DF96 and I see it differently than you I guess. By his comment filtering off the high frequency components of a square wave is harmonic distortion.
I'm saying errors from windowing can get to the -130db level or better if you try hard enough. Dwelling on the fact that the transforms are defined by integrals that go to infinity is a device often used by folks trying to discredit all digital signal processing.
For a longer sample set (e.g, the movement of a symphony), the effect of windowing drops out of sight. Also true if I sample Ed's flute for several milliseconds before and after the note. I'm not sure why people are trying to make this more complex than it actually is.
Andy Conners pointed out to me that a brilliant illustration of what Dave, Scott, and I are saying may be found in figure 7.8 of Antoniou's "Digital Signal Processing: Signals, Systems and Filters."
Andy Conners pointed out to me that a brilliant illustration of what Dave, Scott, and I are saying may be found in figure 7.8 of Antoniou's "Digital Signal Processing: Signals, Systems and Filters."
Filtering using analog linear filters? No. Filtering digitally sampled signal using algorithms? Easy.
Here you go.
An externally hosted image should be here but it was not working when we last tested it.
So, if somebody insists on non-linear distortions, we have to go either back to properties of capacitors and resistors, or ... back to school
So, if somebody insists on non-linear distortions, we have to go either back to properties of capacitors and resistors, or ... back to school
Time for some horse meat sashimi, it should be very tender by now?
Or chicken sushi (reference: "Dr. Jekyll and Mr. Hyde, Together Again")
I hear that there is a spa that propagates turkeys brought over from Admiral Perry so clean they are OK as sashimi.
I ate some of this in a specialist horsemeat restaurant in Matsumoto in Japan a few years ago. Very thinly sliced- delicious. I've never ever told my wife babout it ecause she is a vegetarian on moral grounds. I never get a steak at home - that's something I do when we go out only.
Well don't go to China and sample the dog. You will be in the doghouse for sure.
In Sendai the great delicacy was the mane fat, I did not get it.
Well don't go to China and sample the dog. You will be in the doghouse for sure.
In Sendai the great delicacy was the mane fat, I did not get it.
When business men took me to dinner in China - I ordered something made with pork. They all strongly suggested I dont eat anything with pork in it. Was a danger to my health... I could get deathly ill from some desease it seemed. So, maybe dog is better than pig? They didnt say anything bad about dogs.
But I did eat from a huge plate of french fried mealy worms and didnt get deathly ill.
Last edited:
Scott
We are almost there.
I started by mentioning the laughter about music being considered a continuous waveform.
Now I think it is clear to most it is not. As a result parts may be considered to have infinite bandwidth or a fundamental with all the harmonics
You agree that if we use an RC network to low pass filter this we have a result that is the opposite of what we would normally call harmonic distortion.
However we are talking about a CR coupling network. So what we get is response that is a high pass filter. Below the breakpoint the fundamental is attenuated and the harmonics will have a rising response. That will certainly sound like harmonic distortion
So the anecdoral experience to use much larger coupling capacitors than required by bandwidth considerations follows the math.
If you know the lowest frequency component you can calculate the attenuation and therefore the apparent change in the harmonic content ratio.
- Status
- Not open for further replies.