Orpheus Zero Review (CD Player)

[sheeple]

True. But it does not play CDs...

 

[AnalogSteph]

The DAC is an ancient 24bits Wolfson WM8740

I refuse to think of a WM8740 as "ancient". Released in 2004, it's actually newer than the classic CS4398 which is still in production, and it saw use well into the 2010s. "Dated", maybe. Now a TDA1540, that would truly be ancient.

This is probably a language thing again. If memory serves, ancien in French covers pretty much anything old, while ancient in English usually means it saw the last woolly mammoths walking off or has first-hand experience with the Roman Empire.

 

[NTTY]

That’s what I meant Anything that’s been processing 0s and 1s for 20+ years is basically Stone Age, right?

But yes, I agree with you, and I’m actually glad someone reacted to this.

I see it the same way as you do, aged, not obsolete, especially when looking at the measured performances.

 

[Linnie]

A good opportunity for a non-technicians question, because it's so nicely visible here: What does this red circled little timespan represent? Is this the time the oscillator / transistor needs to "move" from the "positive" status to zero? And what happens during the yellow circled time?
Thanks for making me smarter .

 

[NTTY]

Great question.

Context

This signal is supposed to be only 3 levels, -1, 0 and +1 (the smallest amplitude in PCM 16bits). In the PCM code of the CDA, they are the below (in red):

PCM Code	Decimal Value	Output Voltage (assuming -2V to +2V amplitude)
0111111111111111	+32'767	+2/(32'767/32'768)=+1.9999389V
...	...
0000000000000001	+1	+2/(1/32'768)=+61μV
0000000000000000	0	0V
1111111111111111	-1	-2/(1/32'768)=-61μV
...	...
1000000000000000	-32'768	-2/(32'768/32'768)=-2V

That means, at the output, you'll get only 3 voltage levels (called 3DC), because at this smallest symmetrical amplitude of the signal, it is no longer possible to represent a sine. As you see from the above, there are only three possible levels.

So, when trying to represent a 1kHz (or 997Hz for this test) sine, the 3 levels should be looking like a square, with a half period transition through 0, all being perfectly flat. But it is not what you see.

What you see here are 3 things:

1. The limited bandwidth and the associated Gibbs Phenomenon: A square, or more precisely a flat voltage output in our case, is made of a sum of sine, the fundamental (997Hz here) and an infinite number of odd harmonics of decreasing amplitude. Since the CDA is band limited to 22.05kHz maximum, you'll be missing a lot of odd harmonics and so the flat will not be flat, showing that typical pre/post ringing.
2. The reconstruction (oversampling) filter: To reconstruct the continuous analog signal from the discrete samples, we use the sinus cardinal function to create a link between individual samples. Depending on this filter, you'll see more or less pre/post ringing. Slow filters (that have less pre-ringing) would show less ringing on the left side of the flats.
3. Random Noise and distortion: At this lowest level (few μV) , we are likely to get some added noise and other non linearities (distortion) that disrupt this fragile signal.

I hope it clarifies.

 

[Linnie]

Thank you so much for these insights and your effort answering my question.
I actually read about the sine wave topic in Earl Geddes's informative book about Home Theaters which was recommended here:

https://www.audiosciencereview.com/...roofing-a-room-in-basement.61654/post-2268996

but unfortunately couldn'd make the transfer.