How does Visual C++ compare to old school C++?

[foozball3000]

In my box of many ambitious projects, is the dream to create a white noise screensaver. If I plan on using Visual C#, I need to use Unsafe Code. So, it's a plausable route.

The other option will be to use C++. But that's where my question comes in.. How does Visual C++ compare to old school C++?

 

[greggpb]

It's the same thing.. visual c++ just adds some template's and libaries for interfacing with the OS.

It also depend's what you are going to use for the screen saver if you are gonna use opengl or direct x ? I would stick to visual c++ Ide as you will be able to do things easeir in windows without having to understand all the in's and outs.

 

[greggpb]

C++ or C# ?

 

[foozball3000]

It's the same thing.. visual c++ just adds some template's and libaries for interfacing with the OS.

It also depend's what you are going to use for the screen saver if you are gonna use opengl or direct x ? I would stick to visual c++ Ide as you will be able to do things easeir in windows without having to understand all the in's and outs.

I'm planning on rendering the white noise. (Hardcore!) But I still have no clue on how to do it. My original idea of having a bitmap in the memory was way too slow.

C++ or C# ?

I know all about C#, so I'm exploring the possibility of using C++

 

[dequadin]

If you are worried about portability of your code, you should stick to the standard library.

MS has a lot of "Windows Only" C++ functions. You can still use Visual C++ to code either way...

I assume the hardest part is generating the white noise source? I'm sure there are a couple of algorithms out there that can generate "pretty close to the real thing" Gaussian noise.

You could use GDI+ and C#, I don't see why you have to use unsafe code. Direct3D is also an option...

 

[sn3rd]

White noise? You should look into the Marsaglia Bray algorithm. It's not all that well documented (I've only ever found one proper paper on the topic), but it is easy enough to understand and implement.

In the limit, it produces a very good approximation to additive white Gaussian noise.

You'll need a fairly good uniform pseudorandom number generator to implement this, as the quality of the AWGN generator will depend on the quality of the uniform PRNG.

 

[dequadin]

OR he could sample some random UHF frequency

 

[sn3rd]

OR he could sample some random UHF frequency

True...

Or on Linux, he could use /dev/random

Edit:
Actually, not true...

He can't use "some random UHF frequency". He needs to use ALL frequencies, as that's what "white" noise is.

 

[dequadin]

True...

Or on Linux, he could use /dev/random

Edit:
Actually, not true...

He can't use "some random UHF frequency". He needs to use ALL frequencies, as that's what "white" noise is.

Yes But I assume foozball3000 is going for that TV look, so taking the noise on one frequency component of the spectrum is a good approximation (surely?) Then you could alter the frequency you're sampling and get a cool effect, like tuning an old-school TV...

Hmm if linux has /dev/random, does boost have something similar? Anyone know?

 

[sn3rd]

Yes But I assume foozball3000 is going for that TV look, so taking the noise on one frequency component of the spectrum is a good approximation (surely?) Then you could alter the frequency you're sampling and get a cool effect, like tuning an old-school TV...

Hmm if linux has /dev/random, does boost have something similar? Anyone know?

Well, as you know, simulating true white noise is impossible; it would require an infinite bandwidth. Thus we're always limited to approximating it.

I think it all depends how he wants to display the "white" noise. If he wants the "TV with really bad signal" then it'd be enough to sample the spectrum in the lowest band possible for TV. If he wants the equaliser-type "white" noise display, then it'd be enough to sample in any range of choice and display the Fourier transform...

I love geeking out

 

[dequadin]

Well, as you know, simulating true white noise is impossible; it would require an infinite bandwidth. Thus we're always limited to approximating it.

I think it all depends how he wants to display the "white" noise. If he wants the "TV with really bad signal" then it'd be enough to sample the spectrum in the lowest band possible for TV. If he wants the equaliser-type "white" noise display, then it'd be enough to sample in any range of choice and display the Fourier transform...

I love geeking out

/me should have spent less time drinking and more time paying attention during class

EDIT: I thinkg this is really the way to go. Showing natural white noise, is waaaay cooler than simulating it through code

 

[sn3rd]

/me should have spent less time drinking and more time paying attention during class

Or combined the two?

(I think we've achieved successful derailment)

 

[Necuno]

what about black noise ?

 

[sn3rd]

what about black noise ?

vvvvvvvvvvvvvvvvvvvvvvvv



^^^^^^^^^^^^^^^^^^
There you go...

 

[dequadin]

(I think we've achieved successful derailment)

I'm not so sure this is a derailment, foozball3000 likes overly ambitious projects, he just might consider it

 

[sn3rd]

I'm not so sure this is a derailment, foozball3000 likes overly ambitious projects, he just might consider it

I'm half expecting a "tell me more about how I could attempt to achieve an infinite bandwidth..."

 

[ic]

In my box of many ambitious projects, is the dream to create a white noise screensaver. If I plan on using Visual C#, I need to use Unsafe Code. So, it's a plausable route.

The other option will be to use C++. But that's where my question comes in.. How does Visual C++ compare to old school C++?

My advice to you is this, if you already know C# well enough, then get a working prototype going using unmanaged C# code, and when you start refining your working prototype seriously consider migrating to C++ for performance and flexibility.

Visual C++ does introduce some Microsoft-specific language extensions, but I seriously doubt that you need to worry about them at this stage, for the most part VC++ is C++. If you're considering porting to other OSes then your problem will be API related and not really C++.

what about black noise ?

Noise has a colour? Julius Mampoera is always able to make a lot of noise and should be just the thing you're looking for.

 

[foozball3000]

/starts downloading new vocabulary upgrade... :wtf:

Thanks for the great responses. Now I have a lot of research to do.

Google

But I'll first go and look at the Marsaglia Bray algorithm. Then see if I can do the first version in C#. That way, because a lot of this is new to me, I can focus on how it should work and not how to code it.

 

[dequadin]

Now I have a lot of research to do.

What happened with this?

 

[sn3rd]

/starts downloading new vocabulary upgrade... :wtf:

Thanks for the great responses. Now I have a lot of research to do.

Google

But I'll first go and look at the Marsaglia Bray algorithm. Then see if I can do the first version in C#. That way, because a lot of this is new to me, I can focus on how it should work and not how to code it.

Once you're generating random numbers, you should test them. Download and install GNU Octave (or use Matlab if you have access to it). You can plot a histogram of, say, 1 000 000 random numbers and it should give you something that looks like this with a mean of 0 and a standard deviation of 1.