In the beginning of era of integrated sound cards, they weren’t for any recommendation because of many troubles. Even though the sound quality on certain models was satisfactory, by purchasing quality speakers, you could notice difference in sound quality compared to quality nonintegrated sound cards. There were even some relatively “bad” models of C-media nonintegrated sound cards with better sound quality. Beside sound quality there were also some problems with too big noise, as also sound interruptions during reproduction of audio contents (if someone remembers well known “integrated” PC-Chips motherboards in their own time).

Situation wasn’t much better for video games because most effects were done by software. So every time when sound was emitted or when you set sound on better quality, performance decrease in games happened. All this was happening for years, until Soundstorm was launched. Soundstorm showed that integrated sound cards can give a lot. Today’s situation is for certainly much better and many negative effects are reduced or completely uprooted. Still, the question remains: Are today’s integrated sound cards sufficient for basic work and how much are they weaker from nonintegrated models? And that’s what we wanted to know.


A little bit of theory

Several conditions need to be fulfilled so that one sound card could produce “quality” sound. Good characteristics of some sound card don’t necessary guarantee sound that you’ll like but only that sound will be as much as “pure” as possible. First we’ll start with frequency response that shows how signal amplifies in correlation to its frequency. In ideal conditions, amplifying should be independent from frequency, it should be constant. Still, this will go only for range of our interest and that’s 20Hz – 20KHz (audio range that human ear can hear). If the frequency response is good, than the signals from any frequency in audio range will be amplified similar, which means that we should expect almost flat line around zero (completely flat line in “0” would be ideal but that’s impossible).

Next very important factor is ratio signal/noise which tells us how much is signal “stronger” from noise. This ratio can be measured and valued in dB (decibel), and presented with logarithmic formulas:  10log(Psignal/Pnoise) or 20log(Asignal/Anoise) ( P-power, A-signal intensity). We mentioned this because people often think by mistake that, for example, sound card with ratio signal/noise 80dB has 25% more noise than the one with 100dB. The truth is for this example, that better audio codec has noise intensity weaker for 100x and noise power for 10x. In our case we expect ratio under 75-80dB for every frequency of our interest and ideal one will be endless, which is impossible to get. Not less important factor is THD (total harmonic distortion) that represents ratio of the sum of the powers of all harmonic components to the power of the fundamental frequency. If we have signal from 1KHz than because of imperfections will appear weaker signals on integer multiple of the fundamental signal. These frequencies (2,3,4KHz…) are harmonics and they bring distortion into signal. The point is that the amount of these harmonics should be as less as possible, meaning the THD should be as much as close to 0 (THD=0 is ideal case and it is not possible). Ratio THD + N (N – noise) represents noise power added to power of those signals that create distortion. So, this factor is very important for average user. This factor also is defined in dB, and it always has smaller value than signal/noise ratio. Also there is unwanted intermodulation distortion (IMD) that creates inharmonic tones. Of course the less value the better. In our test we presented to you graphics for signal/noise ratio and (noise+THD)/signal ratio with negative values of the same absolute values. That actually means that audio codecs with values far from “0” have better characteristics.


Rightmark test (theory applied in practice)

To determine values for all parameters that we mentioned above we did tests in newest version of Rightmark 6.1.0. We used for this test armored cable of big length to reproduce as much as possible real results (cable from speakers is already long enough). In test for frequency response Creative X-Fi scores a victory while ALC889a speaker system showed as best from all tested integrated speakers (results can be seen in table). The worst performed ALC662, as we expected, with the weakest characteristics. In this test ALC885 disappointed us even though it has the better characteristics than ALC888. So the lesson is: In case of Realtek’s integrated sound cards bigger mark doesn’t bring better characteristics/performances, and that was the frequent mistake that buyers made while selecting motherboards.
For signal/noise ratio, situation is similar, only now ALC885 is better from stronger ALC889a. IDT 92HD206 turned as the best from all integrated cards, and this one is usually integrated on ECS motherboards but its (noise+THD)/signal ratio was the worst on test along with ADI 2000B. In other performed tests situation was similar, but always X-Fi was the one that wins, which was somewhat expected. 2000B fell on test even though it is the best ADI chip but still it performed better than ALC662 and in range with ALC1200 which was still weaker from ALC88x series. ALC885 had the best (noise+THD)/signal ratio, it also had the biggest intermodulation distortion which was noticeable even on listening, but it didn’t affect much on sound quality. Definitely, in this test ALC889a was the best and none of its parameters looked that bad.

Optics & Wire In case of coaxial cable, S/PDIF signal is transferred through cable and receiver can properly identify signal if disturbances are small enough. For optical S/PDIF, digital signal transfers optically – if there is light “1” is transferred, in reverse “0” will be transferred. Transfer through optics is much safer way because it is more resistant to potential disturbances from surroundings.