Home Propeller Head Plaza

Technical and scientific discussion of amps, cables and other topics.

[ Follow Ups ] Thread: [ Display  All  Email ] [ Propeller Head Plaza ]
[ Alert Moderator ]

Re: Oh Really??

"I wouldn't think the digital filter should be doing anything different than an analog filter in its place (except is it more likely you can do it better and with perfect repeatability with a digital filter)."

The filtering can be analog, and it was back in the infancy of CD playback, when Sony opted to use analog brickwall filters. Digital filters won out because the average phase error at higher frequencies was a lot lower and, at the time, sounded a lot better...

"It should not be averaging the signal, it should be applying as close to an ideal lowpass filter as it can to the signal."

While it is indeed a LPF, in the context of *dithering*, an ancillary function of the digital filter is taking the randomly-triggered LSB and averaging its value with resolution below the LSB. This is how information below the LSB is recovered by dithering.

"This is not averaging it is reconstruction which is why intersample peaking can occur."

Intersample peaking is a totally-different phenomenon. It's when a medium is recorded *hot*, and the calculated values by the digital filter exceed the maximum value within its dynamic envelope, resulting in "clipping" of the output signal. The opposite end of the dynamic spectrum of what dither is acting on.

"If I have a sine wave at exactly 1/4 the sample rate (1/2 the highest allowed frequency) in such a phase relative to 4 samples which represent a wavelength so that that all 4 samples are in the middle amplitudes of the wave (this is easier to sketch than describe) then reconstructed peaks are much larger than the sample peaks."

It's very rare when such peaks are "much" larger... Intersample peaking is rarely an issue with continuous sine waves.... (But then again, it has nothing to do with dithered signals and how digital filters act on those signals.)

"Any good digital or analog reconstruction filter will produce the peaks correctly - averaging would fail this case miserably."

Once again, in the context of dithering and random triggering of the LSB, that's the nature of how the filter works. The filter has greater resolution than the raw data's LSB, so the calculated values will instead of being exactly at the level of the raw data LSB's 0s and 1s, will be at values *between" the raw data LSB's 0s and 1s. This is what I mean by "averaging."


Health Insurance
Healthcare

This post is made possible by the generous support of people like you and our sponsors:
  The Cable Cooker  

Follow Ups Full Thread
Follow Ups
You can not post to an archived thread.

[ Top Top of Page ] [ Contact Us ] [ Support/Wish List ] [ Copyright © 2025, Audio Asylum, all rights reserved ]
[ General ] [ Speakers ] [ Tubes ] [ Vinyl ] [ Digital ] [ Hi-Rez ] [ Video Asylum ] [ Cables ] [ Tweaks/DIY ] [ Music ] [ Films ]