This is a review and detailed measurements of the Quant Asylum QA403 audio analyzer. It was sent to me by the company and costs US $599.
That is one cheerful looking front panel!
It does carry over the odd connectivity of the original QA401 analyzer though. The inputs are inherently balanced but to connect it that way, you need to create a custom cable that goes from 3 pin XLR to dual BNC connections. To use it in unbalanced mode, you can just use one of the BNC but you must, let me repeat, must terminate (short) the other leg/BNC. If you don't noise floor massively increases completely ruining the performance of the analyzer. The manual mentions this but should make a much stronger point about it. An internal relay that would do that shorting for you, or inclusion of the terminator would be helpful in this regard.
Backside shows that the analyzer surprisingly is just powered via USB, making it highly portable:
The user interface actually reports the current usage which is a nice addition.
The software is shockingly efficient. I think it installed in a second or at most two! No tedious setup script, no reboot, no nothing. Click on the download and you almost immediately are ready to use the unit.
What is nice is that there is a REST API and server in the app so you can write your own custom application for it if you don't like its own software. Speaking of that, the interface in the original analyzer was pretty, emulating a hardware scope and such. Alas, it was hard to use because of that. The new version dispenses with all that with a very utilitarian looking UI that is very simple to use for simple functions. Anything else is not easily "discoverable" such as knowing where and when to right click on a button to bring up its settings. As a tool you use all the time, you will learn it quickly but does take way from the initial experience.
There are hardly any products in this space other than the monster Audio Precision APx555 that I use. I asked ChatGPT Pro to create a comparison and it produced this:
Note: THD+N figures use different test conditions and should not be interpreted as a direct apples-to-apples comparison.
You can see the stunning difference in price and that if you never want a software update for AP. If you do, the cost is $700/year with a 3-year commitment! For that money, you do get fair bit more starting with a world-class analog front-end that autoscales the input all the way up to whopping 160 Volts RMS compared with manual scalar in QA403 which limits at 40. This sets a hard barrier for measuring amplifier power (V^2/R).
Most of the time we don't live in that domain though so question is how usable is the QA403 is for assessing engineering excellence and finding design issues. The original did not get there with a SINAD of 99 dB. Let's measure the QA403 and see how it does.
Quant Asylum QA403 Analyzer Measurements
For this testing, I used unbalanced (terminated) connection with 2 volts output from my APX555 as that is the most common scenario we have in testing DACs. First, let's see how APX555 performs feeding itself using its High Performance signal generator and analyzer:
We see that it beats is spec by 1.6 dB, with distortion that is essentially unmeasurable at -155 dB! So noise dominates and that sets SINAD at 121.6 dB. Feeding the same high performance sine generator to QA403 produces quite respectable results:
We are seeing SINAD that meets spec and clocks at nearly 116 dB, again, noise limited as its distortion products are at -125 dB as indicated by pure THD meter (as opposed to THD+N which is equiv. to SINAD). This is a huge step above QA401 and easily makes it possible to use QA403 to find the sources and noise in audio sources in all but state of the art implementations. As a refresher, the blue bucket of SINAD ranking for DACs starts at 110 dB.
Note how the reported level of 1.914 Volt RMS is wrong. The fed input was 2 volts, not 1.91. The error (9.6%) is introduced due to the Hann window function which impacts the accuracy of spikes in FFT where the number is computed. You can select flat top to solve that but then get massive spectrum leakage and elevated noise floor in FFT. You can then counter that with very large FFT and averaging. Once there, you have better accuracy but very slow response. Audio Precision analyzer has a dedicated volt meter that is implemented in time domain so has none of these issues although if you use its FFT, you will face the same issue.
The generator in the QA403 is DAC based and predictably falls short of what AP can generate:
We see a lot of distortion and noise which limit it to SINAD of 109 dB. So good enough to measure vast majority of amplifiers out there with our excellent blue category starting at SINAD of 95. There are high performance sine generators out there that you could use instead but they are usually limited in frequency and at any rate, you would lose the automation in the analyzer.
