3. Is there any way to extract PCM data from a PC better (e.g,, least amount of conversions, cleaner, etc.) than via a USB stream? Not that I disagree with Mike's research/opinions, but if the source is a PC, then I am not sure there is a better option than USB stream.
PCM is digital data - so during transport via USB there is a pretty significant amount of error correction going on if any of the data doesn't reach its destination will be retransmitted. (http://en.wikipedia.org/wiki/USB#Data_packets) To be very clear a packet will fail the CRC check if even one bit is flipped, the data must be 100% the same. Generally if your USB cable is good your going to basically get 100% success rates for data transmission. If you are having frequent errors it will manifest its self in very very obvious ways such as device disconnects.
If you want to be very technical you can read about http://en.wikipedia.org/wiki/Pigeonhole_principle you can understand how it is possible for data to be invalid and pass the CRC check. The CRC is a 16-bit hash.
The real challenge for data transmission however for what is in essence a real-time data application is clocking - When is it appropriate to covert a piece of data into an analog signal. This is why we read so much about re-clocking in DACs and Transports. If your unable to clock the data correctly it doesn't matter how good your PCM stream is, the music is going to sound like crap because the timing won't be right.
Now you may ask if USB is so great why do we need a Wyrd? Your right it is odd - the short answer is power. For data to be transmitted through an medium it must have a carrier - in this case its an electrical signal. In general we have a baseline voltage and then a higher voltage, the baseline is 0 while the higher is 1. If there is a power fluctuation in the USB cable during transmission this is what can cause a bit or bits to flip causing missing data resulting in distortion. Again in normal situations such as a USB drive the receiver will simply not ACK(Acknowledge) the data packet and it will be retransmitted by the sender but since this is real time that data can be ignored.
But wait!
There is yet one more challenge when it comes to digital to analog conversion. As electricity moves through mediums such as copper, silver, or gold it does not move at the speed of light - it moves at some fraction of the speed of light know as its Velocity Factor. If a very high quality DAC the creators will want to match the speeds of the various feeds to output at the exact same time so that the phase of each channel or even frequency(depending on how the DAC is setup) are in line. Generally this is very close but those with more critical ears may hear the difference.
The end up however is really - what we are paying for in a DAC while it is the DAC chips and everything that goes along there we are also paying for the analog portion of the DAC that outputs line level into our pure analog components. A DAC with a PCM9001UberMinsc with a crappy analog stage could sound just as bad as a DAC with Joes Bargin DAC with an awesome analog stage. Everything must match.
With the Yggdrasil since the analog stage outputs +4dBV we have a much lower sound floor which is part of the reason it sounds so good(to my ears)
The answer to your original question however is this. S/PDIF and TOSLINK both do not have error correction and as a result are unable to compensate or even know about any errors. That being said - again remember audio is real time so would you rather have slightly corrupted data into your DAC or missing data into your DAC? For me its simple - i'd rather have missing data. I can hear missing data much more easily than a single-bit error and make corrections to my system.
tl;dr: no not really
![20170826_185949[1].jpg](https://cdn.head-fi.org/a/9992700.jpg)
. 
