Power module question

(Apologies if already asked/answered.)

When I run a non-clipping audio signal through a Power module and Y (exponent) is set to 1, 2, 3, etc. all is well.

But when Y is set to a fractional value like .5, 1.2, 4.7 all hell breaks loose! Specifically, Power outputs a NaN signal which causes (in the example above) the audio output to be maxed out even if the output slider (top right of Reaktor window) is set to 0.

Why?

My guess: Problems performing the floating-point math at the audio rate ... ?

Thanks for your help!

rachMiel

 

a lot of these values are literally not a number, hence the NaN errors.

for instance, check the negative range in this graph:

https://www.wolframalpha.com/input/?i=y+=x^0.5

 

Why would straightforward exponential math yield values that were not numbers? And is there a fix so that I can submit decimal values for Y?

Thanks sala! Good to hear from you, been quite a while.

 

the reason being that, x^0.5 = sqrt(x). and if x is negative, there is no square root of a negative number if you think about it.

to fix, make sure x is always positive by using the absolute function. then you can set the output to be negative afterwards if you need a bipolar output.

yeah, not around these parts too much anymore. to be honest i'm a bit disheartened with the state of development in reaktor and spending my free time working on video games these days.

 

Thanks a lot. I've something like that in my mind. But not so exaclty you describes.

 

And yet sending a -1...1 audio signal through a Square Root module works fine, no NAN. The Power module must be using a different algorithm to do the exponential math.

I tried four ways to avoid inputting negative numbers to the Power module:

1. Rectify them |x| as you suggest.
2. Convert the -1...1 input signal to 0...1.
3 + 4. For both 1 and 2 you can either keep the 0...1 output or convert it back to -1...1.

Each of these four sounds different!

If you're okay with sharing, I'd like to hear why you're disheartened.

What are you doing with video games? I teach UI design in the Gaming department of a nearby university.

 

this post kind of sums it up.

https://www.native-instruments.com/forum/threads/6-3-released.355009/page-9#post-1781387

i'm pretty sure that this post earned me a 'warning' on my account if you can believe that... the next time i logged in, it wouldn't let me do anything until i checked a box saying i understood the terms of service for the the forums. it also happens to have received more 'likes' than any comment i've ever posted here.

such a passive-aggressive response to a rather milquetoast critique irked me more than it should have, probably.

certain features that i've been asking for repeatedly since 2009 are no closer to being implemented now than they were then. let's just say that i'm not waiting another 10 years hoping that they'll listen sometime between now and 2030. i'm moving on in the near future, although i still have work connected to reaktor i won't be committing any free time to it anymore.

for game development, i'm working on an NES game with a woman from sweden. i think i e-mailed it to you a long time ago. it's nearly done now.

 

you´re kidding...

 

I had the same checkbox. I just assumed it had been sent to everyone.

 

i also assumed that, until i spoke to somebody else on the forum he said that hadn't happened to him. maybe he missed it, but looks like mark didn't get one either.

 

I had the same right after the 6.3 heat. If that's the way how they handle criticism...

 

Same here. I think it could have been done in relation to the general bad feeling over the update, but not necessarily SA's specific post.

 

sad for Reaktor world, but I hope less frustration and more fun for you .... Might be connected, I find this forum has been quite dull after the burst of 6.3 release :-/
I had the validation dialog too, so might be global...

 

Yes, got it also. It is an advantage rule, IMO.

 

honestly the forum issue is more like the cherry on top of everything else. feeling like you're being threatened with a ban for sharing your opinion in a way that was intended to be constructive is not great, and certainly if they're going to send such a message they should at least let the person know why. but the real frustration is that reaktor development is not moving in a direction that i'd like, and there is no indication that it will.

we need updates that address the real shortcomings of reaktor as a development environment. where is iteration in core? where is a proper way to declare a function? loops and functions are pretty much the two most basic features a programming language can have, and reaktor can't do either. i'm not interested in any update that doesn't include these features.

 

that's normal :


and have a look into the core macro x(>0)^y:

 

heh, NONE of those are what i mean!

i think this is the equation you want to use:

https://www.wolframalpha.com/input/?i=y+=+sqrt(abs(x))+*+sign(x)

 

i used rachmiel's description
Nevertheless rachmiel's treat-start was going to the wrong direction.

 

my fault for being vague when there is a very precise way to talk about such things

 

Gotcha. I was also quite passionately involved in the push to get the Reaktor UI optimized for builders way back then. I submitted a list of desiderati from the builder's pov. At that time I was still building lots of Reaktor thingies ... and because I teach UI design I felt I was in an ideal position to help the NI Reaktor developers out. But they pretty much shelved everything I said and took the blocks route instead. Being a non-fan of the blocks approach, I promptly lost interest, similarly to how you did.