Help - I've broken my Sinusoidal!! Dither it up please ?

Discussion about: tracktion.com
RELATED
PRODUCTS

Post

I like triangles, but they make my teeth go all funny.
I've joined Lurkers Anonymous.

Post

Lunch Money wrote:Let's say that it takes 8 bits to recreate a perfect sine wave. I mean, a perfect sine wave.
It doesn't. To reproduce a *perfect* sine wave would require an infinte number of bits, and an infinite number of samples per duty cycle.

:P
Someone shot the food. Remember: don't shoot food!

Post

boogooloo1 wrote:I like triangles, but they make my teeth go all funny.
It's the nougat.
Someone shot the food. Remember: don't shoot food!

Post

Lunch Money wrote:Now, the reason I'm not arguing against the difference is that with music, we are NEVER talking about just a set of perfect triangle, square, or sine waves, so it's a moot point.
Actually with music we are always talking about sets of perfect sine waves, any note (whatever its waveform) consists of a fundamental (which is a sine wave) plus a blend of harmonics (which are also sine waves), so a triangle wave is actually a summing of sine waves (as are all other wave shapes) :)

Post

My head hurts.
I've joined Lurkers Anonymous.

Post

valley wrote:
Lunch Money wrote:Let's say that it takes 8 bits to recreate a perfect sine wave. I mean, a perfect sine wave.
It doesn't. To reproduce a *perfect* sine wave would require an infinte number of bits, and an infinite number of samples per duty cycle.

:P
No, you're confusing two things. A visual representation of a sine wave with pixels, would require an infinite number of pixels. People often mistake the pixels in their waveform views for the actual wave itself. They're simply not perfectly analogous.

Let's say you have a sound generator that just needs to be told a few pieces of information: rate, depth, and er... whatever else a sine wave is composed of. That information can be told to the sound generator which will then produce a perfect sine wave.

I'm not saying that's the way the technology currently works. I'm just arguing that theoretically, a perfect sine wave can be recreated in fewer bits than "infinite".

Frankly and honestly, I don't have enough knowledge about how it all works, so I can only talk in theoreticals. :D

On that note, saying that all sound is a combination of perfect sine waves, triangle waves, saw waves, all amalgamated together and influenced/shaped by harmonic content is the same as what I've already said. ;) We don't deal with pure sine waves, triangle waves, and square waves anyhow.

;)

Greg
Image

Post

Lunch Money wrote: No, you're confusing two things.
Trust me Greg, I'm not.
A visual representation of a sine wave with pixels, would require an infinite number of pixels. People often mistake the pixels in their waveform views for the actual wave itself. They're simply not perfectly analogous.
No, but the pixels *do* tell you something. They tell you that the sine wave is *not* perfect. A sine-wave must contain one frequency, and one frequency only. Any distortion, no matter how minor, introduces new spectral content. At that point you cease to have a sine wave. It may sound like a *perfect* sine wave, but it isn't . if you listened to it in an AB with a pure sine, you'd likely be able to tell the two apart as well.

Basically you simply need to understand one fact about sound: abrupt changes in amplitude are caused by very high frequencies. A square wave has a (theoretically) zero attach to the the wave peak. The only way you can creaete such a sound is to have a hell of a lot of high frequency content in the sound. In the case of a square, technically there are in infinite number of upper harmonics.

If you pass a square through a low-pass, you will see that the edges become more rounded as the cut-off approaches the wave's fundamental.

Knowing that, you only need to look at the waveform of a sampled sine to see that the steps represent abrupt increases in amplitude. For a decent bit-depth and sample-rate the effect will be audibly negligible, but it's there, and as you reduce bit-depth and sample-rate it becomes more obvious.
I'm not saying that's the way the technology currently works. I'm just arguing that theoretically, a perfect sine wave can be recreated in fewer bits than "infinite".
A sine wave is circular, and like any circle, it is basically a polygon with an infinte number of vertices. To register the true continuous nature of a sine wave you simply need to capture the change at each vertice - all n->infinity of them. ;)
On that note, saying that all sound is a combination of perfect sine waves, triangle waves, saw waves, all amalgamated together and influenced/shaped by harmonic content is the same as what I've already said. ;) We don't deal with pure sine waves, triangle waves, and square waves anyhow.
No we don't, which is why digital systems simply aspire to provide a decent enough approximation that our senses are convinced of the illusion.

If you go back and read my original post, you'll notice I was specifically drawing attention to your usage of the word 'perfect.' The principles of what you say are valid, but 'perfect' is simply inaccurate.
Someone shot the food. Remember: don't shoot food!

Post

valley wrote:
On that note, saying that all sound is a combination of perfect sine waves, triangle waves, saw waves, all amalgamated together and influenced/shaped by harmonic content is the same as what I've already said. ;) We don't deal with pure sine waves, triangle waves, and square waves anyhow.
No we don't, which is why digital systems simply aspire to provide a decent enough approximation that our senses are convinced of the illusion.

If you go back and read my original post, you'll notice I was specifically drawing attention to your usage of the word 'perfect.' The principles of what you say are valid, but 'perfect' is simply inaccurate.
AHAHAHAHHHAHAAAAHAHAHAAA!!! People are still writing on this? Classic.
I'm gonna rename the thread to "silly silly theory discussion with bread thingies". BTW lossless audio compression techniques like monkey audio and lossy techniques such as mp3/ogg/whatever use multi-sine representation. *Whip-crack sound* NOW BACK TO WORK!!! Write some friggin music damn-your-eyes!!
m@

Post

metamorphosis wrote: AHAHAHAHHHAHAAAAHAHAHAAA!!! People are still writing on this? Classic.
Wow. Like WTF. OMG! The two minutes I spent typing that post like totally wiped out my entire day. Like damn!.

:roll:

It's called a conversation mate. Some people just enjoy discussing things with other people and learning from each other.

That's why this is called a 'forum'. See how that works?

If I'm here keeping an eye on this forum then from time to time I'll reply to threads if they interest me.
I'm gonna rename the thread to "silly silly theory discussion with bread thingies".
If it makes you happy. :shrug:

[qoute] BTW lossless audio compression techniques like monkey audio and lossy techniques such as mp3/ogg/whatever use multi-sine representation. [/quote]

Fascinating! Please do tell me more, I'm intrigued.

:hihi:
Someone shot the food. Remember: don't shoot food!

Post

Fair enough that it's not really working the way I was describing. I can't argue, being devoid of that knowledge.

;)

Greg
Image

Post

valley wrote: Fascinating! Please do tell me more, I'm intrigued.
No offence was intended, mr angry. I just find it amusing, is all...
m@t

Post Reply

Return to “Tracktion”