How to measure the quality of your sound card with Right Audio Analyzer

If you work with audio on a computer, sooner or later you wonder if the The sound card really performs that well. as the manufacturer promises. Luckily, you don't need to spend thousands of euros on laboratory equipment to find out: with programs like RightMark Audio Analyzer (RMAA), ARTA, or STEPS you can to measure quality in a fairly serious way of any audio interface, from the motherboard's integrated chip to a high-end dedicated card.

In this article we're going to see, step by step, how Measure the quality of a sound card using Right Audio Analyzer (RightMark Audio Analyzer), what the results it provides mean (frequency response, noise, distortion, crosstalk, etc.), how other components such as the GPU influence it, and what other complementary tools and tests you can use to get a complete picture of your audio system without going crazy.

What is RightMark Audio Analyzer and what is it used for?

RightMark Audio Analyzer (RMAA) is a test suite designed to evaluate the analog and digital path from virtually any audio device: internal sound cards, external DACs, players MP3Home CD/DVD players, headphone amplifiers, etc. Instead of using an expensive physical analyzer, RMAA relies on playing and re-recording test signals and then analyzing them with signal processing algorithms.

The idea is simple: the program generates a series of test signals (frequency sweeps, pure tones, tone combinations, noise, etc.), passes them through the device you want to evaluate, and then records the output. By comparing the original signal with the recorded one, RMAA calculates a set of technical parameters that describe the audio quality quite accurately.

One of the strengths of RMAA is that it offers both detailed charts and data (spectra, response curves, numerical measurements) as a rating like "Excellent," "Very Good," "Good," or "Average," which is helpful for those unfamiliar with the technical jargon. In other words, you can delve as deeply as you like, but also get a quick overview of how your card performs.

Version 6 of RMAA represented a significant leap forward, adding a A more user-friendly interface and new features of analysis. Among the most relevant improvements are new spectrum display modes (linear, logarithmic, mel), support for additional sampling rates such as 88,2 kHz and 176,4 kHz, a specific parameter for “Harmonic distortion + noise” in dB(A), a mono mode, polarity test and support for local languages ​​in spectrum labels, in addition to serious bug fixes.

How does RMAA actually measure the quality of your sound card?

The operating principle of RMAA is based on what is called loopback measurementIn practice, you physically connect the output of your sound card (usually the speaker output or line out) to an input on the same card (line in or, if there isn't one, a microphone input configured as a line input). The program plays the test files through the output and captures what comes into the input.

By having full control over the original signal, RMAA can Compare input and output to estimate losses and alterations that the device introduces: variations in frequency response, added noise, harmonic distortion, intermodulation, crosstalk between channels, etc. That is the basis of all the measurements you will see in the HTML reports it generates.

It's important to understand that, unless you're using two different interfaces, the loop test measures the quality of output and input of the sound card. If one side is clearly worse (for example, a good output but a mediocre microphone input used as a line input), the overall result can suffer. That's why, when fine-tuning, different input/output combinations are compared.

Another key nuance is that RMAA evaluates the technical quality of the device under specific conditions (gain, cable, electrical environment, GPU load, etc.). It's not measuring whether you "like" the sound, but whether the system introduces measurable noise, distortion, or irregularities. How your ear perceives those changes is another matter, but RMAA serves as an objective reference.

Parameters analyzed by RightMark Audio Analyzer

Each RMAA report includes a series of sections with numerical values ​​and a qualitative assessment. Understanding what each one means will help you to determine if your sound card is up to the task or if it's time to upgrade to a better model or an external DAC.

Frequency response (For example, “40 Hz to 15 kHz, dB”): This indicates how flat the audio output is within the specified range. RMAA typically displays something like +0.00, -0.07 dB. The closer to 0 and the smaller the range between the positive and negative values, the more accurate the reproduction. Values ​​like +0.05, -0.10 dB fall into the “Excellent” category because the variations are virtually inaudible.

Noise level, in dB(A): represents the level of background noise added by the system, measured using A-weighting (approximating the sensitivity of the human ear). A value of -100 dB(A) or less indicates a extremely low noiseIdeal for critical listening or recording. If it drops to -85 dB(A) or worse, you start to notice that the background isn't as clean, especially with sensitive headphones.

