A computer is a good source. The interface it provides gives a lot of control over what can be done. HDMI or S/PDIF provides the option to output to a superior A/V receiver. Using USB as an intermediate only decreases the bandwidth capabilities of the digital data. Its DC power source is provided by the very computer being used. Any highly regarded filtering circuits or ICs it would have would already be on a PCI or PCI-e device. Built in audio uses either of those buses without you knowing it. The interconnect is a pin interface, it can also be traces on a PCB. Whatever "god-like" circuitry an external box has does not have to just be tied to USB. There is
I2S to make it all the same in any type of device.
The person in that video clearly knows nothing about microprocessors and digital signal processing. His
"simplistic" speech is not only breathtakingly wrong, but if what he says were true, the product he's pushing would have the same flaws. I've never witnessed technology this oversimplified before that it's discussed as if it were Flintstones technology with a temperament and catch-phrases.
Also I should point out that CD players are computers. They use electrons through semiconductors in order to operate. They use integrated circuits which process binary data(digital logic). They have
digital to analog converters which must
start with
digital data in order to become analog. If the CD player is using S/PDIF instead of its own DAC, then the whole bitstream is kept digital until it reaches the other device. There aren't tiers quality for that process.
I think this whole topic can only be discussed validly by those who hold an advanced degree in electronic engineering. Everything is just going to be hearsay. A discussion can't be valid if someone just takes a term they think sounds technical enough and claims whatever they desire about it. Here is a datasheet. http://focus.ti.com/lit/ds/symlink/pcm1795.pdf Don't understand all 58 pages? Then how can someone like that be seen as an authority on this?
Just because engineers and physicists design this stuff doesn't mean that it's all a complete failure because they didn't "put their heart into it". ICs are mass produced, not hand carved by a carpenter. They'll be well aware of what would interfere with the signals in their device and know how to compensate for it and merge it into a larger device. I think too many people have had simple computer operator errors and like to blow their problems out of proportion. None of these things can be simplified