan · 24 Mar 2005

Sorry, I thought the code module could handle negation of values, but it didn't. Should've tested my code before posting it, which I seldomly do. heh.

This ought to work:

float val; val = -1.0; val = ((val <= 0.0) & (0.0 - val)) + ((val > 0.0) & val)

stefan · 24 Mar 2005

malc wrote: Or did I?

[EDIT] is this thread getting confusing?

You should've never deleted your reply.

To let people know, this is what happened:

While I was writing my reply to galaa, malc also wrote a reply. He posted his before me. So after I posted mine, I replied to malcs and said "damn you, you beat me to it". When I had posted that reply, malc had removed his reply. So I removed mine, meanwhile, malc replied to my "beat me to it"-post (which I then had removed) and said "Or did I?"

And... Here we are.

edit: ok, malc removed his post....

malc · 24 Mar 2005

Like a soap opera 😉

I will refrain from deleting posts in future.

galaa · 24 Mar 2005

Cheers for that stefan. You've helped me make my first dsp component, its a waveshaper that produces pretty dark,moody overdrive and then once you get past a certain point it goes into major distortion.

polyin in; polyin drive; polyout out;

float k; k=(2*drive)/(1-drive);

float val; val=in; val = ((val <= 0.0) & (0.0 - val)) + ((val > 0.0) & val);

out=(1+k)*in/(1+k*val);

the value for drive should be in the range of 0-0.99.

Sorry if this a failry obvious piece of code but its my first so please be gentle

stefan · 24 Mar 2005

You don't need to use the "float val" btw, you could work directly on the in variable instead.

About "obvious piece of code". Most my own stuff is, imo, very basic. I prefer short simple code snippets / modules that are easily reused instead of big ones solving way too specific problems.

galaa · 24 Mar 2005

Thanks, I'll sort that now. Like you say, the shorter the better

malc · 24 Mar 2005

Great stuff 🙂

jorge · 24 Mar 2005

stefancrs wrote: You don't need to use the "float val" btw, you could work directly on the in variable instead.

That's true but if you assign straight to the "in" variable you'll be affecting any other module it is connected to. If the output connector linked to the In connector of the code component has only one link it should be no problem. I'd probably leave it using the "var" variable eventhough performance will be slightly affected.

galaa · 24 Mar 2005

Just thought, if I work on the 'in' value directly, how would this line work?

out=(1+k)*in/(1+k*val);

The 'in' for this line is not absolute so it could be negative, where as if i worked on 'in' directly then would all instances of 'in' become absolute?

stefan · 24 Mar 2005

@jorge&galaa: Oh. Yeah. Right! My bad 🙂