• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Review and Measurements of QuantAsylum QA401 Audio Analyzer

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
22,563
Likes
34,640
Location
Seattle Area
#1
This is a review and detailed measurements of the QuantAsylum QA401 audio analyzer. It was kindly sent by company principle and member, @QAMatt. Despite being a total hardware and software solution, the QA401 is ridiculously low priced at just US $449. To put things in perspective, my Audio Precision APx555 retails for US $28,000. Even cheaper models in their range and from Prism Sound clock well over US $5,000.

The QA401 comes in a surprisingly small package:

QuantAsylum Analog Input Audio Analyzer Review.png


It is no bigger than many desktop DACs and headphone amps. Better yet, the unit is entirely USB powered although it apparently pushes the limits of USB. I had no issue driving it with the USB 3 hub built into my monitor that I use for testing USB products.

The outputs and inputs are differential but in the form of dual BNC rather than XLR. If you use just one leg, then you need to put termination for the other which is what you see in the above setup.

Of note, there are only analog inputs and outputs. You cannot test digital interfaces such as S/PDIF, Toslink or AES/EBU.

The main competition to the QA401 is using an ordinary sound interface/ADC and here, the QA401 pulls way ahead with ability to tolerate far higher input voltages to the tune of 56 volt peak to peak. This should allow you to test low to medium power amplifiers without issue. Such a test would destroy the input stage and possibly the rest of any standard sound interface without some sort of external attenuation.

The software interface requires some getting used to. I will talk more about this measurement later. For now, let's focus on functionality:

QuantAsylum Analog Input Audio Measurements.png


As you see, the software does NOT use standard windows controls. Instead, a custom set has been created on the right to mimic the look of controls on a virtual hardware interface. I know the appeal of this to hardware engineers and test technicians but I am not a fan. The rotary controls for example show no values. You have to look to the left top text row to see what you have set them too. Rotating them is also hard unless you have a mouse wheel (so a pain on touchpad on a laptop).

The toughest thing for me to figure out was how to bring up the "context menu" to change settings such as dB Relative. Clocking either the left or right mouse buttons on the green controls did nothing. I had read the manual carefully to realize you have to hold the control key down on the keyboard and then click on the green button in the UI. Why? Why isn't this just a right mouse button?

The right pane also doesn't change size which is an issue I also have with the panels on Audio Precision software. It is difficult to make custom controls like this scale in size.

Turning our attention to the left pane, the single-dot aspect of the graph makes it hard to read. On AP you can select the thickness of the line which helps a lot with visibility. There is no way to freeze one graph and add more to it as I commonly show in my measurements. Likewise colors are fixed as yellow and red for the two channels.

On the plus side, there is ability to have the software detect peaks and label/track them. It was a bit tricky as you have to click on the exact pixel of the spurious tone you are interested in. But once there, it nicely puts a tag on it (M0, M1, etc.) and shows their frequency and amplitude. Audio Precision has no such feature and is a glaring hole. Prism Sound dSound III analyzer from what I recall goes further than QA401 by automatically labeling all the spurious tones at once.

Nice to see a button to start/stop the acquisition. On AP, it is always running and sucks a single CPU core dry. Not an issue on desktop but on my laptop, it can start to cook the CPU. So the Q401 approach is much better.

I like that at a glance I can see both THD+N and THD alone (without noise contributions).

The hardware has dual sample rates: 48 and 192 kHz. I performed all of my testing with the former. The latter is very useful for measuring such things as amplifier bandwidth. The sampling rate on AP goes up to 1 Megahertz which allows me to do such things as show you the spectrum of switching frequencies in an amplifier. You can't do that with Q401 or just about any sound interface for that matter.

There is an attenuator that reduces input levels by 20 dB. You have to manually activate it though. On AP, there are multiple attenuators that work automatically to optimize the input level for the internal ADC.

Overall, the QA401 hardware and software is pretty easy to use once you get past the few quirks above. The fact that the hardware is simple means there is less settings and less chances of measuring the wrong thing.

One of the great benefits of QA401 software is its extensibility. You can write a DLL (shared library) that extends its functionality. If you have used a digital audio workstation with effects plug-in, you already know how this work. This is far superior to the extensibility of the Audio Precision which forces you to write a host application to control the AP software as an object. This translates into having to do your own user interface as you may have seen from the crude graphs I post for such things as impedance measurements. With QA401, the app stays in control and can perform such functions.

Q401 Audio Measurements
I started testing by evaluating the performance of the ADC in the QA401. I programmed my Audio Precision analyzer to produce signal using its analog generator which has ultra low distortion. Explaining further, most audio analyzers use a DAC and digital bitstream to produce their analog output. This provides tons of flexibility as far as signals are concerned and is cheap to implement. The downside is that DACs have distortions and those get blended into whatever you are testing. The AP has both choices by the way: DAC and analog. The former is used for more complex signals (e.g. multitone).

