Please Help!! looking to get into computer audio.

To make some terms more clear, and take them out of vendor-speak we are talking about the differences between Asynchronous USB (A-USB) and Asynchronous Sample Rate Conversion (ASRC).  Usually vendors will call ASRC something like upsampling and re-clocking or something like that.  The Wyred4Sound Remedy Recklocker is an example, but many DAC's have this feature built in. 

Both help to reduce jitter but the similarities in how they work and how well they may work in any given circumstance are not the same. Jitter from an A-USB source is often as low as the DAC that is pulling the data, and is bit-perfect.  That is, the DAC receives exactly the bits in the source file, and the output signal is evidence of the very best the DAC can do. There's very little to do to make a no-compromise solution with this interface. 

As you are expecting, ASRC will also improve bad sources, reducing measured jitter somewhat but at a certain point, based on it's implementation, it has to give up. This can be improved upon by bigger input buffers, but it usually does not yield the same level of improvement, especially with things like internet radio, iTunes, Chromecast, etc.  The worse the input signal (i.e. more jittery) the worse the output signal. By contrast, A-USB can tolerate a lot of lag in the input before ANY of it becomes measurable.  Further, ASRC is not bit-perfect.  The entire output signal is a mathematical reconstruction of the input. For fun reading on this look up some of the writing around the closed-form digital filters used by Schiit. Of course, they are all pro-Schiit, but they offer a great education in how ASRC is accomplished, and what kind of compromises are made.  In general, ASRC prefers time precision to bit-perfection. It will sacrifice the data to meet it's time goal, as much as it can. 

Of course, the devil is in the details, and anyone can make a crappy version of any technology.  You can also chain one after the other, yielding very good results at times.

A better option, in my mind, is the latter Bryston DAC's with built in USB 2.0 A-USB support AND upsampling.  Given the option for just 1, I'd take A-USB any day of the week. 

Best,

Erik

Check out the free, open source, Banshee music server.  Looks a lot like JRiver and it's available for Linux, Windows and Mac.  Works fine with flac format, which is what I use.

Did I mention that it's free?

http://banshee.fm/

Erik - Care to comment on the specifics of the reclocking that is going on in the Bryston BDA-1, the DAC in question. I think you are mixing reclocking with upsampling. The BDA-1 reclocks then upsamples. Two different operations.

The jitter can never be better than the clock in the DAC. You somehow think that the timing is better when applied in async USB mode than when implemented internally in the DAC by reclocking. That is just not true.

Please, address the specifics of the reclocking that is going on with the BDA-1. Are you really saying that if the BDA-1 used aysnc USB then the jitter would be lower than using its internal clock to reclock the data?

And, reclocking is bit perfect. It simply takes the data (2 16 bit data points for each sample) and sends that sample on with the timing of the internal clock.

Your discussion of Schiit filters is an entirely different topic than reclocking.

" The worse the input signal (i.e. more jittery) the worse the output signal " Clearly you do not understand the idea of reclocking of signal as is going on in the BDA-1.

Honestly, I think you are throwing around a lot of terms without really understanding the process that the BDA-1 is using, especially the concept of reclocking.

I kind of hate JRiver, and only moderately despise MediaMonkey.  Something about UI designers who are into music.....they are very odd fellows indeed I think. :) 

So, no JRiver clones for me. I'll have to stick with Logitech Media Server and various tagging/retagging tools I could find.

Best,

Erik

Without knowing the actual implementation, they could use a large-ish buffer and then upsample, but these issues have the same limitations as ASRC. The re-clocker has to constantly try to guess the long term stability of the original source clock AND the buffer has to be big enough to handle the difference between the source clock and the DAC clock.

This is why ASRC is actually pretty easy to implement.  You just say F* it to the original clock completely and really only need a very small buffer. 

My guess, is that they are just doing ASRC, like everyone else these days. 

Again, the best solution among the now 3 different types of implementations is Asynch-USB. 

Try this thought experiment.  Let's say your source is feeding out samples every 1/44,110 of a second instead of 1/44,100 of a second. The only one of the three methods that is immune to this problem entirely is Asynch-USB. The source clock has nothing to do with the frequency of data fed to the DAC (within reason, if the PC is running Windows and it's in the middle of an update, chances are it's all going to hell). 

Parasound's CD player, and some digital players have implemented memory-only players. They read the entire file into memory at once, completely doing away with issues of source jitter. There's no evidence it's better than Asynch-USB though. 

PS Audio has a useful article on this too somewhere, of course it's all pro-PS Audio's dac or something. :) 

Best,

Erik