479 reputation
11022
bio website pelletencate.nl
location Netherlands
age 30
visits member for 3 years, 11 months
seen Feb 27 at 9:06

My name is Pelle ten Cate, I am a software engineer from Amersfoort, Netherlands, and I am also a graduated professional recording engineer.

I studied at the Royal Conservatoire in The Hague, and I am currently working for a Chicago-based tech startup called Occasion (occsn.com).


Sep
28
comment Are there any mathematical advantages for higher sampling rates?
I +1'd that. I should not be describing that as a drawback, because it's not. I should however say that both quantisation noise and unshaped dithering noise are audible on a 44/16 mix. My point is that by switching to 24 bits you give yourself the opportunity to change one issue (quantisation noise) for another (dithering noise) that can effectively be diminished by recording in a higher sampling rate.
Sep
28
comment Are there any mathematical advantages for higher sampling rates?
Re first comment: You're absolutely right. I think what I meant to say is that when you're using a digital effect, you can expect the frequency range of its output to be taken care of. Put it this way, if the output comes out aliased in the first place, upping your own sample frequency isn't going to change that. As to your second comment: interesting; it totally depends on the filters used pre-downsampling. If the noise was imaged back, it would obviously mask away the quantisation noise, but it wouldn't sound exactly the same. I think I would shape my noise around the final nyquist freq.
Dec
11
comment Are there any mathematical advantages for higher sampling rates?
While this is not 'untrue', working with 24 bits introduces the drawback that you have to use dithering if you want to go back to 16 bits. Dithering noise can be reduced hugely if applied to a 96 KHz signal when using noise shaping. (see other answer for details)
Dec
10
comment Why doesn't an FFT pitch filter fully 'mute' the selected pitch range?
Normal filters can't basically, since you always deal with 'filter components'. FFT filters are different, they can get any shape possible within the frequency and the dynamic range, and they can have any phase characteristics you wish. (That is why convolution reverb is based on FFT as well.)
Dec
10
comment Why doesn't an FFT pitch filter fully 'mute' the selected pitch range?
Not entirely correct, you can actually get à real brickwall filter using FFT, as long as the FFT reaches the top frequency of your sample rate. That is, you can have an FFT filter that filters 5000Hz with 0dB and 4999Hz with -96dB. One of the advantages of a Fourier filter.
Dec
9
comment Why doesn't an FFT pitch filter fully 'mute' the selected pitch range?
Really muting sound below a certain frequency only happens when using a so-called 'brickwall' filter. Are you sure you are using such a type of filter?