|
Digital Drive Upsamplers, DACs, jitter, shakes and analogue withdrawals, this is it. |
|
|
|
Digital Drive Upsamplers, DACs, jitter, shakes and analogue withdrawals, this is it. |
|
In Reply to: upsampling article posted by solidgore on August 22, 2003 at 08:08:38:
"As will be explained shortly, all
upsampling DACs employ slow roll-off digital reconstruction filters as opposed to
conventional DACs, which employ sharp roll-off or brick wall reconstruction filters."This is not true at all. Most upsamplers use a brickwall at the original fs/2. Even the dCS boxes do so. If you look at the origin of these boxes, studio sample rate conversion, you'll understand why.
"The main reason is that the digital
audio data residing on a CD is already irreparably time smeared. No amount postprocessing
of the digital audio data by the playback system can possibility remove or
reduce this time smearing."This is very much true, assuming that at the ADC either oversampling linear-phase FIR OR high-order fs/2 analogue anti-aliasing filters were used. This assumption is true most of the time.
"For perfect waveform reproduction, sampling theory requires that the digital
reconstruction filter remove all frequency components above the Nyquist frequency or ½
the sampling rate of the input digital data."The is incomplete and/or misleading. Sampling theory does not talk about the *digital* reconstructor. Theory prescribes a very specific reconstructor (Sinc), which just happens to be impossible to realise in the real world.
"The lack of correlation between the baseband audio signal and its ultrasonic
images has important implications."Their correlation is obvious: the image equals the baseband signal but reversed in spectrum.
"The ultrasonic image signal essentially
frustrates any systematic interaction between the DAC’s differential non-linearity errors
and the slow-changing baseband audio signal,"This would be true if the spurious signal really would be uncorrelated and random. But it is correlated. In the lower reaches of audio, where Rife's scheme would be most beneficial, the corresponding HF images are pretty close to fs. But even slow-rolloff digital filters (see Wadia or the optional slow filter on the DF1704) have a lot of attentuation at fs. So the 'dithering' just don't, or barely, happens.
"However, this feature is rarely enabled because
the noise floor of the DAC is increased depending upon the amount of differential and
integral non-linearity present in the DAC as well as the level of the ultrasonic dither
selected."This is irrelevant and probably based on false assumptions.
"It was
shown that the ultrasonic images output by this “poor†filter is responsible for the
improved sound quality,"No, it wasn't. What was shown was a simple thought experiment that would be easy to assert or refute by anyone having access to an Audio Precision and e.g. Burr Brown's DF1704/PCM1704 demoboard.
But again, quite irrelevant as modern oversampling filters include noise-shaped dithering of their output anyhow.
This post is made possible by the generous support of people like you and our sponsors:
Follow Ups
- Re: upsampling article - Werner 08/27/0305:31:36 08/27/03 (0)
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
