This is also a very good resource:
https://linearaudio.nl/sites/linear...ound Card Instrument Manual & FFT windows.pdf
Characterization of 60 different windows, their transfer functions + example apps.
Jan
Last edited:
Huh? Those Germans call it a Hanning window? Grrrr. A good reason to not read that stuff.
More seriously, no window at all (i.e. a rectangular window) gives also rise to leakage (depending on the continuity of the input). Applying windows is just a matter of distributing the leakage in another way over the spectrum. Some windows increase the resolution, while other windows increase the accuracy (of the magnitudes).
The Virtins list talks about a Hamming window, a Bartlett-Hann window and a Hanning window. Different windows.
Jan
#offtopic
I do have problems calling it a Hanning window. According this 'logic' one also should talk about a Hamminging, Kaisering, or Barletting window. Rubbish of course.
His name is von Hann period.
Only an illiterate idiot would call it a Hanning window.
NB: Also Richard W. Hamming calls it a Hann window (in "The Art of Doing Science and Engineering") and he knows what he is talking about.
Cheers,
E.
edit: No! See page 4 of your link in post 1041.
I do have problems calling it a Hanning window. According this 'logic' one also should talk about a Hamminging, Kaisering, or Barletting window. Rubbish of course.
His name is von Hann period.
Only an illiterate idiot would call it a Hanning window.NB: Also Richard W. Hamming calls it a Hann window (in "The Art of Doing Science and Engineering") and he knows what he is talking about.
Cheers,
E.
Last edited:
edit: No! See page 4 of your link in post 1041.
Cheers,
E.
That's just a type of an illiterate idiot. Only one. ;-)
Jan
Attachments
Notch filter compensation
Back to business.
Cheers,
E.
Back to business.
See comment below.
Yes.
Sure.
You mean the discrepancy between THD and coherent plot, don't you? Well, the answer simple (and embarrassing): In the current version, the notch compensation doesn't work yet on the wide band spectra. Although it was on my to-do list, I simply forget to implement this feature (probably because I considered these wide band spectra of minor importance). Anyhow, I will correct this omission (next week I hope).
Cheers,
E.
Notch filter compensation
Sure, it does do something: adapting the residual. But do you mean it also does something with the raw (wide band) spectra? In that case something is wrong, as it should do nothing (currently).
BTW, now I remember why I "forgot" to implement the compensation of the wide band spectra. The difficulty is that the notch characterization file only contains data of the fundamental and its harmonics. So pretty useless for compensation of the these spectra. Although I can solve this by means of interpolation for frequencies above the fundamental, but not below that point. That's the reason I've postponed the compensation for these spectra. A possible solution might be to simulate the response of the notch filter, provided that it is a standard passive one.
Anyhow, for the moment, only the THD spectrum is correctly compensated.
Cheers,
E.
Hi Demian,
Sure, it does do something: adapting the residual. But do you mean it also does something with the raw (wide band) spectra? In that case something is wrong, as it should do nothing (currently).
BTW, now I remember why I "forgot" to implement the compensation of the wide band spectra. The difficulty is that the notch characterization file only contains data of the fundamental and its harmonics. So pretty useless for compensation of the these spectra. Although I can solve this by means of interpolation for frequencies above the fundamental, but not below that point. That's the reason I've postponed the compensation for these spectra. A possible solution might be to simulate the response of the notch filter, provided that it is a standard passive one.
Anyhow, for the moment, only the THD spectrum is correctly compensated.
Cheers,
E.
Having the transfer function of the notch filter in the num/den polynomial format would probably simplify the calculations in DiAna, and the filter definition input would reduce to the polynomial coefficients only
The transfer function can be derived for the filter topology used, and is probably available in the standard texts.
I've got it for my Hall topology filter, and it agrees very well with the measured one.
Just a thought.
Regards,
Braca
The transfer function can be derived for the filter topology used, and is probably available in the standard texts.
I've got it for my Hall topology filter, and it agrees very well with the measured one.
Just a thought.
Regards,
Braca
Hello Edmond,
Yes, this is the essence of the idea.
For the data entry, it would probably be easier to use the rational function form attached below.
For a third-order filter, it would be M = N = 3; and the polynomial coefficients would have to be provided by the user.
Using complex arithmetic in your program, it is then easy to calculate the gain and phase at a given frequency.
Regards,
Braca
Attachments
I have started the tests here:
https://www.diyaudio.com/forums/parts/354641-linearities-cored-inductors.html#post6208395
Having a set of theoretically determined coefficients also provides for great starting values for a nonlinear regression.
Jan do you do anything to assure the null is at a bin center, I have found that you can have several dB's of error especially if the null is exactly between bins. There is an argument to deliberately de-Q a notch to lower the artifacts when the center drifts a little with temp/time.
Drift
If the drift is relative small (a few ppm) the impact on the insertion loss and phase of the harmonics is small and has very little impact on the reconstruction of the residual. As for the fundamental, it depends. A passive notch filter at 1kHz with a notch depth of -60dB, a frequency mismatch of 3Hz will result in a decreased damping of about 1 to1.5dB. Does it matter? Only if you do use this figure to reconstruct the fundamental or if you use it to calculate the relative distortion (WRT the fundamental), otherwise who cares. (sorry for the late response)
Cheers,
E.
Hi Jan,
If the drift is relative small (a few ppm) the impact on the insertion loss and phase of the harmonics is small and has very little impact on the reconstruction of the residual. As for the fundamental, it depends. A passive notch filter at 1kHz with a notch depth of -60dB, a frequency mismatch of 3Hz will result in a decreased damping of about 1 to1.5dB. Does it matter? Only if you do use this figure to reconstruct the fundamental or if you use it to calculate the relative distortion (WRT the fundamental), otherwise who cares. (sorry for the late response)
Cheers,
E.