Anyway, I turned off the attenuator on the QA401 and adjusted the levels to keep it from clipping (it has a nice red indicator that comes on the UI) while trying to get the best performance:
QuantAsylum Analog Input Audio Measurements.png


This was about 0.7 volt. I think Matt told me that the ideal value is around 0.64 volts or so. At this input level, the SINAD (THD+N) is 99 dB. Telling the AP to analyze itself resulted in 116 dB. So we have quite a shortfall there compared to much more expensive AP.

Assuming that you want your analyzer to be 10 dB better than what you are testing, you would be limited to measuring audio products that have a SINAD below 90 dB or so. This would only let you measure DACs in forth (worst) quadrant of units I have tested. So you are out of commission testing mid to high performance DACs.

On the other hand, many amplifiers struggle to reach such performance levels so you can test them fine with QA401.

Next, I switched chairs, making the QA401 the source generator and the APx555 the analyzer. This is what I get:


QuantAsylum Analog Input Analog Generator Performance Audio Measurements.png


I could not average the FFT to get rid of the noise. Unlike QA401, the AP has poor buffering and falls behind, corrupting its received data with QA401 running concurrently. That doesn't impact the distortion measurement though showing a second harmonic of -100 dB. This should be fine again for measuring amplifiers and such.

Looping back the QA401 (with a cable) generating its own signal and measuring it gives us this:
QuantAsylum Analog Input From Own Generator Audio Measurements.png


The total THD+N/SINAD of 101.5 dB shows that there is no need for a cleaner generator as we are limited by what the ADC can do. The number also perfectly matches the specifications provided by QuantAsylum.

Conclusions
The QuantAsylum QA401 is crazy good value in audio analyzer hardware and software. While you may be tempted to use a sound card/interface and independent software to analyze the data, let me warn you that they are universally hard to use compared to QA401. You will be stuck in what I call "open-loop" analysis in that you play one thing, and then measure it. You cannot perform synchronous analysis, e.g. sweeps, such as you can with a dedicated analyzer.

If you have any aspiration of duplicating what I do here, you have no path with QA401. The hardware performance is just not there even though the functionality is mostly there. It would have been wonderful if the performance was that good for 50 times less money. :) Our playback equipment has gotten so good that it requires very high-end measurement gear to handle it.

If you are manufacturing audio hardware, the QA401 would be a wonderful QC system. If something is not manufactured right, it will break good and not just degrade a bit in performance so the QA401 hardware precision is more than good enough. QuantAsylum has a batch system called "tractor" which lets you automate your suite of tests.

Overall, I am happy to recommend the QuantAsylum for its incredible value, functionality, size and overall ease of use.

------------
As always, questions, comments, recommendations, etc. are welcome.

Looking ahead to winter, the panthers want to learn to ski so they have something to do beside just posing with audio hardware. I need to buy them the gear and practice lessons so please donate generously using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
 
Last edited:

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
22,563
Likes
34,640
Location
Seattle Area
#2
Oh, I forgot to post one more measurement performed. There was some talk about noise modulation with Q401. That is, if the source signal goes up in level, so does the level in measurements. So I tested the Q401 using a much lower amplitude signal at just 50 millivolts:

QuantAsylum Analog Input 50 mv Audio Measurements.png


Noise floor does go down with signal to the tune of 10 dB. I tested the same on AP and saw maybe half as much. So some is there but I don't think it is excessive.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
22,563
Likes
34,640
Location
Seattle Area
#3
And if you want to know how the APx555 analyzer does with the original signal at 0.7 volt, here it is:

Audio Precision APx555 Loopback Audio Measurements.png


Only distortion is second harmonic and it stays almost buried in the noise floor at around -146 dB. :) So with respect to seeing distortion in FFT, the APx555 is 40 dB better.
 

Hugo9000

Active Member
Joined
Jul 21, 2018
Messages
296
Likes
722
#4
Thanks, @amirm , interesting device! I've seen a few mentions of it here and a couple of other places, so it's good to see your tests and review, and discussion of the proper applications for an analyzer like this. It's definitely well worth the price simply for verifying quality control of a lot of devices, an important function that you mentioned. That's one thing that has always struck me about the anti-measurement people, they either don't know or forget/ignore that measurements are an essential tool in verifying proper operation as part of standard quality control/quality assurance in good manufacturing, no matter the field. Key of course is using the right tool for the job, and someone who knows how to correctly interpret the results. Even those who think "tuning by ear" of audio equipment is the best thing ever should realize how useful/necessary measurements are for a multitude of reasons.

Thanks again, and thanks also to @QAMatt for sending the unit in for testing!
 

JohnYang1997

Major Contributor
Joined
Dec 28, 2018
Messages
1,737
Likes
1,078
Location
Sydney - Australia
#9
Would like to see thd(no n) vs amplitude for input and output individually. But that may be very time consuming to test. I see this qa401 to be disappointing.
 

Rja4000

Addicted to Fun and Learning
Joined
May 31, 2019
Messages
523
Likes
354
Location
Liège, Belgium
#13
Top Bottom