|
Propeller Head Plaza Technical and scientific discussion of amps, cables and other topics. |
|
|
|
Propeller Head Plaza Technical and scientific discussion of amps, cables and other topics. |
|
In Reply to: May not be undesirable posted by Dave Kingsland on December 3, 2006 at 16:02:58:
> Some time back, another poster here implemented an FFT-based CD
> upsampling approach ...[]... and burned the resulting 24/176.4 track
> to a DVD-A. Or something like that anyway.If he is sampling at 176kHz then there will be almost no energy to be filtered above the Nyquist frequency due to low energy in the music here plus the low pass filtering of the microphones. Consequently, a perfect brick wall filter in frequency space will have both little to operate on and the resulting ringing is going to be almost wholly at completely inaudible frequencies.
> Obviously, it's not something that can be done in real time with
> current technology,It is fairly cheap and easy to implement with current signal processing hardware (so long as you can program the hardware). Also a quick look at benchmarks for modern DFTs (e.g. www.fftw.org) will show that for a stereo audio stream performing the DFTs themselves on a sequence of time windows is not a problem. One would use a sequence of time windows because the length of the samples only determines the resolution at the low frequencies and resolving at sub-10Hz is not necessary for music. In practise the windows would probably overlap raising the required sampling rate a bit.
> BTW, I think it's questionable that reducing the spread of the
> ringing is desirable.Why? If there is silence before a transient in the original and audible ringing before the transient after filtering then this would seem rather undesirable to me. The question would seem to revolve around whether the ringing is audible. The ear/brain extracts a lot of information from transients but is relatively insensitive to phase information at high frequencies. I cannot say which will win without looking it up, writing a 10-20 line script to find out or posting a message on a computer audio group. Perhaps one could find out via google but information like this on the web risks input from those with audiophile beliefs.
> In a properly band-limited square wave, the ringing is supposed to
> be there.Indeed. Viewing the signal and its processing in frequency space is convenient in many ways but some aspects can be open to misinterpretation for the unfamiliar. There also seems to be a tendency to view filters as the solution to the problem and, therefore, perfect filters are the perfect solution. Of course, the actual problem is not how to filter but how to sample an audio signal and regenerate that audio signal without introducing audible artifacts.
This post is made possible by the generous support of people like you and our sponsors:
Follow Ups
- Re: May not be undesirable - andy19191 04:08:31 12/05/06 (10)
- I agree DVD-A 176KHz sampling does the trick. - cheap-Jack 11:54:26 12/05/06 (2)
- Re: I agree DVD-A 176KHz sampling does the trick. - Todd Krieger 00:33:20 12/06/06 (1)
- Steep brickwall filters are need for specific applications, - cheap-Jack 09:12:55 12/06/06 (0)
- Re: May not be undesirable - Dave Kingsland 11:05:30 12/05/06 (6)
- Re: May not be undesirable - Todd Krieger 00:17:11 12/06/06 (5)
- Re: May not be undesirable - Dave Kingsland 09:09:28 12/06/06 (4)
- Re: May not be undesirable - Todd Krieger 08:59:28 12/07/06 (3)
- Re: May not be undesirable - Dave Kingsland 09:44:49 12/07/06 (2)
- Re: May not be undesirable - Todd Krieger 00:09:23 12/08/06 (1)
- Re: May not be undesirable - Dave Kingsland 09:06:44 12/08/06 (0)
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
