DC offset removal through FPGA
in [DSP]
|
Prev: SHARC debug
Next: Who Invented the Z Transform
From: Steve Underwood on 29 Jun 2005 22:05 robert bristow-johnson wrote: >in article d9uc20$3s8$1(a)nnews.pacific.net.hk, Steve Underwood at >steveu(a)dis.org wrote on 06/29/2005 10:43: > > >>How is "phase linear" of relevance in a signal so narrowly focussed >>around DC? I find DC rather hard to phase shift. >> >> > >there's the transition band. it's funny, on one hand there are people >saying we can't hear phase at all (see other thread about it) and those who >say that the phase non-linearity of a DC blocking HPF (hell, we can put the >cutoff at 0.1 Hz, the filter will be pretty flattened out by 20 Hz). both >extremes are extreme. > > That is why I said "so narrowly focussed". If you can notice the transition band you haven't done it right. The design process should be something like: Filter will be z = z + (x - z)*A A = 0.01 while (you can notice any phase or amplitude disturbance in the band of interest) { A = A/2 if (A is so small that (x - z)*A drops off the end of the register) increase the register length a bit } 20Hz isn't really purist. I think he most extended bass pedal on organs is 16.something Hz. So, with the corner at 0.1Hz (which is actually fairly high for such a filter) we are over 7 octaves below that bass pedal.That's maybe 45dB. I somehow doubt a signal 45dB down is going to make a significant phase or amplitude difference. If it does, use the above design procedure. So, like I said, how is "phase linear" of relevance? Regards, Steve
From: robert bristow-johnson on 30 Jun 2005 00:32 in article d9vk1n$rlt$1(a)home.itg.ti.com, Steve Underwood at steveu(a)dis.org wrote on 06/29/2005 22:05: > So, like I said, how is "phase linear" of relevance? i agree. i really just wanted to make clear that, although it's true that is pretty hard to shift the phase of DC, it's less hard to shift the phase in the transition band. i think an even symmetry FIR DC removal filter is not necessary (it would have to be awful long), but it is a way to make sure one is phase linear the whole way. i still think it might be interesting to test a DC removal filter that LPFs to get DC and then before subtracting, median filtering the DC and then maybe LPFing that as well before subtracting it from the input. dunno. it's just an idea i had when considering cases where subtracting DC is better than an HPF. -- r b-j rbj(a)audioimagination.com "Imagination is more important than knowledge."
From: Jerry Avins on 30 Jun 2005 11:22 robert bristow-johnson wrote: > in article d9vk1n$rlt$1(a)home.itg.ti.com, Steve Underwood at steveu(a)dis.org > wrote on 06/29/2005 22:05: > > >>So, like I said, how is "phase linear" of relevance? > > > i agree. i really just wanted to make clear that, although it's true that > is pretty hard to shift the phase of DC, it's less hard to shift the phase > in the transition band. i think an even symmetry FIR DC removal filter is > not necessary (it would have to be awful long), but it is a way to make sure > one is phase linear the whole way. > > i still think it might be interesting to test a DC removal filter that LPFs > to get DC and then before subtracting, median filtering the DC and then > maybe LPFing that as well before subtracting it from the input. dunno. > it's just an idea i had when considering cases where subtracting DC is > better than an HPF. If the subtracted signal is to be contemporaneous, the delay has to be large, and the result of the subtraction _IS_ a high-passed version. Because the signal to be subtracted changes so slowly, a delayed version is good enough. Then the subtraction method shines: good (enough) DC removal with little or no delay. Jerry -- Engineering is the art of making what you want from things you can get. ýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýý
From: Andor on 1 Jul 2005 03:48 Robert wrote: > i still think it might be interesting to test a DC removal filter that LPFs > to get DC and then before subtracting, median filtering the DC and then > maybe LPFing that as well before subtracting it from the input. I had a look at median filters the other day. One of their properties is that for monotone signals, the output is equal to the input with a delay of N/2, where N is the frame length of the median filter. For example, a step signal is unaffected by this filter. As very low frequency signals are not likely to jump around wildly, the median filter will not have any effect most of the time. When it does affect the signal, I think the distortion is audible because of the low frequencies involved. By the way, a sliding median filter with frame length N can be implemented using a O(log N) algorithm - that is, there are better ways than the Quick Median algorithm (which basically is just one half of a Quick Sort) which has O(N log N). I took me while to figure out how :-). OTH, adaptive median filters (ie. only replace an output sample by the median of the frame if the difference exceeds a certain threshold) might make some sense for removing single sample "outliers" from an audio signal. Regards, Andor
From: Rick Lyons on 3 Jul 2005 21:24
On Wed, 29 Jun 2005 22:43:12 +0800, Steve Underwood <steveu(a)dis.org> wrote: >Rick Lyons wrote: >> On Sun, 26 Jun 2005 09:55:58 -0400, Jerry Avins <jya(a)ieee.org> wrote: >> >> >> >>>Why is direct superior to going around the other side of the barn? If >>>the final result is the same, the simplest approach is usually to be >>>preferred. Are there hidden assumptions we don't share? >>> >>>Jerry >> >> >> Hi, >> This thread interests me because I don't know any >> simple way to remove the DC portion of a signal in real >> time. It seems to me that any digital IIR filter >> method would induce some sort of unpleasant phase >> nonlinearity. >> >> It sounds like the "subtract the output of a >> super-narrow (linear phase) lowpass filter from the >> original signal" method should work. >> I wonder if it does. > >How is "phase linear" of relevance in a signal so narrowly focussed >around DC? I find DC rather hard to phase shift. > >Regards, >Steve Hi, I was just thinkin' about a super narrowband linear-phase lowpass filter centered at zero Hz. That's all. See Ya', [-Rick-] |