Digital to Analogue Conversion, DACs: How Digital Audio Quality Can Vary Between Devices
When sound is first recorded it is it is captured using a microphone, which is an analogue device. The sound waves are made into analogue waveform, which is then later processed into binary digital data for storage and transmission.
The analogue to digital conversion process works by repeatedly measuring the analogue waveform, then outputting these measurements as a long list of binary data, to create an image of the sound.
The quality of this conversion is determined by two things; how frequently these measurements of the waveform are made, and how accurately they are measured, also known as the word length.
Each measurement of the sound wave is called a sample. The sample rate or frequency is the number of times each measurement is made per second. With CD quality the sample rate is 44.1KHz, which means that the converter has measured the sound wave 44100 times each second.
The sample rate of a recording is consistent, it does not vary throughout a single recording. The higher the sample rate, the better the quality, but the more data is required to store the file. With a lower sample rate, you loose quality, but have a file which is smaller and easier to store or transmit.
The accuracy of each measurement is known as the word length and measured as a Bit rate. The Bit rate corresponds to how many binary numbers make up each measurement of the waveform. For example, a 4-Bit sample might look like this: 1111, this: 1010, or any four digit combination of 1s and 0s. There is a maximum combination of 16 four digit numbers which can be made, taken from a 4-Bit measurement.
A 8-Bit sample is made up from a combination of eight 1s and 0s, for example: 11110000. This allows for 256 possible combinations.
A 16-Bit sample is made up from sixteen 1s and 0s, (ie: 1110011000110001) allowing for 65,536 possible combinations.
Imagine for a moment that you have to continuously physically measure the analogue sound wave with a special ruler. Then note down your measurement (or sample) on a piece of paper, for playback later. You have three rulers, each of the same overall length. Your 4-Bit ruler has 16 graduations spaced out along it. Your 8-Bit ruler has 256 graduations and your16-Bit ruler is divided into 65,536 graduations.
When you measure using your 4-Bit or even your 8-Bit ruler, it is very rare for the sound-wave to actually come into line with a measurement on the ruler. You have to round-up or down to get the nearest sample. This is noted down on your paper, but when it is replayed later it does not sound exactly as it did to begin with, it is the nearest sound.
The samples you made with the 8-Bit ruler are more convincing then the samples made with your 4-Bit ruler, as each measurement was much closer to the original point on the sound wave. When played back, the results from the 4-Bit ruler are rougher and more distorted then the 8-Bit.
The samples made with the 16-Bit ruler provides a sound which is considerably closer to the original and far less distorted, as the measurements are made far more accurately.
CD quality is 44.1KHz/16-Bit. This means that each second the converter will produce 44,100 16 digit numbers, each number being a sampled measurement of the original analogue waveform.
There are High Resolution Audio formats available these days which are superior to CD. Even though CD quality is very good. High Resolution Audio is often 192KHz/24Bit.
So music is recorded as an analogue waveform and the Analogue to Digital Converter makes it into data. The data is then stored and distributed, then your Digital to Analogue Converter reads the data, converts it back into an analogue waveform. This waveform is then amplified and made back into sound-waves by your speakers.
As we’ve explained, the accuracy of the Analogue to Digital conversion is an important factor to the sound quality. So to is the accuracy of how that data is then read and converted back.
Not all Digital to Analogue Converters (DACs) are created equally. The DAC which powers the speakers and headphone socket on your phone, costs a couple of pence to manufacture and is capable of nothing close to the performance of a DAC which costs hundreds or thousands.
There are several factors which can effect the performance of a DAC and the resulting sound.
Firstly, a basic DAC might not support the full range of file data rate. Your cheap DAC might not be able to process 24Bit/192KHz data, or even CD quality 16Bit/44.1Khz.
Then there is jitter, or digital timing errors. Remember how we said that when a analogue signal is measured, the samples are made with consistent timing? At the stage of Analogue to Digital conversion an accurate clock is used to time when these samples are made. You need a clock which is equally as accurate in your DAC to ensure that the readings are made with the same consistency. A cheap DAC is also likely to introduce unwanted noise to the playback, due to poorly designed circuitry.
It is worth being mindful of the quality of the DAC in products you are using. Many products such as CD players, computers, TVs and streaming devices have the means to output their audio digitally. This makes it possible to by-pass the internal DAC and instead use a stand-alone DAC of a higher quality.
If you are looking for the best audio performance from a variety of sources, it might be worth considering a separate DAC to add to your HiFi rack.