I'm beginning to fear that the wonderful sounding Pioneer PD-75, the best sounding digital I've ever heard in my home, is adding something. It's like it's adding extra energy, a fuzz which makes attacks higher and so on. It seems to be making my Krell work harder too. This barely registers below 20kHz were it is far far far quieter than the DVP-9000ES playing SACD. The pioneer shows nothing above -90dB below 15kHz. However, I fear, a bit, that the rise I see starting at 15kHz continues up for a bit, but my Behringer display stops at 20kHz so I'm not sure how far it goes. However...the noise is unlikely to go as far as with DSD, precisely because it can't, the aliases would become too big. So, in the end, some kind of digital filtering must being done, in the case of a 16mhz bitstream pwm, an interesting kind for sure. Julian Hirsch measured no more than 0.004% distortion 20-20kHz, so the distortion isn't huge, though digital distortion is especially undesireable because it becomes proportionately larger at low levels. But meanwhile Sony was getting 0.0015% distortion with similar technology, but not sacrificing much to pulse coherency. Now 0.004% is comparable to many modern time coherent slow filters, while the culty NOS gets you above 1%.
So, that's Yang, extra positive energy radiating out beyond the envelope for the Pioneer--though very little extra, 0.004%, what's with the Yin Sony?
The Sony SACD as implemented in the DVP-9000ES with the same Sony bitstream chips as the SCD-1, which may have been the last bitstream models as bitstream models generate a lot of heat and noise which must be managed, generates quiet, better than 16bit quieting in the midrange, from 50% noise (1-bit 2.8mhz sigma delta operating at no more than 50% duty cycle) by creating holes, holes in time of the stream of pulses, holes and holes made of holes That's all it can do, the pulses can't get any larger..
It can't get more obvious than that, can it?
Does the "holeyness" of DSD get better in highly oversampled and overquantized sigma delta implementation as nowadays? Yes, buuut...the holeyness is there, it's made of holeyness, all the blur can't make it go away.
Or, you could just say the Pioneer adds distortion--extra energy, DSD has perfect linearity in principle but ginormous HF noise, which obscures because it's random.
If you can't hear the noise or distortion, does it matter? Yes, I suspect it may, you may not hear the noise or distortion itself but I suspect it can still add (if it's distortion) to attacks and such, or obscure if it's pure noise.
So the trick of eliminating the need for severe HF filtering is removed in DSD and replaced by huge clouds of noise (which aren't even Gaussian noise, but slightly tonal). These mostly inaudible but obscuring clouds give DSD it's Yin character, while alias leakage gives a Yang character.
So, that's Yang, extra positive energy radiating out beyond the envelope for the Pioneer--though very little extra, 0.004%, what's with the Yin Sony?
The Sony SACD as implemented in the DVP-9000ES with the same Sony bitstream chips as the SCD-1, which may have been the last bitstream models as bitstream models generate a lot of heat and noise which must be managed, generates quiet, better than 16bit quieting in the midrange, from 50% noise (1-bit 2.8mhz sigma delta operating at no more than 50% duty cycle) by creating holes, holes in time of the stream of pulses, holes and holes made of holes That's all it can do, the pulses can't get any larger..
It can't get more obvious than that, can it?
Does the "holeyness" of DSD get better in highly oversampled and overquantized sigma delta implementation as nowadays? Yes, buuut...the holeyness is there, it's made of holeyness, all the blur can't make it go away.
Or, you could just say the Pioneer adds distortion--extra energy, DSD has perfect linearity in principle but ginormous HF noise, which obscures because it's random.
If you can't hear the noise or distortion, does it matter? Yes, I suspect it may, you may not hear the noise or distortion itself but I suspect it can still add (if it's distortion) to attacks and such, or obscure if it's pure noise.
So the trick of eliminating the need for severe HF filtering is removed in DSD and replaced by huge clouds of noise (which aren't even Gaussian noise, but slightly tonal). These mostly inaudible but obscuring clouds give DSD it's Yin character, while alias leakage gives a Yang character.
No comments:
Post a Comment