The New Sound

Out of all of my systems, one of the components that I’ve upgraded the *least* over the last 11 years is the sound card. I first got an SB16 way back in the mid-’90s and continued to use it alone until a couple years ago, when the lack of ISA slots in a new motherboard forced an upgrade to an SBLive. Even then, the SB16 continued to live on in my server box. The sound card is just one of those parts that I never really felt an urgent need to upgrade. It produces sound…what more do I need? Whereas the clarity of a new video card’s higher resolutions or the speed of a new processor are easy to appreciate, the subtleties of a different sound are harder to quantify to a tone-deaf musical ignoramus like me.

Nonetheless, upgrade time has come again and a shiny new Audigy 2 ZS has kicked the SBLive out and down the hand-me-down chain into the server box. The reasons are somewhat more practical than audible, though: the old SB16 in the server box was simply annoying the hell out of me.

Unfortunately, audio support under Linux isn’t quite perfectly seamless yet. It’s easy enough to add a sound card, load its device driver module, and start using it, but with older cards like the SB16 there’s a single-audio-stream limit and thus only one program at a time can be producing sound. This quickly becomes a problem whenever I’m playing MP3s on that box, since then any incoming IM messages don’t produce an arrival sound. Instead they all get queued up until the end of the song and then I’m suddenly deluged with ‘bingbingbingbingbing…’ if someone’s feeling chatty all of a sudden. It also makes ‘xmms’ stop playing and spit out an error and I have to go and restart the playlist. Annoying.

The usual solution is to run a sound daemon like ‘esd’ or ‘arts’, or a plugin for ALSA like ‘dmix’, that will do all of the mixing of multiple sound streams in software and then feed the single mixed stream to the sound card. It works, but with drawbacks: they increase the amount of memory and processor overhead needed (important on a less-powerful box like this server), and they introduce some latency into the audio stream. When xmms is using the sound card directly, all actions like starting and stopping and skipping tracks are instantaneous, but when it’s directed through a sound daemon instead, there’s a slight lag between clicking a button and when it actually takes effect. In particular, it *really* slows down attempts to skip quickly through a few tracks in a shuffled playlist. Annoying.

So, instead, I’ve taken the easy way out and moved the SBLive into the server. The Live can mix streams in hardware instead, so programs can continue to use the sound devices directly, they can play simultaneously, and I still get instantaneous responsiveness without any extra overhead. Of course, that left my gaming system without a sound card, and *that’s* the real reason why I picked up the Audigy 2.

Now that I have a spiffy new sound card though, do I notice any difference? Well…maybe. It came with a sample DVD-Audio disc with 24-bit 192-kHz songs (versus your usual 16-bit 44.1-kHz CD), and some apps demonstrating the different environmental effects available. I *think* the songs on the disc are coming through a lot clearer than what I’m normally used to, and I jumped into BF1942 for a bit and I *think* I’m hearing better positioning and realism in the environmental sounds, but I’m not 100% sure. It could be an actual increase in sound quality, or it could be the psychological effect of expecting to hear something better kicking in, and it’s hard to say for sure with no easy way to compare the old and new output (I’m certainly not about to continually swap cards in and out just to try and compare them).

Maybe improved sound quality is just one of those things that’s harder to appreciate until you’re forced to do without…