On one side, you have a PC running its task from a CPU, which granted is very powerfull but also meant to perform a whole range of processing tasks. Audio being one of them.
On the other hand, you have a dedicated processor (called Digital Signal Processor) like our miniDSP kit, dedicated to running only one type of operation, audio signal processing.
What's the difference between the two? The main drawback, that even the most powerful PC will struggle to come around is the minimum overall latency.
What's latency? As a general term, it's the overal time it takes for the signal to travel from the input to the output. Taking the case of an analog input, it first needs to be converted to digital by an ADC, then the signal being processed by a processor (CPU or DSP), to finally get converted back to analog by the DAC.
The major issue in a PC environment using a filter solution is that a couple of variables needs to be added to this path. The audio drivers (even with Asio at best 5 to 25ms), the processing by a software running on an OS that may not be a Real Time OS, and finally feeding the signal back to the sound card using the driver again all add up together. The end result is an overall latency potentially streching to the order of 100ms or more. Compare this specification to the miniDSP fixed overall hardware latency (input to outputs, inc ADC/DAC + processing) of about 1.5 to 2ms (depending on the plug-in), it is clearly obvious that dedicated processing is the way to go for minimizing latency by a 100x order.
Hope this explanation makes sense and feel free to send us your comments on the best latency you may have achieved recently using filter software applications. The information listed here was written to the best knowledge of the experience we had over the years. With technology going so fast, who knows what it