Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Creative Labs (CREAF) -- Ignore unavailable to you. Want to Upgrade?

To: Andy Chong who wrote (5890)	11/18/1997 10:27:00 PM
From: Jon Tara	Read Replies (3) \| Respond to of 13925

Andy, I'm not sure that he really meant "technically", as in "the electronics are screwed up". But just in case that IS what he meant, I will offer my opinion. I DO think that the AWE series is technically backwards in how it gets data back and forth to the waveform audio subsystem. They retained the old Soundblaster standard using conventional DMA. This imposes a significant CPU overhead at high sampling rates, and can even lead to dropouts. It is NOT an easy task to program the drivers for this, either! Many years ago, Turtle Beach recognized that conventional (ISA) DMA is obsolete, and that they could actually impose less CPU overhead by using a FIFO mapped to a memory location or locations, and using block-move instructions to move stuff in and out. While DMA was more efficient on older PCs (8088, and maybe 80286 machines), DMA is now (on 80386 and higher machines) actually SLOWER and has a HIGHER OVERHEAD than programmed I/O done with block moves! This is because the DMA cycle speed was never increased as processor speed increased, and now those slow DMA cycles that the bus has to wait for constitute a significant wait. FWIW, we (I/OMagic, a company I did development work for) used a similar technique in designing the first PCMCIA audio card and video capture card. While we did it mainly because DMA was not possible on PCMCIA at the time, we did the math to satisfy ourselves that this was the right way, and that there was no reason (other than compatability) to rush to DMA once/if it became available. In fact, doing block I/O to an I/O port (Turtle Beach uses a memory location, which is much faster than I/O) is even faster and more efficient than DMA. Things got even dicier for Soundblaster wave I/O when the Internet bought upon us all these full-duplex applciations (Internet phone, etc.) Creative had to dig back even FURTHER into ancient history to dig-up an 8-bit DMA channel to handle the input side of things! (Output is 16-bit DMA.) This means that quality of the outbound signal suffers, as they only take 8-bit samples when capturing in full-duplex operation. (They COULD do 16-bit samples, and move them in with 8-bit DMA, but I don't beleive that they do this. Somebody who knows more about this please correct me if I'm wrong though.) Suffice it to say that full-duplex with any of the current Soundblaster cards is a royal kludge! Now, did any of this really matter, though? Creative is now coming out with PCI products (did anyone catch that they are introducing a PCI card, though only in an OEM version?) which will likely be even faster and more efficient than the mapped memory/FIFO approach. Creative managed to bamboozle their customers with a basically meaningless -64 suffix (which basically meant "same old thing, with some software to add an extra 32 synthesizer voices, and some improved sound quality) and was able to skip the engineer and marketing costs of the generation in the middle. While sales of AWE64 have been brisk, if you look beneath the surface, there really isn't that much reason to upgrade. Creative has been lucky that their customers have perceived "wow" where there really is no "wow". But with the new products, there really ARE some major improvements and advantages, and some good reasons to upgrade. This could temporarly backfire, if customers feel that they should now wait for new products, and delay purchase of AWE64s. However, this may be precisely why Creative chose to obfuscate the new products by giving them AWE64-xx suffixes instead of new product names. It will take a while for these new products to sink into the public's mind, and hopefully they will have a way to hammer home the point that these are all new once they are ramped-up for production and have old product cleared out of the channel. (Not that AWE64 is likely to go away, and will likely remain as the low-end/commodity product.) With the price of the AWE64 brought down to commodity levels, and some real meat - not just marketing spin - on the new high-end products, Creative now has both the low-end and the high-end of the market covered - and have really created a whole new high-end (5-channel) to boot. (Though they don't cover the super-low-end, where Aztech and others will continue to scratch-out a meager existence. One could consider, though, that these second-tier manufacturers are just Creative's forward sales force. :)