Could someone explain the RMAA test?

[helloworld]

There's a plenty of RMAA test results in those tech reviews website. Could someone explain to me the difference between RMAA and APX analyzer? What's the limitation of the RMAA test? How accurate is it?

 

[SIY]

RMAA isn't a test, it's a piece of free software. With RMAA and a soundcard, you can run some very basic tests (frequency response, crosstalk, distortion), but it's not as versatile or reliable as good software like ARTA or Virtins MI. And you're limited by the performance of the soundcard and will also need to add some signal conditioning. I wrote a series of articles on using soundcards and basic test software which can be found at the AudioXpress website and will give you an idea of what's involved in setting up and properly using these sorts of measurement tools. A smart user can do excellent measurements with the right setup, but it's easy to generate lots of garbage data.

An Audio Precision (AP) analyzer is an expensive (think the price of a good new car) piece of professional gear. Better resolution, versatility, and reliability, as you'd expect for pro equipment.

 

[sergeauckland]

I've found RMAA to be very flaky under Windows 7 and 10 on my Lexicon USB audio interface, whereas it works OK on Windows XP using a Digigram PCMCIA audio interface. Even then, not as useful as ARTA, I haven't tried Virtins.

Probably what I use most is Soundcard Scope. It's limited to 44.1kHz and 16 bit, but the way I see it, if it's too low to measure, I don't need to worry about it. Very useful for quick rough and ready pass/fail type measurements, and I only need to go further into it if it fails.
S.

 

[bennetng]

There's a plenty of RMAA test results in those tech reviews website. Could someone explain to me the difference between RMAA and APX analyzer? What's the limitation of the RMAA test? How accurate is it?

An important reason that made RMAA results unreliable is a lot of users are unable to perform correct measurement procedures, for example the false frequency response mentioned in the article below:
http://audio.rightmark.org/downloads/audigy-setup.zip

NwAvGuy's opinion about RMAA:
http://nwavguy.blogspot.com/2011/02/rightmark-audio-analyzer-rmaa.html
Read the comment section as well since Maxim Liadov, the chief RMAA developer also commented in the blog.

The most reliable RMAA-based measurement website is probably iXBT, since they are the maker of RMAA.
http://prosound.ixbt.com/interfaces.html

English version of iXBT, but old products only.
http://ixbtlabs.com/sound.html

 

[amirm]

One of the key things in audio measurements is knowing what is correct and what is not. Simple configuration issues can lead to incorrect results. The problem with RMAA in that regard is that it is automated and generates a bunch of data in the hands of people who don't know what is or is not right. And so creates a lot of false data.

I have only tried it in limited tests but gave up at the time because I could not generate high-confidence results. I do mean to circle back to it at some point as it does have some good functionality.

As far as the larger question, it has been answered already. Using any sound card without knowing its performance limits is a recipe for problems. It can for example be over driven. And certainly can't be used to test anything other than DACs and pre-amps as power products (headphone or speaker) produce such high voltages that it easily blows up the sound card. Audio Precision has automatic input scalars that allow it to accept up to a whopping 150 volt input.

 

[sergeauckland]

Just to add to Amir's post, a sound card can be used for measuring high voltages, provide you make yourself a suitable attenuator which ensures that the audio card never clips. An external USB audio card with input level metering is useful, but at very least a 'clip' light is essential. In my own case, I've built a set of dummy loads with variable attenuators so I can measure the output of power amps up to the limit of the dummy load power handling, and somewhat more if I keep measurement time short so the loads don't overheat

One can of course calibrate the input of the sound card and use a calibrated attenuator to make it suitable for measuring voltages, but I find it a lot easier to use an external audio millivoltmeter for that function, and use the sound card for relative measurements, like Distortion, frequency response and S/N, none of which require an absolute voltage measurement. Monitoring the waveform on an oscilloscope I think is also essential with any measurement, whether with a sound card or a dedicated meter, as it's too easy to be fooled by a clipped signal.

S.

 

[SIY]

Just to add to Amir's post, a sound card can be used for measuring high voltages, provide you make yourself a suitable attenuator which ensures that the audio card never clips. An external USB audio card with input level metering is useful, but at very least a 'clip' light is essential. In my own case, I've built a set of dummy loads with variable attenuators so I can measure the output of power amps up to the limit of the dummy load power handling, and somewhat more if I keep measurement time short so the loads don't overheat

One can of course calibrate the input of the sound card and use a calibrated attenuator to make it suitable for measuring voltages, but I find it a lot easier to use an external audio millivoltmeter for that function, and use the sound card for relative measurements, like Distortion, frequency response and S/N, none of which require an absolute voltage measurement. Monitoring the waveform on an oscilloscope I think is also essential with any measurement, whether with a sound card or a dedicated meter, as it's too easy to be fooled by a clipped signal.