Dynamic rangeAlso in dB(A): This measures how much difference the system can handle between the weakest and strongest signals without distortion or excessive noise. A value above 100 dB(A) is considered Very good performance for domestic and semi-professional useValues ​​around 85-90 dB(A) are rather "normal" nowadays, sufficient for gaming or multimedia but not so spectacular in numbers.

THD (Total Harmonic Distortion)Expressed as a percentage, this measures the amount of harmonics not present in the original signal that the system introduces. Numbers like 0,001-0,002% are Excellent for a consumer sound cardIf you go up to 0,006%, you're still in a very good range for real-world use, but the system isn't as "clean" in the strictest sense. Even so, these differences are usually difficult to notice in normal listening.

THD + noiseIn dB(A): this combines harmonic distortion and noise into a single parameter. Values ​​around -83 dB(A) fall within the "Good" range, while below -87 dB(A) we usually speak of Excellent results for consumer cardThis is where it becomes very noticeable if the input is poor or if the electrical environment is "dirty".

IMD + noiseAlso expressed as a percentage, this measures intermodulation distortion, that is, the additional components that appear when two tones are played simultaneously. Values ​​of 0,007–0,01% are very good, especially on modern integrated sound cards. Slightly higher values ​​indicate that the system introduces more artifacts when mixing complex signals.

IMD at 10 kHz%: This is a specific case of the high-frequency intermodulation test. It serves to show how the system handles signals in the high part of the spectrum, where some designs show more weaknesses. Reducing this figure means cleaner treble response.

Stereo crosstalk, in dB: indicates how much signal from one channel "leaks" into the other. A crosstalk of -94 dB means that the separation between channels is very high and the The stereo image will be sharp and well-defined.Values ​​of -79 / -84 dB are still good, but there you start to notice a little less isolation between L and R, although in practice it is still more than acceptable.

Real-world example: integrated chip vs dedicated sound card

To see all this in practice, it is very useful to look at a real case where a Integrated Realtek ALC1220 audio chip (marketed by ASUS as SupremeFX S1220 on its ROG motherboards) with a dedicated Creative Sound BlasterX AE-5 sound card. The behavior of both was tested both at idle and with the GPU working at full capacity using a graphics benchmark.

The configuration was fairly typical: a computer with an AMD Ryzen 5 2600X processor, ASUS ROG Strix X470-F Gaming motherboard, AMD Radeon VII graphics card (replacing a previous Vega 56) and a 650W power supply. The operating system was Windows 10 of 64 bits and the integrated card was set to 24 bits and 96 kHz in the two devices tested, which is important because the first battery of tests had been done with a different configuration and the results were not comparable.

The test procedure was as follows: a 1,2m gold-plated auxiliary cable between the speaker output and line input on the motherboard for the Realtek, and between the speaker output and microphone input (configured as line) on the AE-5. This detail is critical: Creative's software allows you to specify whether the rear input works as a microphone or as a line, and if it is not adjusted correctly, the results degrade noticeably.

To force the GPU to work and see how that affected the audio signal, the following was used: Unigine Heaven benchmark in windowed mode while RMAA was running. Two rounds of testing were performed per device: one with the system idle on the desktop and another with the graphics card under heavy load, simulating a game session or demanding 3D work.

The results showed that, at rest, the integrated Realtek ALC1220 audio offered a virtually perfect frequency response (for example, +0.00, -0.07 dB between 40 Hz and 15 kHz) and noise levels and dynamic range in the -100 dB(A) / 101 dB(A) range, rated as “Excellent”. Harmonic distortion was very low (on the order of 0,001%), and intermodulation and crosstalk received ratings from “Very Good” to “Excellent”. In other words, under ideal conditions and with a good board design, the integrated circuit performed very well.

However, when the GPU was put under load, those same parameters suffered: the The noise level rose to around -85 dB(A)The dynamic range consequently dropped to 84-85 dB(A). Harmonic distortion remained excellent, but the THD + noise combination worsened to around -77 dB(A), resulting in a "Normal" rating in that area. Crosstalk and IMD remained in "Very Good" ranges, but the increase in noise was noticeable.

