Pondering on Sound Card Configuration

3 posts / 0 new
Last post
W7OWO
Online
Last seen: 8 min 19 sec ago
Joined: 2016/07/23 - 02:40
Pondering on Sound Card Configuration

I’ve got a situation that I may need some clarification on to see if my reasoning is correct. As I firmly believe I am in the category with this topic of “I know enough to be dangerous”.

I’ve been using the WSPR mode for about 1 1/2 years, have used it with multiple SDR receivers, and about 5 different transceivers between my three, and two owned by friends in the same radio club. Until recently, I followed the instructions provided by Joe Taylor for configuring the playback and recording advanced options of the sound card, 16 bits and 48000 Hz. My most recent transceiver is a fairly inexpensive SDR that I use with HDSDR, and an external high quality sound card. What surprised me is that the instructions specified that I set the sound card to 24 bits and 96000 Hz. My computer I run WSPR on is also set up to my receiving antenna that I use an SDRplay on for spotting WSPR, and my FT857 I use for WINMOR and UHF/VHF packet with Winlink express. With all those USB items in use, needless to say I get an USB resource issue error. This occurs when using any configuration of USB port usage I can try when attempting a 96000 Hz setting for the sound card. So I use 48000 Hz, and to my point of view everything looks good. Which has got me to wondering….

As I understand it, the sound card frequency determines the band width that can be processed, and as a minimum, I should have a sampling frequency twice the highest frequency component. I’ve read that they chose the 31 bps rate for PSK31 as it was easily resolved from the 8000 Hz DSP systems that were available at the time. With that in mind, since I have HDSDR filtered down to between 1000 Hz and 2000 Hz to cover the 1400 – 1600 Hz range that WSPR uses, I’ve got plenty of room to spare, as at a minimum I should need a sampling rate of 2000 Hz. If my thinking here is correct, since SSB Phone should be limited to 2500-3000 Hz range, so its minimum sampling should be 6000 Hz. Again, this should be plenty of headroom at 48000 Hz.

From strictly an audio point of view 24 bits makes since, as I can see that increasing the bit depth can increase the dynamics of the conversion, improving the signal to noise ratio. But with AFSK modulation is that level necessary, since the tones used in WSPR have a spacing of around 1.4648 Hz. A bit depth of 16 gives 65536 unique values, from my calculations it should be able to discern to the nearest 0.0001 Hz over a 6 Hz frequency. Though I’ve used the 24 bit depth where I could, and it hasn’t caused any issues.

I may be tilting at wind mills here, but to go to the 24 bit 96000 Hz sound card setting I believe I will need to add a USB card to the PC to that doesn’t have much activity on it. But all card slots are used up. Don’t want to get a new machine at this time. Right now I’m going with the thoughts of “If it ain’t broke, don’t fix it”.