SACD fundamentally flawed?

Excellent, Markus!

Remember that the sampling theorem requires only slightly more than two samples per cycle to PERFECTLY represent a sine wave that is in band.

_________
Thom

 

And some are not happy with over/up/under/sideways/down sampling - for example AudioNote (UK). And their DACs (I've heard the model 4 and 5 and signature versions) sound extremely realistic. We are still learning.

 

Steve: We know you master the DSD and PCM as split feeds rather than merely make a DSD master and derive the PCM sample from that.

But have you given the DSD into 16 bit PCM sample a careful listen? Can you share your opinion based upon the most recent tools being used?

Richard.

 

I created a sine wave using Wavegen (I think if you google Wavegen21 you can still download it). It's an old Win3.1 signal generator program.

 

Let's do an experiment. Look closely at your graph. Study the fifth and the sixth sampling points (the ones just above 0). Notice how close they are in amplitude?

Now, consider a DAC. How can the DAC know that between those two points a peak exist? The answer is - it doesn't know, so the analogue output follows an almost horizontal line.

Ergo, the line that connects the fifth and sixth sampling points should be an almost straight line, not a parabole. We know there's a peak in there because we know the signal is a sinewave. But a DAC doesn't (unless there's a DSP to rebuild the original signal).

Thomh, the Nyquist theorem does not infer a perfect reconstruction of the original signal if the signal does not exceed half of the sampling frequency.

http://www.digital-recordings.com/publ/pubneq.html

If I understood correctly (and didn't sleep during my classes), the Nyquist theorem does describe perfectly a signal's frequency, but not the shape of it.

Here's an example of what I'm talking about. Check out the "A few rules" section.

http://members.aol.com/ajaynejr/nyquist.htm

 

I took the liberty to correct your sampled signal. You'll see in red the real trace you are feeding to your DAC.

 

It doesn't go like that. The DAC knows it's a sine wave. All other waveforms are composed out of an infinite number of sine waves.

Thanks for the links, Javier! That first link looks interesting.

 

I just thought of a better example...

At exactly Fs/2, a sampled sawtooth signal and a sinusoidal signal would look exactly the same... yet they are two completely different signals.

 

This is in violation with the Nyquist theorem and therefore irellevant.

_________
Thom

 

Yes, that's why we need slightly more than two samples per cycle.

 

They are different signals, but do they sound the same? Since the sawtooth is distinguished from the sine by higher harmonics beyond the range of hearing, the ear perceives them both as sine. The reaction of electronic equipment to the higher frequencies (aka distortion if the original signal was in fact a sine) may be another matter.

 

http://www.its.bldrdoc.gov/fs-1037/dir-025/_3621.htm

What part of the Nyquist theorem did my example violate?

 

For an 18kHz wave to be a sawtooth it must contain higher harmonics beyond the Nyquist limit, therefore it is not a valid input. It would be brickwall filtered before entering the ADC and subsequently turned into a sine.

 

Absolutely true! Thanks for the clarification.

 

Gardo, might I offer up a more representative quote (without permission) from that four-page article:

... SACD uses a proprietary fixed delta-sigma modulation code, known as DSD. As delta-sigma modulation technology is well-known to audiophiles, I will only briefly summarize it here.

In the simplest form of delta-sigma modulation, a 1-bit converter is oversampled in a feedback loop. By definition, a 1-bit system—it's either on or off—suffers from very high quantization noise. However, the feedback loop shapes the noise energy so that noise is acceptably low in the audible band but high outside it.

Audiophiles have been suspicious of negative feedback for years, but for reasons I don't understand, they appear to have given delta-sigma modulation their implicit approval. Unlike feedback loops in amplifiers, where the loop circuit is already somewhat linear, the feedback loop in a delta-sigma converter is wrapped around the most nonlinear component we can imagine: a 1-bit A/D converter.

So, tell me again about the miracle sound of DSD?


FFF

P.S. THANK GOD (ONCE AGAIN) FOR STEVE'S ZOMBIES SA-CD!

 

Steve: Sony provides tools to take a DSD sample and directly convert it into 16 bit PCM. Many hybrid disks out there have taken advantage of this PCM sample for their CD layers. You use your art to master your own PCM layer directly from the analog source. My question is if you have any experience with the DSD into 16 bit sample and if you were at liberity to express your opinion.

Richard.

 

"Trickle down"??? Could you translate that, please???

 

Is this what BMG/RCA is doing with their Elvis reissues???

 

I think the releases that feature the "trickle down" technology have a logo on the back that reads "Direct SBM". That's the technique Sony uses to do DSD -> 44.1/16 bit PCM.

 

The Nyquist criteria requires that you sample at GREATER THAN twice the bandwidth, not greater than or equal twice the bandwidth.

__________
Thom

 

I did say the "real world" and not the "hi-end" world

Call me old fashioned, but I feel a reconstruction filter should be part of any properly-designed DAC and/or CD player.

IIRC, Hi-Fi News had a review of the DAC 5 and while reporting great sonics, the measurements (as expected), showed more distortion products than actual signal.

__________
Thom