• Channels: 2, 4, 6 (5.1), 8 (7.1) or 10 (8+2).
• Resolution: 16-, 18-, 20- or 24-bit.
• Sampling rate: 44.1, 48, 96 or 192 KHz.
• Signal-to-noise ratio (SNR): 80 to over 100 dB.

It is important to notice that several codecs offer one value for its inputs and another value for its outputs – for example, offering up to 192 KHz sampling rate for its outputs but only up to 96 KHz to its inputs (which is the case of Realtek ALC888S codec portrayed on Figure 8, by the way). This happens because while all users use the motherboard audio output and can judge audio quality by listening to audio produced by his or her PC, only a few use the “line in” input offered by the motherboard (several users use the “mic in” input for applications like Skype but since this input is used only for human voice, the quality doesn’t need to be the best). Based on this fact codec manufacturers provide cheaper codecs with lower specs for their inputs, which are chosen by motherboard manufacturers in order to cut costs. In fact you will only find codecs with high quality for their inputs on a few very high-end (and very expensive) motherboards.

By the way, on the codec datasheet or specs page on the manufacturer website the outputs are usually referred as “DAC” and the inputs are usually referred as “ADC”. This will help you to find out the specs of a given codec.

Let’s now explain in details each item from our list above.

Channels are the number of independent outputs a codec offers. In the past sound cards were only able to provide two channels – left and right, i.e. standard stereo sound. Then came the basic surround sound with four channels, with two speakers (left and right) on the front and two speakers (left and right) on the rear. Nowadays these two options are only found on very old motherboards or on very low-end ones.

The minimum number of channels you will find today is six, also known as 5.1. This is the same audio standard used by DVDs. Here we have two front speakers, two rear speakers, one center speaker (mainly used for voice; for example, on news broadcast or on a movie when the characters are talking the sound will come from this speaker) and one subwoofer channel – also known as LFE (Low Frequency Effects) –, which is in charge of creating a more realistic experience by reinforcing low frequency (bass) sounds.

Eight-channel audio, also known as 7.1, provides the same specs of 5.1 audio, but adding two more speakers between the front speaker and the rear speaker, one at each side (left and right). These speakers are also known as middle speakers. This is the same standard used by high-def discs, i.e. HD-DVD and Blu-Ray.

And finally the latest addition is the 10-channel audio, which is a misnomer, as it doesn’t add two more channels to the surround system but provides two independent channels (left and right) for headphones. This system is better described as 8+2. On all other systems the headphone jack located on the frontal panel of the case is connected in parallel with the line out (front speakers) output and thus provides the exact same sound that is being sent to the front speakers. On the 8+2 system the frontal headphone jack is independent from the front speakers output and programs can send a different audio signal to the headphones while playing another sound on the frontal speakers. Think of games that use VoIP technology like Battlefield, where users can talk to each other thru a headphone/microphone set, and you will quickly see the benefits of this system: users won’t listen on the headphone the sounds of the game like explosions and gunshots, only their conversation with other users, while keeping the game sounds playing at the same time on all speakers.

Resolution and sampling rate are two parameters that would take us a long time to explain exactly what they are. We have already written a tutorial called How Analog-to-Digital Converter (ADC) Works where we explained them in details. If you don’t want to read this tutorial all you need to know is that the higher these numbers, the better the audio quality is.

And finally we have signal-to-noise ratio, also called SNR, which measures the noise level produced by the codec. In our opinion this is one parameter that average users don’t pay a lot of attention, but they should.

Keep in mind that what we are talking here is the noise produced by the codec, not the noise that is already embedded on the audio source. This parameter has nothing to do with the noise that is already inside a video file or a MP3 file that you may have.

To understand what we are talking about, make a simple experience. At home, find a sound system that still has a tape deck. Put a tape to play, press the pause key and increase the volume. You will hear a lot of noise (technically called white noise) that is being produced by the tape deck and isn’t inside the tape you are playing (as the tape is stopped). That is the noise we are talking about. Of course you don’t want a motherboard that inserts noise on your sound.

The signal-to-noise ratio is measured on a unit called decibel (dB). The higher this number, the better. Ideally you should have a motherboard with a codec that has a signal-to-noise ratio of at least 100 dB for the best audio quality possible. Mainstream codecs offer a signal-to-noise ratio of 95 dB or 97 dB, which is “almost there” and will provide a very good audio quality for Average Joe.

The problem, however, is the signal-to-noise ratio for the audio inputs. The signal-to-noise ratio advertised by the motherboard or codec manufacturer is for the codec analog outputs only (DAC SNR). Like we explained, motherboard manufacturers will usually pick a codec with input specs lower than output specs.

If you plan to professionally work capturing and editing analog audio – converting VHS tapes to DVD, converting tapes and LP’s into MP3 or CD and similar tasks – then you should buy a motherboard (or an add-on sound card) that uses a codec providing at least 95 dB SNR. Below that you will hear noise on your work, which is bad as it will give you a bad name with your customers. Unfortunately mainstream codecs will provide an input SNR of 90 dB or even 85 dB, which is really bad. Of course you only need to worry about this if you are going to work capturing and editing analog audio. If this isn’t your case – like it isn’t for almost all users – you should not worry about this.

Now that you know the basic specs of a codec you will be able to compare the audio quality of different codecs (and thus motherboards) available on the market. To make things easier, we compiled a series of tables listing the basic specs of the most popular audio codecs found on the market.