[Lowfer] Re: Observations while copying WOLF

[Alberto di Bene]

> Regarding Dick's comment about the SDR and WOLF on the same computer, I 
> can only pass along the observation that WOLF and some of his friends 
> are touchy when multitasking. But I thought that the bulk of the SDR 
> processing was being done on the sound card, and that the CPU load would 
> be reasonable. I wonder if the computer also had other tasks like email 
> running...

No, the sound card doesn't do *any* processing, apart from performing the analog-to-digital conversion. And this is 
valid for all the SDR programs around. That is a common misconception. Some sound cards have a so-called DSP chip, but 
that acronym means Digital *Sound* Processor and not Digital Signal Processor.

100% of the processing done on the input signal, after the conversion to the digital realm, is done exclusively by the 
main CPU, be it an AMD or Intel.

Using an SDR, the wanted portion of the spectrum is brought to baseband by first buffering the digital samples, then 
applying a quadrature mixing, possibly followed by a resampling. Care must be exercised to not having any buffer 
underflows during this process, maybe caused by Windows temporarily advocating to itself all the CPU cycles to perform 
its housekeepings. Receiving SSB, that maybe is not even perceived, or with just a small click. But with phase sensitive 
demodulators, like in the case of the Wolf encoding, that will play havoc and ruin the correctness of the demodulation.

One possible solution, that works in many cases but not all, is to keep the priority of the process that implements the 
SDR, and of the Wolf program, above the normal level. This usually helps a lot.

73  Alberto  I2PHD