Computer and SDR

[SCPD]

Hello all. I have a few question. Sorry in advance if i does not make sense but here we go.


I have the RFSPACE SDR-IQ running it with HDSDR hooked up to a dualcore with 1 gig of ram. The audio card that is playing the IQ data is a diamond 5.1 set to playback at 16bit 44khz now does the playback effect the filters and bandwith if so should i use my onboard audio which allows for 192 khz of playback or it does not matter, Also which one would be less taxing on the CPU in HDSDR 44hhz playback or 192khz; Sorry if does not make sense but im trying to optimize my setup for the host software and the sdr


Thanks
Mike

 

[SCPD]

Nobody? Token, Bradon, Ridge... i fell a little lonely in this thread

 

[Napalm]

Your onboard would probably not be as good quality.

With my crappy onboard sound I get a deep interference pattern at 0Hz (or basically lots of ground noise in the middle of the I/Q display).

 

[SCPD]

Sorry it took me so long to reply. I have post all over the place on here. Thank you for taking the time to answer my question.

 

[kc2rgw]

Lower bandwidth is easier on the CPU, less data to have to process, of course you also see less.

There are different implementations of 192kHz and you will find that only the highest end pro-audio cards start to truly approach it for full useful width. I had an Edirol FA-66 here for a while and it was really good, but they are expensive. That card also requires firewire, which is also good as it offloads bandwidth from the USB bus. If you go USB, optimally put the soundcard on a dedicated USB expansion card.

I didn't have internal slots for adding a soundcard, thus why I went firewire. Tons of info on various soundcards people have tried on the flex radio mailing lists on yahoo etc.

 

[Token]

OK, some of this is supposition, and if anyone knows better I would appreciate the information also. But…

The SDR-IQ is a DDC SDR. The signal is digitized in the SDR–IQ and the I and Q data are sent to the PC via the USB. The information is then processed in the CPU and the demodulated digital audio, with filtering already applied, is sent to the sound card for playback.

Now, if the above is correct and my understanding not skewed the sound card will have no real affect on things like filter performance.

What I can say for sure is I have a PC here with three sound cards in it, one for each SDR that is hooked to that machine from time to time (normally it only has two SDRs on it) so that I have separate audios out for each. One of those sound cards is an ISA basic old 16 bit soundblaster compatible knock-off. One of the is PCI 128 Soundblaster, and the other is a high-end 24 bit 192 kHz unit with 7.1 Dolby EX/DTS and fiber optic in and out. And no matter what sound card I use in this setup the filter performance does not seem to change. Now, for sure there is an audio quality difference from the top to the bottom end, but not a filter performance difference that I can tell.

Now, this is a great question to be posted to the SDR-IQ Yahoo group. Do you belong to that Xeno?

T!

 

[SCPD]

Wow thank you token. I just bought a new quad core with 4 gigs of ram so hopefully sdr-console will preform a lot better than it does now on my dual core. I do not use yahoo groups anymore. I actually do belong to the sdr-iq group but im never on it. Mostly on here and #wunclub on starnet

 

[Token]

Got a response from RFSpace on the SDR-IQ Yahoo group to confirm that my description was correct. Yes, the sound card or the sound card quality/sample rate has nothing to do with the performance of the filters in regards to the SDR-IQ. End of RFSpace confirmation, beyond this is mine.

The above is most likely a general statement for all DDC SDRs (Excalibur, Perseus, QS1R, etc) and software suites as I can not see any software developer writing processing to the sound card when he/she cannot know what sound card the user will have.

With sound card based SDRs (SoftRock, Soft66AD, SDR-1000, Lazy Dog SD-1, etc) the sound card quality will have an impact on filter performance as channel-to-channel phase tracking (and thus I/Q phase tracking) and the sample rate will be significantly less good than DDC even with the best sound card possible. What that means is that the sound card SDR hardware itself might be able to perform quite well but the sound card sampling itself becomes the limiter on filter performance and dynamic range. That also probably accounts for the wide variations in reported sound card SDR (even within a given model) performance. One person might have a killer sound card and the other a so-so one, but the problem gets blamed on the SDR.

T!

 

[SCPD]

Thank you Token. I finally built the new tower for my SDR-IQ. Amazing difference from what i was using. I decided to go with the on-board audio from the motherboard and like your said there is no difference from running it at 24bit 192khz rec/playback to 16bit 44.1khz rec/playback. My previous system was a dualcore with 2 gigs of ram. Now im running a quadcore with 8gigs of ram and 4 terabytes. Lag is a thing of the past now.