How to measure audio performance

Equipment and methods

Abstract

According to my observation, in the audio scene there are two fractions: Subjectivists and objectivists. The Subjectivists claim that the ear can best judge audio performance and the objectivists claim that metrics matter most. I'm somewhere in the middle between both positions, but closer to the objectivists.

The objectivist part of me is highly interested in how well the equipment that I build performs in terms of usual audio metrics. In this article I will explore options and methods to measure audio equipment performance.

Professional measurement equipment

Professional measurement equipment offers stellar performance, but is totally unobtainable for DIY audio enthusiasts. The Audio Precision APx555 costs around 28000 USD and this is well above the budget for any DIY audio hobbyist.

Semi-Professional measurement equipment

I came across the Quant Asylum QA402 audio analyzer, which sells for only 500 USD.

Here is the predecessor product QA401 review on Audiosciencereview.

There is also a discussion about the QA401 on DIYaudio.

Quant Asylum also provides dedicated software along with their products.

While audio performance of the QA401 does not seem to be that exciting, it has some very compelling features specific for audio measurements like wide input voltage range and differential inputs and outputs.

Chis Russel from Hifisonix wrote a review and application of the Quantasylum QA-402. A good part of this applies to audio measurement setups in general and I highly recommend reading his advice.

Meanwhile, the QA402 is history again since it relied on AKM chips, that became unavailable after the AKM fab was destroyed. Since February 2022, there is the predecessor QA403 available, which uses ESS chips like any other high end equipment nowadays. Inside the QA403, there is a collection of the finest chips one could use to build audio equipment. Given they are implemented well, this could be a stellar piece of equipment for only 549 USD. By all means, the QA403 performs much better than its predecessor, the QA402. The QA403 looks pretty good from afar.

In case I had not recently spent roughly 300€ on separate DAC and ADC already, I would get the QA403 immediately. On the other hand side, the equipment I have looks more nicely in the living room as dual use for listening music. The QA403 does not look good in the living room from an aesthetic point of view.

There is a discussion about the QA403 on the "DIY Audio" forum.

Sound cards as measurement equipment

Using a sound card for doing measurements seems tempting because a good sound card may already be available and this would double as measurement equipment free of charge. A sound card is designed for low voltage level input and output and this restricts the measurement range to line level signals. For testing power amplifiers, an attenuator network would be required at the sound card input.

I investigated some sound cards for building my own budget measurement setup and found the audio performance to be totally unsuitable. I had the ESI U24 XL at home, which fails to meet even the most basic audio quality standards. I bought the Focusrite Scarlet 2i2 because I heard positive reviews from DIYers who used this sound card for doing measurements. I found the Focusrite sound card to be only marginally better than the ESI and returned it.

Sound cards are simply not designed to deliver performance high enough to be used as measurement equipment. In case the device under test has significantly more distortion than the sound card, sound cards may still be an option. For example for measuring loudspeakers.

I gave up on the idea using a sound card as part of my lab equipment until I bought the Topping D10 balanced. This DAC delivers pretty good performance for a reasonable price. Distortion is low enough to not swamp the distortion of the DUT.

And I came across a very promising ADC: The E1DA Cosmos ADC. As the name implies, this is an ADC only, but designed to be suitable for audio measurements. Here is a review of the E1DA Cosmos ADC. And there is a discussion about the E1DA Cosmos ADC on Audiosciencereview. According to the reviews, performance seems very good. The only minor drawback seems the low input impedance. I bought an E1DA ADC and can confirm it is very good. Here is my review of the E1DA Cosmos ADC.

Topping D10 balanced DAC and E1DA Cosmos ADC

I bought the Topping D10 balanced DAC and the E1DA Cosmos ADC both for listening or recording music and also for dual use as measurement equipment in my audio electronics lab.

Below are some plots recorded with REW. Both devices were run with 192kHz sample rate. The E1DA Cosmos ADC was set to mono mode and the XLR input was set to allow 10VRMS, which loads the DAC with 3.48kΩ impedance.

The plots speak for themselves: Low noise floor and distortion make this setup useful for testing power amplifiers. The E1DA ADC has ten times less distortion than the Topping DAC. I don't like the multitude of spikes in the higher frequency range. There would be DACs with better performance, but not in this price class.

Signal generators

Samuel Groner has made a helpful comparison of low distortion oscillators.

Software

Audio Precision supports sound cards as measurement equipment and recommends this only for applications where the device under test has significant more distortion than the sound card, like loudspeakers.

I find REW easy to set up and use. While REW is optimized for room acoustics, it can also be used for some general audio measurement setups.

ARTA (Audio Real Time Analysis) is also a very popular software for doing audio measurements. The software is commercial, but allows a free demo mode. I have not used ARTA yet since I'm happy with REW. With a personal license available for 79€ this would be an affordable option though.

AudioTester is another option I haven't had a closer look at yet.

Conclusion

It is no surprise that good measurement equipment costs accordingly. However, there is good equipment available for a rather low budget.