Guest Post: Digital to Analog Conversion Basics

A DAC or DA converter (digital to analog converter), is essentially the outputs of your sound card. When you read about monitoring in a studio environment you see a lot written about loudspeakers and amplifiers but digital to analog conversion is less frequently discussed. Of course we also have ADC which is analog to digital conversion which relates to the recording process. In this context we will focus on the DAC as everyone will be using those, including electronic only musicians.

Even the cheapest sound card has a DAC chip inside which converts the digital data stream into an audio signal which can then be amplified to a level which is appropriate for line level. There are many different DAC chips that can operate at various bit depths and sample rates. Some common ones are made by Cirrus Logic, Analog Devices, Burr Brown, Wolfson, and AKM. DAC chips are released fairly frequently and each development tries to improve on cost and/or sound quality.

All semi-pro sound cards/interfaces have a DAC chip for each of outputs available (often one per stereo pair), and usually the more you spend, the better the specifications of the audio devices within the audio interface. It is possible to pay $2,000.00 for a mere 2 channels of high end stereo L and R output or a mere $20.00 for a SoundBlaster.

Apogee Symphony IO is one of many well respected AD/DA converters for professional studio use

Do DACs sound different and why?

The specifications of a DAC chip give an idea of the technical performance, the kind of specifications you can expect to see are: dynamic range; THD (Total harmonic distortion); noise levels; cross talk; and frequency response. When you purchase a sound card you will get a specification based on all the components in the circuit including the DAC chip so the specs here will differ from the specs of the DAC chip which has been employed.

The resultant sound from a DAC is a complex combination of the chip used, clocking circuitry, power supply, filters and the analog amplification stages used on the way to the audio output. The circuit design and layout can have a significant bearing on the sound of even the same DAC chip. So a design can use the same parts but sound different. Design of a competent DAC is a highly technical and skilled electronic profession.

It is worth stating that specifications of a chip or sound card only give you some basic details of expected performance. The reality of the situation is that the subjective sound can vary wildly.

I have a music recording room at home and did a few experiments with changing some of the audio opamp chips (an analog device which buffers/amplifies the analog output stage of the DAC chip). When I tried listening to the outputs with different opamp chips in them, the differences were nothing less than startling. The original chips were JRC2068 and I inserted a fast and highly specified chip I have had very good results with in the past. By inserting this chip in this given circuit the sound became so wildly different from the original that I considered it unusable. The stereo image had widened considerably to a point where I could not actually trust it. So a device that has previously sounded great in one circuit can easily not work well in another. This is not condemnation of a specific opamp chip but it outlines the potential for one component change to radically change the sound of the DAC.

A word about clocking

As a general rule, any given DAC can never be technically superior when clocked externally. Inside a sound card there is a clocking circuit which is locked with very high accuracy to a crystal chip. As long as this circuit is well implemented it should provide the ultimate timing accuracy for the DAC clocking, keeping jitter to a minimum. This is contrary to the experience of some people, they clock a DAC externally and are certain they are hearing differences that amount to an improvement. Well I will leave that for you to decide ultimately as this is a rather subjective area. But as theory dictates a crystal clock is more stable than the clock produced by a PLL (phase locked loop) circuit. I will however add some extremely costly DACs have circuits which can supposedly reduce external clock/PLL jitter to very, very low levels by using a “re-clocking” circuit. My own experience thus far is that it is preferable to use the internal crystal clock unless you need to synchronize multiple digital devices via AES/EBU or word clock. There are often practical situations that mean external clocking is inevitable but fortunately most higher end DAC’s provide circuits which counter any externally born jitter.

I would say that when it comes to monitoring, the most important factors would be loudspeakers, acoustics, amplification and then digital to analog conversion. However I do not think digital to analog conversion can be ignored because there are marked differences in sonic quality which may not be revealed by looking at a specification sheet. That includes high frequency smoothness and detail, stereo image, transient accuracy, highs, mid range and low end reproduction and even perception of front to back depth. (i.e. where instruments sit from front to back of a sound stage)

If you are considering spending a small fortune upgrading your conversion, it would be wise to ensure that you can try out the equipment before buying it outright. Maybe you can work an “on approval” deal out with the retailer on the basis you will purchase the right equipment once you have heard it.

Thanks for reading. Hopefully this has sparked some interest in this lesser discussed technical aspect of monitoring. If you have any questions or comments you can enter them below.

Barry Gardner operates SafeandSound Mastering a low cost high end online mastering studio. SafeandSound Mastering

3 thoughts on “Guest Post: Digital to Analog Conversion Basics”

  1. Thanks your article was interesting i have an older dac with a very beefed up anolog output stage. I have some newer digital players that still benefit from running through it.

Leave a Reply