My Summary about Computer Audio Music Server

[marce]

On 9/8/2018 at 6:36 AM, Em2016 said:

Eeep, from that link:
 

"CEC sounds smooth and analog like NOT because of the belt thing, but because CEC engineers modified the SPDIF trace to be unrecognizable as square. They filtered it and reshaped it to be without the low square harmonic fundamental. So DACS read it but lack bass. This is the fake analog smoothness which was created to fool people into the belt bullshit story.  And NOBODY EVER discovered this."

 

The problem is that statement is incorrect (in fact its complete and utter twaddle), the shape of the SPDIF wave has nothing to do with the resulting analogue bass output, the analogue signal is buried in the bits, surrounded by extra data such as reed-solomon information etc.

[marce]

On 9/11/2018 at 12:22 PM, marce said:

The problem is that statement is incorrect (in fact its complete and utter twaddle), the shape of the SPDIF wave has nothing to do with the resulting analogue bass output, the analogue signal is buried in the bits, surrounded by extra data such as reed-solomon information etc.

Further to my above post, the square wave looks to have been passed through an integrating circuit which is... a low pass filter so he can't even get his filtering right...

[marce]

Most people who are not involved with digital design or signal integrity think a square wave is the text book image of a square wave complete with 90 deg. corners. In reality most are not, DDR memory being a prime example. In SI the worse case situation is a none monotonic rising or falling edge on the wave causing spurious switching, a perfect square wave is also not desired, the rise and fall times should be chosen to be as gentle as possible without introducing other problems such as ISS. The slower the rise times the less high frequency energy being fed into the line and the more the wave edge will slope.

What worry;s me is that someone who improves digital equipment could come out with such a statement, that is wrong, the frequency content of the transmitted digital information has no relevance to the bandwidth of the resulting audio analogue signal after conversion. Also the fundemental frequencies are present, it is the high frequency harmonics that have been filtered from the wave shown, resulting in a very slow rise and fall time, but still leaving a wave that will cause periodic switching, thus the data gets through.

[marce]

Square waves don't effect the sound, they carry the data, that is converted at the DAC to an analogue signal. What we ndo understand in detail is how that data is delivered, bit perfect 0 and 1's all the way.

[marce]

No they don't, the shape of the square wave as long as the data gets through has ZERO (no) effect of the resultant analogue conversion... The information is carries in the 0's and 1's not in the wave, the shape of the wave as long as it is within the required tolerances is separate from the analogue output.

[marce]

It goes on about data errors, hence the signal integrity of the square wave is not up to the task, that should not happen these days, a badly engineered interface...

[marce]

Read it, if the digital data gets to the DAC bit perfect then everything is hunky dory, the shape of the square wave as long as it allows the data to be read has NO bearing on the resultant analogue output, its not an analogue signal it is a digital waveform transmitting a packet of information as bits. Most digital waveforms are not perfectly square, DDR memory is a perfect example, the data gets through, thats how it works, whether its an audio DAC or a high speed 24 bit DAC working in MHz's. As long as the digital wave is withing the signal integrity limitations for the interface then the data gets through and how that digital wave looks has no bearing on the analogue output.