The Creative Sound BlasterX AE-5, for its part, showed very similar figures at rest (flat frequency response, noise below -100 dB(A), slightly superior dynamic range, good crosstalk and distortion around 0,006%). In raw terms, the AE-5 didn't crush the integrated amplifier in every parameter, but it was at least at the same level or slightly above in most.

The big difference appeared when the test was repeated with the GPU under stressWhile the integrated sound card saw a clear degradation in noise and dynamic range, the Sound BlasterX AE-5 virtually mirrored its idle performance: the noise level remained around -100 dB(A), the dynamic range around 100-101 dB(A), and the other parameters barely changed. In practice, this means that the sound you get while gaming or with the GPU under heavy load is almost identical to what you hear on your desktop.

Impact of the GPU and the electrical environment on the measurements

The previous case illustrates something that is often overlooked: the audio quality measured by RMAA depends not only on the sound card itself, but also on how is everything integrated? hardware on your PCComponents such as the GPU can introduce electromagnetic noise that couples to audio lines, especially when these lines pass physically close together or share power supply stages.

In the real-world example, the user discovered that their previous GPU (an AMD RX Vega 56) was generating a significant amount of interference, to the point of severely degrading the integrated audio performance in initial tests. Upon switching to a Radeon VII, which consumes similar power, the performance improved considerably, revealing that the problem wasn't solely with the audio chip, but also with... how the graphics card electrically polluted the system.

This explains why two people with the same motherboard and same audio codec Different results can be obtained: the GPU model, power supply, chassis, internal cabling, power management, etc., can all change. Every PC is unique. Therefore, it's best not to draw universal conclusions from a single measurement, but rather to understand it within the context of your own system.

Another important aspect is the type of input used in the test. On the AE-5, the rear input is a connector that can function as either a microphone or a line input. If it's left in microphone mode when you're actually sending a line-level signal, the circuit operates outside its optimal range and More noise and distortion appearOnly by specifying in the software that the input is used as a line can figures comparable to the theoretical values ​​of the device be obtained.

This sensitivity to the environment is also seen when you connect the output of one card to input of anotherIn an expanded test, the AE-5's pure loopback was compared to itself against the AE-5's output fed into the Realtek ALC1220's line input. The result was very interesting: distortion and IMD were reduced from 0,006% to values ​​in the range of 0,002-0,004%, with a clear improvement in THD+N, while crosstalk changed slightly. This suggests that the AE-5's output is very good, but its rear mic/line input is... the weakest link in the set.

When is it appropriate to use an audio tester?

Beyond laboratory tests with RMAA, ARTA or similar, in everyday life it is very useful to be clear about when it makes sense run an audio check Quickly to avoid unpleasant surprises.

Before a voice recording, it is almost mandatory to do at least one micro testIt costs nothing to record a few seconds, listen, and check for obvious background noise, distortion, saturation, or proximity issues. Detecting these early prevents having to repeat entire takes later because the voiceover is unintelligible or the background noise is unbearable.

During the setup of a live event (concert, talk, streamingetc.) a good practice is to implement a full sound test with all microphones and speakers. There, the level of each channel is checked, feedback is looked for, compressors and equalization are adjusted, and it is verified that everything works together, not just separately.

When editing audio on the computer, a small battery of tests helps you to clearly hear the changes you are applying. detect subtle flaws (clicks, pops, glitches, desynchronization) before finalizing the project. If you're editing video with embedded audio, checking that the voice is properly centered in the mix and that the music doesn't overpower the dialogue is essential for a professional-sounding final result.

It is also highly recommended to do a quick check when you change a device in the chain: new interface USBA new pair of monitors, a mic preamp, etc. Not only do you confirm that everything is connected and routed correctly, but you can also adjust levels to avoid overloading inputs or falling short of the signal. This way, you avoid surprises on the day you really need everything to work perfectly.

Finally, if you ever notice crackling, static, buzzing, or suspiciously low volume, using an audio tester will help you quickly locate the source of the problemSometimes it's a faulty cable, other times a USB port with poor contact, a conflicting driver, or an absurd gain setting. The sooner you identify the problem, the less time you waste struggling with the equipment.