Aren't Sound Blasters what they used to

Question

Hello everyone,
I have been a Creative Labs user for several years, especially before 2000 (Soundblaster 2.0, SB6 Vibra ISA (2x), SB Pro .0), and was absolutely pleased with them, especially with DOS games. Those were the happy times...
Anyway, in more recent times, after years of listening to trustworthy but not really performing AC'97 integrated soundcards, I decided to purchase an "audio accelerator" in the form of an Audigy 4 (the one with the EMU 0300 chip and no firewire/remote control, if that means anything).
Anyway, it worked fine for a couple of weeks or so, I got none of the SCP problems posted here, however that warning in the manual about the "VIA hard disk corruption problems" troubled me...after some research I saw it was mainly about older VIA chipsets (I had an A64 3200+ with KT800T VIA chipset so I relaxed and said "this cannot happen to me, seems like a problem solved long ago". How wrong I was....
One day, out of the blue, while playing Battlefield 942 I got a horrible BSOD and had to reboot. After than, and after a VIA drivers update, I noticed that the sound stumbled and HARD DISK SAVING stumbled too, which was much more dangerous. I reverted to an older version of VIA drivers that didn't cause that problem, but one day I got another BSOD coupled with disk corruption, and, apparently, what later revealed to be a SATA CONTROLLER PHYSICAL DAMAGE!!!
To make the long story short, I had to replace my motherboard, and after checking these forums I discovered that disk corruption, PCI Bus Mastering and SCP problems have been plaguing the whole soundblaster product line ever since the Li've! series and were never really solved, plus it's not clear if they are purely software or hardware, or both.
In any case, this card caused me a lot of expenses, and after reading the horror stories here, I am really afraid to plug it into any of my PCs, in fact I'll try to sell it hoping that the manufacturer's reputation doesn't make it impossible.
So really, what's the deal here? Is it just me, or Sound Blasters just aren't what they used to be? How can a whole class of hardware problems persist for so long, even trhough the X-Fi series? If I had seen those forums BEFORE handing out 70 Euros for that d4mn thing, I would have probably kept my money and stick to AC'97 with rock-solid drivers and hardware!Message Edited by CreativeSkeptic on 03-5-200704:20 PM

Answer 1

Well, I don't know how the X-Fi series scores in any of those fields, as I can only speak for my experiences with the Audigy 4.
Apart from the times where it obviously glitched and/or interacted badly with the rest of the hardware, it played quite smoothly, and MIDI playback took noticeably less CPU time than with the software synth (at least according to task manager, and obviously not accounting for PCI bus utilisation), which was one of the reasons I initially bought this card.
Exactly what makes MIDI playback different than playing sound effects in games? Since both are made from samples read from memory and sent to the sound card in different streams (up to 32 or 64 depending on your model), they should be exactly the same (well OK, midi is perhaps sort of more "continuous", but a game laden with sound effects may generate 0s of sound events per second).
By admitting that such "hardware accelerated" playback is beyond the capabilities of the current PCI bus is ridiculous, and obviously defeats the purpose of an accelerator sound card. Then why should one ever buy one, if in the end it's going to perform worse than software mixing and synthesis?
The numbers also don't convince me: the PCI bus has a 33 MHz clock speed and a 32-bit data bus, for a nominal data rate of 33 MB/sec. Even in ths case you're using 92 KHz, 24 bit resolution soundfonts for 64 channels at once, That means a "mere" 36.684 MB/sec at once. OK, using stereo soundfont would actually double that, and reach a whooping 73+ MB/sec, which is probably impractical.
For more "normal" game sound effects and soundfonts, admitting they use 48 KHz, 6 bit samples in stereo, that's a measly 2.28 MB/sec for using all 64 channels. Since virtually no midi really uses 64 channels, as don't most games, that leaves a margin of 0: compared to the PCI bus capabilities. To make a comparison, a TV tuner card I have, when capturing a PAL signal, produces something like a continuous stream of 33 MB/sec of raw data that can go on for hours, yet I never had any problems RECORDING that, with real-time MPEG compression (in software, it was a cheap TV tuner) while recording the sound with the integrated soundcard.
If the Audigy or the X-Fi or whatever cannot handle similar or inferior data rates, then something is basically flawed. Sure, there is still subdivision of the data stream into PCI packets, a maximum allowed bus mastering time and timing-sensitivity in the case of MIDI/sound effects, but still, glitchy playback and/or multi-second pausing indicates a much worse error than "missing" control of the PCI bus for a few cycles.
Regarding the AWE32 comments...does this imply than all or most of the current Sound Blaster lineup is technically inferior to an ISA card released 0 years ago? Then the "hardware acceleration" is but an illusion, and we should all perhaps stick to integrated or "normal" soundcards?