advertisement


Timing and frequency response

onlyconnect

pfm Member
From this paper:

http://www.temporalcoherence.nl/cms/images/docs/HelsinkiTC.pdf

Extended frequency response is not required because our ears can detect waves of such frequencies, but is required to resolve the reproduced sound sufficiently in time to satisfy the temporal resolution of our hearing.

I am not sure that I fully understand his argument though, and he does kick off with a lot of anecdotal assertions.

What do the experts think?

Tim
 
I am not sure that I fully understand his argument though, and he does kick off with a lot of anecdotal assertions.

Yes, I don't quite get it (after a quick reading), and some of his references are definitely suspect, but in any case his theories should be reasonably easy to test and verify.

He clearly isn't a neutral researcher either, but is pushing his own rather exotic speaker design.
 
His argument seems to be that you don't need high sample rates to reproduce signals above 22 kHz, but to allow for sufficient temporal resolution.

As we already know that even a 44.1 kHz sample rate system can reproduce timing differences down to less than a nanosecond (way shorter than the time intervals he is talking about), he seems to argue that it is the low pass filter rather than the sample rate that causes the "time smear" he is talking about.

Thus his arguments would apply equally well to analog systems with a low pass filter behaviour - conventional tweeters, anyone?
 
I didn't read the pdf but the quote is self-contradictory and/or deliberately obfuscating the issue.
Either you get the frequency response (without distortion, delays, etc. - in the audible range) or you don't.
A different but similarly confused paper was referenced in the "vinyl vs. cd" thread.
 
Highly incoherent. And with gems like "The apparent differences between "live" and reproduced sound can be explained by looking at the
impulse response of audio components" also quite clueless.
 


advertisement


Back
Top