RMAA (Rightmark) users--a couple questions!

[bongomania]

With noise ratings, they are given in negative numbers. Is a greater absolute magnitude lower noise or higher?

With THD, all my tests of audio cables show a value of about 0.019%. This makes me think either my testing is flawed, or all of these cables have identically incredible THD. Which do you think, and should I be trying a different approach?

And the "impulse response" test doesn't seem to show any results, and there's not even a window for a result to show in. Can you explain RMAA's impulse test, and whether it is useful for cable testing, and where the results should be found?

Thanks guys!

[fdeck]

Personally, I don't like RMAA, because you don't know the absolute signal levels unless you have a true-RMS voltmeter. Instead, I prefer to use Virtual Analyzer with pure sine waves, and measure the signal levels with a voltmeter. VA will read out THD in a waveform (assuming a sinusoidal input). You can also see the levels of the harmonics in the analyzer, as a way to check that the measurement is meaningful.

I appreciate what you're doing, but...

You know going into it that the distortion and noise levels introduced by a cable will be unmeasurable with mainstream audio gear.

I've generated test signals that are similar to those used by RMAA for frequency response, and they really improve the accuracy of measurements made with VA. Give me a week (and then start bugging me), and I'll upload the audio files to my web page. It turns out that a MP3 player is a really handy signal generator. Whatever MP3 does to the audio, doesn't seem to affect routine measurements like frequency response.

[bongomania]

Originally Posted by fdeck

You know going into it that the distortion and noise levels introduced by a cable will be unmeasurable with mainstream audio gear.

It was wishful thinking I guess.

I do have a waveform generator and a voltmeter. Also an old analog scope, and I have a new Rigol digital scope coming in soon. I was already planning on using those for testing the claims of "better transient response in the low frequencies" of some of the cables... It hadn't occurred to me to use them for THD and noise. More suggestions/tips please?

[Passinwind]

Originally Posted by bongomania

More suggestions/tips please?

Did you ever fix the horrendous loopback signal/noise figures you posted in the first thread?

With my soundcard, THD reads at just under .001% in loopback tests. So to me, those figures you quoted also look pretty bad.

Re: fdeck's comments, at least one true RMS meter has to be a given in any real testing rig. I double check levels with my o'scope from time to time as well.

I'll revisit my RMAA test protocols as soon as I get my old computer back in operation. I'm posting on a Win7 machine that is useless for any of my audio analysis/production stuff. I'm about to build a new test jig to interface with my soundcard in the old 'puter, hopefully in the next few days.

Francis: thanks for the heads up on Virtual Analyzer. I'll give it a try as soon as I get my test rig up and going again. I've always regarded RMAA as useful for relative measurements more than for absolute ones.

[fdeck]

It's impossible to read technical meaning into pseudo-technical marketing verbiage. "Transient response in the low frequencies" is meaningless.

I personally think that the emphasis on transient response is just a red herring. Cable makers were confronted by the fact that all cables have the same frequency response when driven by low impedance sources, so they had to invent some other issue in order to maintain the aura of mystery and confusion. Voila, transient response. Yet it's extremely unlikely that two passive circuits, both with flat frequency response, will have measurably different transient response. Effectively (and yes there are exceptions) the frequency response curves are telling you that there is nothing interesting going on in the time domain. If you have the complex response curve, then you can compute the transient response function.

I've played around with a program called GNU Octave, which one might be tempted to call a MatLab clone. It is capable of reading and writing numerical arrays in the form of .WAV files, thus allowing for the computation of arbitrary waveforms, and their analysis.

[12bass]

If attempting to test cables using analog loopback, the results shown are probably mostly just the sound card's D/A-A/D figures. In comparison to the converters, I'd guess that the cable has a negligible measurable impact.

