CAS 2: Audiophile Myth

(1) Chapter One Review:

Last time we talked about how digital audio works. In every second of CD playback, there are 44,100 digital samples, which contain 16bit word-length (00000 to 65535). A lot of data is read from the disc, and these data can be perfectly transfer from one media to another. For example, we just post one of our upcoming album’s song “Lush Life” for audiophile download and preview free of charge. The DXD version is over 600MB. Every second of this particular file carries 384,000 digital samples, that contains 24bit word-length each. A Canada audiophile took 2 hours to download this file to his harddisk, When we compare the file in his harddisk, with our server, they are 100% identical. 


(2) Okay, if things are such simple and perfect, every digital source should has the same sounding?

Things are more complex than this, however we shall not forget the principal of digital audio existence. If the data is kept in digital domain, you can PERFECTLY clone, compare and verify.  When you extract the data and playback through a sound interface, the interface will more or less affect the playback quality. 


(3) That’s starting more likely an audiophile article rather than plain pure boring digital theory.

When digital data is read and send from a soundcard through normal digital audio standard connections (AES/EBU, SPDIF, TOSLINK) it is necessary to include some extra information (how many bits, what sample rate) with the pure binary code, otherwise the DAC will not know how to process the incoming data. The timing information error becomes a very well know audiophile term “jitter”. 


(4) Jitter = timing error

Timing information for digital audio is like a music conductor in musical term. It controls the timing between different sections. Let’s say all members in an orchestra play perfectly with their own parts, but a bad conductor completely messes up with their timing, the orchestra performance will obviously poor.


(5) How can this timing information create?

Another name of this timing information is called “Clock signal’. In most digital devices, crystals are installed to generate this timing (clock) signal. A better crystal can provide more accurate timing information. However, we must be careful that the accuracy of crystal does not show the whole picture. The problem most of the time lies on other areas such as power supply, temperature, clock signal path etc. 


(5) Will that show a soundcard can never be good source compares to my ultra expansive CD transport?

That depends on how we define what is a “good source”. Moreover, a lucky thing that jitter can be recovered and cleaned by good de-jittering devices and DACs. In all DACs, after it received the incoming digital data and timing information, it will match the clock input with its own reference timing, to form a pool. This is a basic PLL (Phase Lock Loop).


(6) Does it mean there is another clock inside the DAC?

Yes, DAC actually has 2 clocks (crystals). One is based on 44.1kHz, the other is based on 48kHz. If you meet a problematic DAC, try lock it with different sampling rates. If the problem exists on sampling rate 44.1kHz, 88.2kHz, 176.4kHz, and working fine with 48kHz, 96kHz, 192kHz, you may have a bad crystal.


(7) Would you suggest a better DAC, or a better CD Transport? 

Good DAC can clean up incoming jitter from bad sources. It is hard to define the term “good” Transport. First of all it’s a pure digital output device, the data should always be the same. If we talk about the jitter output, even if it provides perfect shape, it will(can) be alternate by the DAC clock/pool. Jitter can be cleaned/polluted in between the whole signal path. What’s matter is how accurate this timing information reaches the digital audio conversion point. (That’s means the clock input of DAC chipset) There were audiophiles who chained multi de-jittering boxes together a head of a DAC. If the DAC PLL circuit is poorly designed, jitter can be even higher than before it takes any de-jittering stage. 


(8) Does transport mechanism vibration, power supply affect the jitter performance? If they do, then computer must not be a good thing for audio playback. 

Yes they can affect the jitter performance. Again a good DAC can clean the incoming jitter. Computer is standard machine with standard power supply. The soundcard design is rather more directly related issue. Please do not custom made MIT caps for your 450Watt ATX power supply. This can not improve the audio performance. 


(9) Can a faster CPU, newer graphic card, more ram or faster harddisk affects audio performance?

They do not affect the standard stereo playback quality. Stereo audio playback demands very little resources with today computer standard. Audiophile should not worry about the motherboard, CPU and computer performances. In recording, this case is rather different. The channel limitation of how one digital audio workstation (DAW) can recorded/playback is determine by the computer power (or specialized DSP soundcard such as Protools/Pyramix/Sadie). 


(10) What about digital dropout that I heard from computer playback?

When a soundcard is playing/recording audio, it needs buffer to store a small amount of data. The larger the buffer size, the longer the latency. The buffer prevents audio dropout. The higher performance computer can reaches smaller buffer size, hence shorter latency. This lightly affects audiophile 2 channels playback. In studio recording/mixing world, multichannel playback/record latency becomes a critical elements. On Windows PC domain, there is a DPC latency checker that analyses the capabilities of a computer system to handle real-time data streams properly. It may help to find the cause for interruptions in real-time audio and video streams, also known as drop-outs. 


(11) Is that Ram buffering? There are programs which store the entire audio song(s) in the computer ram and playback from there. Can this improve audio quality because of its isolation and no mechanical movement? 

The audio data will be sent out by the soundcard, which determines the clock signal accuracy. The data storage location has no effect on audio quality. The latest SSD ram drive will not improve audio quality over SATA hard disk. 


That’s it for now.

Next time we will focus on newly development computer interface, USB, firewire IEEE1394, and take a closer look to various clocking methods and playback software/processing. If you have any question would like to discuss, please feel free to tell.