Paul Swirhun (@swirhun)
Ok. Now that I read it more carefully, I think AC_RND_CONV does exactly that: each successive 0.5-quantum point gets rounded in the opposite direction. Thanks...
I haven't seen VHDL used (in the USA), but I mostly get it. So when truncating N bits off, you round up if X[N-1] is 1 and X[N:0]/(2**(N-1)) is greater than or...
Hi Kaz,I'm definitely open to your suggestions. All you said was "round it" though. Even with rounding, you get truncation error, but it's just half the magnitude...
Anyway, thanks for the guidance on this. I guess the figure is correct as-is, and we just have to deal with the clipping (or wider output word) for pathological...
I extended all time periods by 100x. They both have the same qualtative behavior of trending towards zero. The average signal values in the "stepped down"...
Ah, that makes sense, so you're just referring to the exponential trend towards zero. Yes, I can lengthen those times and check it out, but I think my spreadsheet-based...
I decided to derive the transfer functions to see if that helps me understand it at all. Sorry if the forum spams you a few times while I edit these Latex equations....
To be more specific, I don't understand your point. There is a much larger DC offset in the original implementation. I've annotated it with an arrow similar to how...
I'm not sure I follow that reasoning. If you look at the filter as drawn in the textbook ("Screenshot 2018-08-17 at 11.34.15 PM.png") there is a gigantic DC offset...
As a DSP beginner implementing a CIC filter recently, I wanted to add something to this. I know this is a very old post, but I'll add to it anyway.I found Rick's...
Thanks Rick; no rush at all. Sorry those images didn't work so well; I now have fuller access to the forums so I am trying to attach a better image now. Blue is...
Use this form to contact swirhun
Before you can contact a member of the *Related Sites:
- You must be logged in (register here)
- You must confirm you email address
