Topic: Sound Blaster ZS Notebook - PCMCIA

[it-goes-to-eleven]

I thought I'd start a thread to document my experiences with this card.  I've been looking for a 24/96 solution to use with my Minime and laptop taping.  I can currently do 24/44.1 reliably via USB.  These are on sale for $99 at best buy and I picked one up on Friday. I figure I'll shake it down and make a decision on whether to keep it.

I hope to ditch the lappie for a Flash Recorder (or similar) when they come out. As a result, I'm not interested in spending much time or money on getting 24/96.

There have been a couple of different versions of PCMCIA sound cards from SB over the past year.

http://www.bestbuy.com/site/olspage.jsp?skuId=6991641&type=product&id=1099392683811

This review describes the substantial shortcomings and audio problems when operated at 44.1K:

http://www.extremetech.com/article2/0,1558,1684450,00.asp

Anything this card can do beyond bit correct 24/96 captures would be gravy. Also, the card must work reliably under Linux.  There is some Linux support for this line of cards but my initial efforts went poorly. In the meantime, I've been doing some testing under windows. Because if it can't do bit correct captures in that environment, Linux is moot.

Analog loopback testing of the card was very disappointing.  The same issues that extremetech encountered were demonstrated.  The analog performance is, in a word, crap. It has been suggested that this card only really supports 48k and 96k and that 44.1 is resampled.  That might explain some of the noise. I don't care much about this feature, but it would be nice to have the option.   96k analog capture was also mediocre.  All card connections to this card, including those on the pigtail, are via 3.5mm miniplug.

For 24/96 testing, I sent a 24/96 digital sample from my desktop system via USB to a UA-5 and then via coax to miniplug to the SB card in my laptop.

One immediate issue is that the card seems to be changing the signal levels.  The source file is down perhaps 6-9 dB from peak. Yet when captured at 100% volume, levels come up to peak.  I tried SB's record utility as well as Sound Forge.  I'm scratching my head over that.   Obviously, this is a show-stopper if it can't be resolved.  The card has two modes of operation. In high power mode, EAC processing is handled by the card.  In low power mode, it is handled by the host in software.  I hoped that low power mode might be less likely to mess with my signal but it made no difference.

I normalized both the source and the capture data in an effort to see how closely they would match when merged via invert. For the right channel, things went well and it appeared that the signal was correct.  The left channel had a noise floor that suggested an issue.  There was also an odd spike that was reproduced each time I ran the test.  I could find no reason for the spike in the original reference file.  The noise floor could be due to a normalization issue (I did normalize independently).

The result on the right channel is encouraging.  But the issue with levels being altered is not.  Anyone have any thoughts on that?