Re: [SDRSharp] Re:: Added bladeRF support


jdow <jdow@...>
 

I think God may have had it right in Genesis when he muttered about "all the days of your life" regarding snakes crawling on their bellies, women suffering painful childbirth, and people using DVB-T dongles and expecting more from them than they can give.

There is a reason for this. First the analog RF filtering in the dongles is relatively low Q if it exists at all and it is very wide to get reasonable flatness across a TV bandwidth. Second the digital filters used are very "soft". You can see this on the fully zoomed out waterfall display. Note that the edges have lower amplitude than the center. I believe but cannot prove that the DVB dongles sample at a much higher rate than they can push out the USB port. Downsampling requires decimation filters. And all evidence suggests they use quite weak decimation filters. So the aliasing effects are a "feature". There is little or nothing Youssef can do to remove them.

That said, Youssef elected to use somewhat softer filtering in his decimators than he could have, basically in the interest of CPU usage and to a degree in the interest of reducing latency. (The audio comes out of SDR designs running on Windows, at least, a LONG time after audio comes out of the equivalent analog radios when tuned to the same frequency. One second delays are not unusual. This is definitely true with DVB-T dongles. It may or may not be true with other tools. Nor am I sure the transport medium, USB vs Ethernet, makes a big difference or not.)

{o.o} Joanne

On 2014-08-03 09:13, scott@scan-ne.net [SDRSharp] wrote:


Working very nicely overall. I'm still trying to reduce the amount of false
imaging I'm receiving, but aside from that it seems to be quite stable at this
point.

What are the Auto (1dB) and Auto (3dB) options in the filter bank? Trying them
seems to make no difference here.

Scott

Join airspy@groups.io to automatically receive all group messages.