Looping back the generator into the QA403, we get its end to end performance:
This is a more practical limit of what you can get in testing a purely analog audio device (pre/power amp). So again, more than good enough to determine if a device has serious performance issues.
You could use a Comos ADC to get better results but it comes with a heap of usability issues and measurement constraints:
Again, using the loop back we see a nice and flat frequency response:
Note that the above frequency response is generated using a continuous "CHIRP" signal. This gives you very fast response but again, is subject to FFT length issues -- something that the app warns you about (if FFT is too small).
Net, net, you need to know what you are doing as with any other analyzer to understand exactly what you are seeing. This is much better than Cosmos but novice players should take caution and ask questions before going public with measrements.
Conclusions
ASR has resurrected the importance of measurements after years of neglect by the industry and publications that are supposed to watch over them. It is easy now to blame a poorly measuring audio device while not appreciating how expensive it is to buy an audio analyzer. My APX555 was expensive when I bought it but now it has gone up another $15,000! And has pretty hefty yearly maintenance fee to boot. It is easier than QA403 and is much more capable but the cost barrier is formidable. It is such a wonderful relief to see the performance of QA403 at such a ridiculously low price. If you are designing hardware, just a prototype would cost you more than this tool! Every designer should have one with no excuse to do without.
There is also large demand for these low cost analyzers for factory floor QA/production testing. Companies buy way out of production AP hardware for that but they lack software support and won't run on modern PCs. QA403 nicely solves that problem especially given its ease of programming for custom testing.
For a hobbyist, the QA403 is a wonderful option that enables you to easily separate bad audio products from good and even excellent. And at a cost that is more than reasonable.
It is my pleasure to recommend the QA403 audio analyzer.
-----------
As always, questions, comments, recommendations, etc. are welcome.
Any donations are much appreciated using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
That is one cheerful looking front panel!
Backside shows that the analyzer surprisingly is just powered via USB, making it highly portable:
The user interface actually reports the current usage which is a nice addition.
The software is shockingly efficient. I think it installed in a second or at most two! No tedious setup script, no reboot, no nothing. Click on the download and you almost immediately are ready to use the unit.
What is nice is that there is a REST API and server in the app so you can write your own custom application for it if you don't like its own software. Speaking of that, the interface in the original analyzer was pretty, emulating a hardware scope and such. Alas, it was hard to use because of that. The new version dispenses with all that with a very utilitarian looking UI that is very simple to use for simple functions. Anything else is not easily "discoverable" such as knowing where and when to right click on a button to bring up its settings. As a tool you use all the time, you will learn it quickly but does take way from the initial experience.
There are hardly any products in this space other than the monster Audio Precision APx555 that I use. I asked ChatGPT Pro to create a comparison and it produced this:
| Specification | QA403 | APx555 B Series |
|---|---|---|
| Analog channels | 2 in / 2 out | 2 in / 2 out |
| Architecture | 32-bit ADC/DAC, FFT-based | Precision analog + digital |
| Sampling rate | 48 / 96 / 192 kHz | Up to 200 kHz digital |
| Measurement bandwidth | ~80 kHz usable | >1 MHz |
| Residual THD+N | -105 dB (loopback) | Typically < -120 dB |
| Residual THD | -115 dB (loopback) | Typically < -130 dB, midband |
| Input noise | -115 dBV (specified configuration) | ≤1 µVrms, 22 kHz BW |
| Maximum input voltage | 40 Vrms | 160 Vrms unbalanced / 300 Vrms balanced |
| Maximum output voltage | 8 Vrms SE / 16 Vrms balanced | 13.33 Vrms SE / 26.66 Vrms balanced |
| Amplitude accuracy | ±2% absolute | ±0.03 dB (~0.35%) |
| Analog input coupling | AC only | AC / DC |
| Digital audio interfaces | Limited I²S expansion | AES, S/PDIF, optical; optional interfaces |
| Calibration | Factory calibrated | Traceable calibration available |
| Price | $599 | $45,325 with no software update |
Note: THD+N figures use different test conditions and should not be interpreted as a direct apples-to-apples comparison.