[fdeck]

Thanks for the PM, Charlie. The program I was referring to is called "Visual Analyzer."

http://www.sillanumsoft.org/prod01.htm

For a time, I had my own RTA program that I had written, but keeping up with successive versions of MS Windows using my archaic copy of Visual Basic got to be too much of a headache.

[bongomania]

Originally Posted by fdeck

It's impossible to read technical meaning into pseudo-technical marketing verbiage. "Transient response in the low frequencies" is meaningless.

I'm right there with you in principle--I just wish there was a feasible way to definitively show it up as bunk.

[bongomania]

Originally Posted by 12bass

If attempting to test cables using analog loopback, the results shown are probably mostly just the sound card's D/A-A/D figures. In comparison to the converters, I'd guess that the cable has a negligible measurable impact.

You're probably right, but the point of my exercise is to prove it.

Interestingly, while the THD remained nearly identical for all cables, the noise figure varied between them--and the results were consisently repeatable. So while the A/D doubtless has its own noise, and the A/D plus the software lacks the finesse to do "real" hair-splitting noise testing, I'm reasonably certain I am getting useful data of some sort.

Charlie, as to whether the loopback noise specs are bad, I don't remember what numbers I quoted the first time--but the current batch tests at between -62 dBA and -69 dBA. Please do let me know if that indicates a probable bad test.

[12bass]

FWIW, my modded Gina24 reports noise at around -110 dBA, THD as 0.0004 %, and IMD as 0.0014 %. This is in analog loopback mode using a short piece of Canare GS-6 with Neutrik 1/4" plugs. Not sure about the noise figures above... as they seem kind of high.

[bongomania]

Hmmm... Now I have to figure out whether the PreSonus has unacceptable noise, or if it's just very high EMF levels in my workspace causing those numbers.

[fdeck]

Originally Posted by bongomania

Hmmm... Now I have to figure out whether the PreSonus has unacceptable noise, or if it's just very high EMF levels in my workspace causing those numbers.

Is it a desktop computer? I found that my old HP desktop has more than one set of in/out jacks, and the front panel jacks had a lot more interference noise than the rear panel.

When I'm doing noise measurements, I've added a +20 dB gain amp to the input of my system, using an op amp with very low input noise. It's battery powered so I don't have yet another ground loop to contend with. And if possible I put the entire system being tested into a grounded metal box.

[bongomania]

The computer's a desktop, but the audio converter's an external module connected to the computer via firewire.

Now I just have to find a suitable metal box. What if I was to wrap the cables up in a lead apron, the kind x-ray techs wear?

[12bass]

Maybe try a digital loopback and short analog loopback to check the interface. Should be seeing -100 dBA or better for a 24-bit unit. My test above was 32-bit/48 kHz.

[Passinwind]

Originally Posted by 12bass

FWIW, my modded Gina24 reports noise at around -110 dBA, THD as 0.0004 %, and IMD as 0.0014 %. This is in analog loopback mode using a short piece of Canare GS-6 with Neutrik 1/4" plugs. Not sure about the noise figures above... as they seem kind of high.

Yeah, time to take a hard look at this I think. My EMU 1820 reads -115 to -120, but at the lowest figures cable orientation and stuff like that even shows up more than I'd like.

I'll post some more thoughts once I get my test jig up and running. It's been a few months since I've done any audio analysis work.

[bongomania]

A bit of Googling tells me the lead apron would not have helped. Instead, I just bought a 12" x 5" mu-metal cannister with 1/16" walls, off Ebay. That oughtta do 'er.

[fdeck]

You guys have me beat. I'm plodding along with the internal soundcard of my notebook computer, with around 95 dB dynamic range.

I'll have to look into a higher quality thingumajig.

[Codger]

Originally Posted by bongomania

With THD, all my tests of audio cables show a value of about 0.019%. This makes me think either my testing is flawed, or all of these cables have identically incredible THD. Which do you think, and should I be trying a different approach?

Is that the THD of your source?

[bongomania]

There's only one way for me to test that... by putting a cable from the input to the output.