Noise Shaping

  • Thread starter Thread starter thajeremy
  • Start date Start date
thajeremy

thajeremy

registered abuser
Ok...i have read a few threads that talk about noise shaping....where is this adjusted in CEP?
 
When you dither down your mixdown right at the end of the whole process, you click on Convert Sample Type, right? The popup that appears has options for Noise Shaping. Not exactly a Great Big Issue though IMO.
 
i have heard the term "dither" a million and one times but not quite sure what it means...is it just the conversion from 96000 to 44100??
 
More or less, yeah. That, plus the conversion from whatever your bitdepth was when you were mixing, to 16-bit.

But dithering is a calculation as well, or rather a number of calculations - there are algorithms and math that go into dithering so that the digital chunks of information don't sound ugly - the math 'smooths off the jagged bits at the end of each chunk' of digital information so that it sounds smooth and not ugly.
 
well..i havent been converting to 44.1 with cooledit. I just left it at 96000 and my CD burning software converts it before buring...would i have better results by using cooledit to convert?
 
Dithering is not for changing sample rates (96k to 44.1k) it is for changing bit depth (24bit to 16bit). It is the process of adding low level noise to give you the illusion that you are still listening at the higher bit depth. The noise shaping options are just a few different types of dithering noise to choose from. You need really good ears to pick up on this and the type of music will dictate which one is better (or if it matters at all)
 
"Dithering is not for changing sample rates"

Really! I never knew that. So when you change the sample rate, it's just a straightforward conversion and there's no possibility of nasty artifacts?
 
There is a posibility for artifacts but dithering has nothing to do with it. There are different conversion algorythms that all sound different. The best way I have found to go from a high sample rate to a low one is to use a high quality d/a converter hooked to a a high quality a/d converter at the new sample rate.
 
For all the difference there is between 48,000 Hz and 44,100 Hz, I choose to record at 44.1K and avoid the conversion. If everything moves up to higher resolutions, I might have to make a tougher decision.

But if I had the gear to track at 88,200 Hz, then the conversion back down to CD quality would be more or less free of dangerous math, correct?
 
From what I understand, the conversion is not as simple as removing every other sample. So it still might contain dangerous math, it also may not be the biggest deal in the world. A good deal of the listening public will be listening to the mp3 version anyway. The conversion from 88.2k to 44.1k isn't the conversion that bugs me.

P.S. If my memory serves... 48k was come up with because it synced with video (it was a multiple of video frame rate or something) The difference in the nyquist frequency (22.05k to 24k) I don't think represents a half step.
The difference in the nyquist frequency between 44.1k and 88.2 is a full octave.
 
Where can I get material about niquist theory, bit-depths, sample rates etc? I think my knowledge is very superficial.

Any suggestions? Would prefer online...
 
You must log in or register to reply here.
Back
Top