Great to hear, ac97 has held audio back for far too long. Long live aza...(whatever its called).Natoma said:http://www.intel.com/pressroom/archive/releases/20040415tech.htm
It's about freakin time.
very soon, i believe intel's next chipset (for tejas) will have it built in.K.I.L.E.R said:So when will this be implemented into mobos?
To avoid aliasing, you need sampling rates well above the actual frequencies being sampled. With an assumed audio bandwidth of, say, 15kHz, Intel's rate of 192kHz is clearly (pardon the pun) going to be better than anything with a rate of 44kHz. Of course, whether the end user will be able to tell the difference using £30 desktop speakers is another discussion entirely .CosmoKramer said:Great. Now even your dog can enjoy listening to computer audio. Because you are aware that those samplerates are only useful to represent ultrasound?
epicstruggle said:very soon, i believe intel's next chipset (for tejas) will have it built in.K.I.L.E.R said:So when will this be implemented into mobos?
later,
epic
IST said:Tejas is next year. It might be in the Socket 775 boards though.
K.I.L.E.R said:So when will this be implemented into mobos?
epicstruggle said:Great to hear, ac97 has held audio back for far too long. Long live aza...(whatever its called).Natoma said:http://www.intel.com/pressroom/archive/releases/20040415tech.htm
It's about freakin time.
epic
ps this should be in the hardware forum
God, this just needs to be put in my sig.Natoma said:Doh! I think you're right. Oh wells.
Neeyik said:To avoid aliasing, you need sampling rates well above the actual frequencies being sampled. With an assumed audio bandwidth of, say, 15kHz, Intel's rate of 192kHz is clearly (pardon the pun) going to be better than anything with a rate of 44kHz. Of course, whether the end user will be able to tell the difference using £30 desktop speakers is another discussion entirely .CosmoKramer said:Great. Now even your dog can enjoy listening to computer audio. Because you are aware that those samplerates are only useful to represent ultrasound?
Agreed, hence my comment about £30 desktop speakers! That's, of course, assuming the filters in current AC97 implementations are up to scratch. If there was no appreciable performance overhead between 44.1k or 192k sampling per channel, in any hardware device supporting HDA 1.0, which would you run with anyway? Guaranteed there will be plenty of users who will swear that it sounds better. By the way, if one looks hard enough, there are always plenty of links to be found supporting the claims that sampling rates higher than 44.1kHz are worthwhile and audibly better. I teach the Nyquist theorem in the above context to A-level students, who then go off to exam it in great detail using the appropriate software and equipment; I get a different set of responses back each time though so I can't quantify any of the statements made in such links.CosmoKramer said:The AC97 standard is 48 kHz which has a fs/2 = 24 kHz which is sufficiently high for the anti aliasing filters to be gentle enough that any artifacts are inaudible.
Bolloxoid said:Resampling everything to 48 kHz is the unforgivable crime of AC97.