As mentioned in the title..I keep hearing about this thing...'Dithering'.

What be this critter?

Eddie Bazil from samplecraze.com gave me the best explanation other than stuff that has been shared by bitflipper and Drewfx.

Eddie writes: "Dither is used when you need to reduce the number of bits. The best example, and one that is commonly used, is when dithering down from 24 bits to 16 bits or 16 bits down to 8 etc...Most commonly, dithering from a higher bit resolution to a lower one takes place when a project you are working on needs to be bounced down from 24 bits to 16 bits using dithering algorithms.
So, what is the process; in mortal language of course?

A very basic explanation is we add random noise to the waveform when we dither, to remove noise. When we truncate the bits (lowering the bit resolution), ie in this case we cut down the least significant bits, and the fact that we are always left with the stepped like waveforms in the digital process, by adding noise we create a more evenly flowing waveform instead of the stepped like waveform. It sounds crazy, but the noise we add results in the dithered waveform having a lower noise floor. This waveform, with the noise, is then filtered at the output stage. I could go into this in a much deeper context using graphs and diagrams and talking about probability density functions(PDF) and resultant square waves and bias of quantisation towards one bit over another. But you don’t need to know that now. What you do need to know is that dither is used when lowering the bit resolution and that this is an algorithmic process, ie using a pre determined set of mathematical formulae.

If we take the 24 bit project scenario and select to bounce the resultant audio without dithering then the last eight bits (also know as Least Significant Bits) of every 24-bit sample are discarded. In terms of audio integrity you will not only lose resolution but also introduce Quantisation Noise.


Because dithering adds random noise to the lower eight bits of the 24 bit signal whilst maintaining stereo separation the quantisation noise is dramatically reduced.

It then makes sense to dither from 24 bits to 16 bits rather than bounce without it.

How well the process is executed is down to how good the dithering algorithms are. But to be honest these algorithms are so good nowadays that even standard audio sequencing suites (Cubase, Logic etc) will perform dithering tasks without much problem.
My recommendation is to always work in 24 bit and dither down to 16 bit for the resultant file, as CD format is still 16 bits."

-Danny

Thanks for the information!!

I'd just add that it's only recommended to dither once. i.e. reduce bit depth only once. I keep all my stuff at a minimum of 24 bits until bounce to CD time.

The short technical answer is this:

When you quantize a signal to a set number of values/steps, for each sample there's a difference between the actual value and the quantized value known as quantization error. The more bits you have, the smaller the quantization error.

Generally speaking quantization error sounds like a weird, nasty distortion (assuming you can hear it all).

But if you add just enough random noise (this is the dither) before quantizing or reducing bit depth, it changes the weird nasty distortion to sound like noise, which is much less annoying (again assuming you can hear it at all).

Dither also allows signals smaller than the smallest bit value to be preserved, and some of this signal is audible below the noise floor.

Different types of dither just refer to the nature of the dither noise and use tricks to try to make the dither noise less audible while still doing its job.

i'm hoping ...and thinking ..... that Sonar does this automatically when I export my project with 24 bit recordings to the CD level of 16 bit...... 

Otherwise.... I'm screwed. 

I lost Danny right after " A very basic explanation is ...." 

Guitarhacker


i'm hoping ...and thinking ..... that Sonar does this automatically when I export my project with 24 bit recordings to the CD level of 16 bit...... 

Otherwise.... I'm screwed. 

I lost Danny right after " A very basic explanation is ...." 

Probably is the answer. Most of the export presets usually have one of the dithering presets selected but it is something a user can change if they wish.

FastBikerBoy

Guitarhacker


i'm hoping ...and thinking ..... that Sonar does this automatically when I export my project with 24 bit recordings to the CD level of 16 bit...... 

Otherwise.... I'm screwed. 

I lost Danny right after " A very basic explanation is ...." 

Probably is the answer. Most of the export presets usually have one of the dithering presets selected but it is something a user can change if they wish.

I think I saw this in the export/bounce window.....actually pretty sure about that..... but I never bother to change the settings from factory on that dithering thing and everything works fine. 


I actually did do some research on this early on, because I did wanna get it right.... and the default settings are good all around general purpose settings designed to work fine. 


I'm really not as clueless as that post above might tend to indicate......   close.... but still....... oh never mind.... 

red book standard.
that's why.