Audio signals in a computer are processed in real time. This is the case regardless of whether you use JPLAY and memory playback engines, or any method.
The first common misconception is that data on the computer is just ones and zeros. This is not the case. The audio data is a square wave, which is an analog representation of a digital signal. This square wave, a pulse width modulation, contains both amplitude and timing information.
Why does this matter when it comes to choosing between Mac and PC? Several considerations.
First, where does distortion come from in the computer?
Because audio is a real time process, the square wave in the computer is duplicated and modified based on software algorithms thousands of times. Each new "version" of the square wave is generated from voltage in the power supply.
Noise radiated from the computer and noise from the power supply, specifically high frequency noise, gets folded over into the square wave. This introduces harmonic content that shouldn't exist in that signal and results in the digital sound many hear in computer audio.
The longer it takes the computer to process the music, the longer the signal can pick up noise and the more amplitude distortion that occurs.
The next misconception is that the computer doesn't matter if you are streaming over the network or if you have a high-end DAC that fixes Jitter. This is not true.
While it's possible to fix timing information and jitter, it's impossible to fix amplitude distortion on the square wave. This amplitude distortion becomes the square wave and no amount of filtering will do anything besides create other forms of distortion and coloration.
Even if you are streaming to a network player or DAC, all you have done is removed the cable. This can, of course, reduce noise, but it will not eliminate your source as a vital component to your sound.
The signal created in the computer must still be processed before being sent down the network. Once the data hits the network output it is converted into packets, which are buffered. These packets are then sent off to the DAC or network player for further processing. The idea of a network player makes little sense to me as the signal is then being processed by two computers instead of one. Digital doesn't like handshaking.
Ok, so what makes the computer process data faster and what impacts the sound of the computer?
The CPU clock speed means almost nothing. Modern CPUs process data SO quickly that there really isn't much difference between 2ghz and 3.5ghz. Sort of, there are a lot of other differences that play into those clock speed.
First, and probably the most important factor in computer/audio performance is the amount of L3 cache in the CPU.
This cache interfaces with the memory controller and determines how quickly the CPU can PREFETCH data for processing.
Here's the problem, the CPU must load data into memory for processing. That means it will always be limited by the memory speed and bandwidth.
Memory speed is limited by its CAS latency, which is how many CPU clock cycles must pass before a new string of data can be loaded for processing. Lower, of course, is better.
When the CPU must wait for data to be loaded into RAM (and this is constantly), the CPU incurs a wait-state. The wait state is exactly what it sounds like, the CPU is waiting to process. The longer it waits, the longer the square wave floats and picks up noise and distortion.
Comparing the Mac Mini and Zuma...
The newest Mac Minis can have up to 6MB L3 cache, whereas the Zuma (being that it's just a computer) can have anywhere from 6MB on up depending on the CPUs used. For instance, a 2GHZ Xeon with 24mB L3 cache will sonically outperform a 3.5GHZ Core i7 with 8MB L3 Cache.
So the PC Route has better CPU potential, and I believe the CPU used in the Zuma has 8MB L3 cache.
Power Supply...
This is probably the most mission critical element of music server and computer audio performance.
Using a Linear power supply on a mac mini is not the same as using a linear power supply on a PC like the zuma.
The Mac Mini operates off a 12V power supply where as a computer operates off an ATX Specification power supply. The ATX Specification requires (simplified) 12V, 3.3V, 5V, 5Vsb, -12V and various logic circuits to function.
With a mac mini it's as simple as using an external linear power supply. With a PC, that same linear power supply faces a few more obstacles to the best performance.
With a PC, if you are using a 12V Linear Supply, you will be limited in several ways. First, that 12V supply will have to go through a DC-ATX converter, which is a switching power supply, to generate the additional voltages. This will limit your hardware selection due to current limitations and be exceptionally noisy even with the best DC-ATX converters and linear supplies.
If someone was to compare the same linear power supply (assuming it's perfect) on a modded mac mini vs a PC with a DC-ATX then the Mac Mini will win.
In my experience, the only way for a PC to outperform a decked out Mac Mini is with an ATX Specification linear power supply.
Comparing stock Zuma vs Stock Mac Mini...
This is a no-brainer... the Zuma is a more powerful machine running a more customizable operating system with more upgradeable hardware.
Comparing Upgraded Mac Mini to Stock Zuma...
The Mac Mini, with the right design, will significantly outperform the Zuma.
Comparing Zuma with ATX Linear Supply and fully modded mac mini...
The Zuma will significantly outperform the upgraded Mac Mini.
I have a lot more information about this on my website. Hopefully this will help you make a decision between the two units. Both have a lot of potential and are quite good.
Courage,
Ryan
Core Audio Technology