You can see the stunning difference in price and that if you never want a software update for AP. If you do, the cost is $700/year with a 3-year commitment! For that money, you do get fair bit more starting with a world-class analog front-end that autoscales the input all the way up to whopping 160 Volts RMS compared with manual scalar in QA403 which limits at 40. This sets a hard barrier for measuring amplifier power (V^2/R).
Most of the time we don't live in that domain though so question is how usable is the QA403 is for assessing engineering excellence and finding design issues. The original did not get there with a SINAD of 99 dB. Let's measure the QA403 and see how it does.
Quant Asylum QA403 Analyzer Measurements
For this testing, I used unbalanced (terminated) connection with 2 volts output from my APX555 as that is the most common scenario we have in testing DACs. First, let's see how APX555 performs feeding itself using its High Performance signal generator and analyzer:
We see that it beats is spec by 1.6 dB, with distortion that is essentially unmeasurable at -155 dB! So noise dominates and that sets SINAD at 121.6 dB. Feeding the same high performance sine generator to QA403 produces quite respectable results:
We are seeing SINAD that meets spec and clocks at nearly 116 dB, again, noise limited as its distortion products are at -125 dB as indicated by pure THD meter (as opposed to THD+N which is equiv. to SINAD). This is a huge step above QA401 and easily makes it possible to use QA403 to find the sources and noise in audio sources in all but state of the art implementations. As a refresher, the blue bucket of SINAD ranking for DACs starts at 110 dB.
Note how the reported level of 1.914 Volt RMS is wrong. The fed input was 2 volts, not 1.91. The error (9.6%) is introduced due to the Hann window function which impacts the accuracy of spikes in FFT where the number is computed. You can select flat top to solve that but then get massive spectrum leakage and elevated noise floor in FFT. You can then counter that with very large FFT and averaging. Once there, you have better accuracy but very slow response. Audio Precision analyzer has a dedicated volt meter that is implemented in time domain so has none of these issues although if you use its FFT, you will face the same issue.
The generator in the QA403 is DAC based and predictably falls short of what AP can generate:
We see a lot of distortion and noise which limit it to SINAD of 109 dB. So good enough to measure vast majority of amplifiers out there with our excellent blue category starting at SINAD of 95. There are high performance sine generators out there that you could use instead but they are usually limited in frequency and at any rate, you would lose the automation in the analyzer.
Looping back the generator into the QA403, we get its end to end performance:
This is a more practical limit of what you can get in testing a purely analog audio device (pre/power amp). So again, more than good enough to determine if a device has serious performance issues.
You could use a Comos ADC to get better results but it comes with a heap of usability issues and measurement constraints:
Again, using the loop back we see a nice and flat frequency response:
Note that the above frequency response is generated using a continuous "CHIRP" signal. This gives you very fast response but again, is subject to FFT length issues -- something that the app warns you about (if FFT is too small).
Net, net, you need to know what you are doing as with any other analyzer to understand exactly what you are seeing. This is much better than Cosmos but novice players should take caution and ask questions before going public with measrements.
Conclusions
ASR has resurrected the importance of measurements after years of neglect by the industry and publications that are supposed to watch over them. It is easy now to blame a poorly measuring audio device while not appreciating how expensive it is to buy an audio analyzer. My APX555 was expensive when I bought it but now it has gone up another $15,000! And has pretty hefty yearly maintenance fee to boot. It is easier than QA403 and is much more capable but the cost barrier is formidable. It is such a wonderful relief to see the performance of QA403 at such a ridiculously low price. If you are designing hardware, just a prototype would cost you more than this tool! Every designer should have one with no excuse to do without.
There is also large demand for these low cost analyzers for factory floor QA/production testing. Companies buy way out of production AP hardware for that but they lack software support and won't run on modern PCs. QA403 nicely solves that problem especially given its ease of programming for custom testing.
For a hobbyist, the QA403 is a wonderful option that enables you to easily separate bad audio products from good and even excellent. And at a cost that is more than reasonable.
It is my pleasure to recommend the QA403 audio analyzer.
-----------
As always, questions, comments, recommendations, etc. are welcome.
Any donations are much appreciated using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/