Alternatively, you can use one of these. I have one in my lab and it works flawlessly, giving accurate voltage readings and making sure the measured signal is in range for the soundcard.

 

[danadam]

I wrote a series of articles on using soundcards and basic test software which can be found at the AudioXpress website and will give you an idea of what's involved in setting up and properly using these sorts of measurement tools.

Practical Test & Measurement I assume?

 

[SIY]

Yes, that was part 1. I think it ran 7 or 8 parts.

 

[mansr]

I have only tried it in limited tests but gave up at the time because I could not generate high-confidence results. I do mean to circle back to it at some point as it does have some good functionality.

I don't like RMAA. The metrics it produces are useful as such, but without knowing exactly how it operates, interpreting them can be perilous, and it has a tendency to blow up for unclear reasons. To top it off, the user interface is terrible.

 

[Blumlein 88]

I haven't had trouble with RMAA blowing up. It even runs in Linux via Wine fine. Now I've found it to work better when you generate its test file, play and record that file and let RMAA analyze that file for results. I don't use it in a live loopback mode.

You can figure out how it is doing most measurements looking at the test file it uses. One of the problems is some of the distortion measures are using a signal 6 to 9 db below max. So if you've set it up to have max level properly for max output of the DUT some of these distortion numbers are going to look better because they didn't use max levels.

Obviously the quality of the results is dependent upon the quality of the ADC, and the understanding of the person who sets things up. What RMAA tries to do is be automatic and useful for someone who doesn't know all that much about doing testing themselves. As usual that results in compromises. The way it runs a signal and does all the analyzing for you could be handy. I wish there were a pro or plus version that let the user select and set some parameters which would make it much more useful by a large number of people who would like solid simple results when it doesn't justify owning an AP unit or Prism D scope or similar.

I suppose if I were a software writer that wouldn't be hard to do. I've made files to do a thorough job of testing DACs and line level gear. Interpreting that is hands on. In my mind it wouldn't be hard to automate it, but then everything is simple if someone else is doing it.

 

[restorer-john]

...but then everything is simple if someone else is doing it...

Yes, that's so true. I tried to de-compile and 'fix' the software for my ancient AudioLab PC driven hardware. I figured, it can't be that hard, I wrote tons of simple software 'back in the day'. If the designer could do it, why couldn't I etc.

Total fiasco. Lucky I kept original backups or I would have bricked it completely. Lesson learned.

 

[mitchco]

There is a RMAA Pro version and does allow more changing of parameters: http://audio.rightmark.org/index_new.shtml

The RMAA 6.4.5 PRO is also available with a number of features not present in the freeware version. Updates from any older 6 PRO version are free of charge. Use the same link when you bought PRO version. If you have no PRO version yet, may be it is a good time to upgrade your RMAA.

RMAA PRO features:

• 4 times more detailed spectrums (FFT Size option);
• changable test parameters in test options;
• GUI skin support (15 skins bundled).

Archimago uses it with good results, including a number of other tests. http://archimago.blogspot.com/ With the addition of his RME ADI -2 Pro FS ADC, it will be interesting to see how his measures compare with traditional analysers...

 

[Blumlein 88]

There is a RMAA Pro version and does allow more changing of parameters: http://audio.rightmark.org/index_new.shtml

The RMAA 6.4.5 PRO is also available with a number of features not present in the freeware version. Updates from any older 6 PRO version are free of charge. Use the same link when you bought PRO version. If you have no PRO version yet, may be it is a good time to upgrade your RMAA.

RMAA PRO features:

• 4 times more detailed spectrums (FFT Size option);
• changable test parameters in test options;
• GUI skin support (15 skins bundled).

Archimago uses it with good results, including a number of other tests. http://archimago.blogspot.com/ With the addition of his RME ADI -2 Pro FS ADC, it will be interesting to see how his measures compare with traditional analysers...

I vaguely remembered a pay version, but not what paying added to it. Thanks for pointing this out.

 

[mansr]

I haven't had trouble with RMAA blowing up. It even runs in Linux via Wine fine.

By blowing up, I meant producing obviously bogus results.

You can figure out how it is doing most measurements looking at the test file it uses.

The test signal is obviously easy to inspect. I was referring to the analysis, things like FFT size and window function.