Sign in to follow this  
Followers 0
the_mad_mirror

question about anti-aliasing filter

5 posts in this topic

what causes the high end frequency boost which occures when an antialiasing filter is applied?

Bit more info please? Real time plug-in or destructive DSP on the file?

And also what are the orig and new sample rates?

In theory the high frequency audio can get pretty nasty and squarish at the Nyquist threshold and so smoothing filters are often applied to get rid of them. Maybe the plug-in is set to curve the high frequencies in such a way as to smooth the threshold, but in return boost certain frequencies just below it.

Share this post


Link to post
Share on other sites

according to a friend I was chattin to it occures will all antialiasing filters.  when the top end is rolled off a barely noticable bump occures up around 20k.  he didnt know what caused it but assured me it was definitly there with any antialiasing filter.  has something to do with the rate of roll off, he said that its why they make sampling rates above 44.1k so that they could have a more gradual roll off and reduce this boost effect.  i am wondering what causes it. havin trouble finding info on it on the net. 

I think ur idea about a boosted band spilling below the cutoff sounds like a pretty good possible explanation

Share this post


Link to post
Share on other sites

Ah ok, the more general issue of nyquist in normal sampling...

Your mate is correct, at bang on the nyquist frequency (exactly half the sample rate), if a hard brick wall filter cut it off then a distortion of the highest frequency occurs, given there are so few samples to capture the wave cycle up there. If you try converting a 20Hz to 11.025kHz sine sweep from 44.1kHz to 22.05kHz in any decent editor, you'll likely find that the resultant wave is anything but sine at certain points, especially the top end. Tried it myself once for curiosity sake, was pretty obvious.

By raising the sample rate to more than 2 x desired Nyquist, the filters can be more gentle and therefore more smooth (and cheap). At 44.1kHz, the nyquist is 22.05kHz - higher than human hearing. So if any distortion did occur, theoretically we wouldn't hear it anyway.

that leads us to oversampling, which improves that scenario again, but it's late and I have to go to bed :(

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now
Sign in to follow this  
Followers 0