Re: [SDRSharp] Make it understandble PLEASE!


Alberto I2PHD
 

On 8/15/2013 8:03 AM, Alan wrote:

I have seen references of places to look for learning the basics, perhaps someone will repeat them for those who want to learn?

I could suggest the following books, the first of which is also available as a free download in PDF format.

http://tinyurl.com/jwqcfhf��� available also as free download at� http://www.dspguide.com

http://tinyurl.com/kynjne4��� a must-have. A book written in a tutorial way, easy to understand
�������������������������������������� but full of precious information.

http://tinyurl.com/mqy8537�� also a classic. A bit dated, but the algorithms aren't. The only pity
��������������������������������������� are the many typographic errors, especially in mathematical formulas.



On 8/15/2013 4:52 AM, Akos Czermann wrote:

Will you please face your life and accept that you do know more / have more money than the rest of the world
I would like that the underlined sentence were true, but unfortunately it isn't...�� :-)
I appreciate your effort writing a few hundred words, but next time write what you want to say,
then revise / rewrite on a level understandable to your non-radio neighbour.�
I am afraid that explaining in a tutorial way how an SDR works would take much more space than that
allowed to a forum message.� After all, the Gerald Youngblood's series of articles on the subject took
four issues of QEX...�


On 8/15/2013 4:35 AM, alan_r_cam wrote:
With all due respect, Alberto-
You have described one way of performing an SDR. There are others.
Surely. I described the most common way of implementing an SDR. Presently there are only a few deviations from
this architecture, the most notable being the latest line from FlexRadio, the Flex-6000 Signature Series, with
the Flex-6700 model (6700 being also a lower limit of its price in USD...� :-) ).
But note that all the signal processing algorithms are NOT performed in an FPGA, which, with its shortage of
multipliers and RAM, is the least desirable place where to perform FFTs. Instead that radio sports a Texas Instruments
powerful DSP chip, that takes the role of the PC, leaving to this latter only the mundane tasks of showing
a spectrum and a waterfall, and that of implementing the GUI.

The concept of FFT fast-convolution filtering is in fact mentioned in "A Software Defined Radio For The Masses: Part 3"
by Gerald Youngblood, AC5OG. He takes a small section of signal and translates it to baseband, effectively frequency shifting it.
�
You can do a frequency shift using an FFT (albeit there are disadvantages in this method), but not with a fast convolution.
The fast convolution algorithm is used for filtering, where it can replace a very long FIR, using just a fraction
of the processing power otherwise needed for it.

--
73 Alberto I2PHD

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