You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Very fancy tools! It works great. I grab the IQ to upload in a completely different generator harware and it works great! Unfortunately it's a big IQ file... but there is no way to reduce it. The code is very powerfull I dig it!
I just have a quick question, I would like to add an offset in the IQ, as the DC component generates a big line during the RF frame. I think I just have to add a exp(-2pishift*t) terms in the FFT ... But is there a way to do it straight forward? For example in the processing of the array?
For now the workaround I have it to leave black pixel on the left of the image up to the middle in order to unsure that the DC component won't interfere with my image. But it's actually the same thing I can do in coding... But the drawback is that the IQ will remains bigger in bigger.
So my idea was to post process the IQ to keep the same length but just offset by some frequency to allows the user to program the shift.
What do you think
Thanks,
Tuxgvf
The text was updated successfully, but these errors were encountered:
Hi Polygon,
Very fancy tools! It works great. I grab the IQ to upload in a completely different generator harware and it works great! Unfortunately it's a big IQ file... but there is no way to reduce it. The code is very powerfull I dig it!
I just have a quick question, I would like to add an offset in the IQ, as the DC component generates a big line during the RF frame. I think I just have to add a exp(-2pishift*t) terms in the FFT ... But is there a way to do it straight forward? For example in the processing of the array?
For now the workaround I have it to leave black pixel on the left of the image up to the middle in order to unsure that the DC component won't interfere with my image. But it's actually the same thing I can do in coding... But the drawback is that the IQ will remains bigger in bigger.
So my idea was to post process the IQ to keep the same length but just offset by some frequency to allows the user to program the shift.
What do you think
Thanks,
Tuxgvf
The text was updated successfully, but these errors were encountered: