Rightmark result interpretion question


2011-05-13 11:33 am
Hi everyone.

After 25 years i finally got the chance to measure my old trusty line amplifier. It was created using only ohm's law, my ears, dmm and music, and does not make use of feedback.

Last night I borrowed a Soundblaster Pro X-Fi 5.1 usb soundcard and used rmaa with asio4all drivers setup. The test includes all of the amplifier, meaning everything from phono connectors, volume pot, amp itself, output relays.
I then setup a loopback test to see what the soundcard was capable of. I ran these tests a few times to see if I got consistent results, which i did. All tests run at 24bit 96KHz.

And here the question arises:
If you look at the attached picture, you can see that distortion about doubles from what the soundcard has on its own. I spoke to a friend about it, who thinks logic dictates that the soundcard thd should be subtracted.
But that doesn't seem right to me. If the harmonics for the soundcard and the lineamp are distributed the same way, would the harmonics from the amp not mask those from the soundcard? Or do they still add up ?

I hope someone can enlighten me and my friend here


  • Lineamp.PNG
    37.5 KB · Views: 115
No, it's not generally that simple. It's more like vector addition of the individual harmonics. Even seeing less distortion is not out of the question of the nonlinearities complement each other.

For further insight, I would generally look at the graphs as well. Cancellation tends to show up in the relative amplitude of harmonics.

Looks like you would at the very least have to use the Mic In to get a good idea about output noise level.


Paid Member
2004-10-07 6:05 am
I don't think they would cancel out. In fact there is a thread here somewhere about doing just that, using math to remove the soundcard from the measurement.

Using an FFT plot can show you a lot. ARTA is a good software for that. You can easily look at the noise floor, noise spikes, and individual harmonics. Testing at different signal levels and frequencies is also useful. You'll keep yourself occupied for days. :D


2011-05-13 11:33 am
Thank you Pano / Dankeschön sgrossklass

I will install Arta and take some screenshots of it's FFT window when I get the chance. ( at 1KHz and 10KHz if Arta allows that )
I'm thinking one image of the soundcard and one of the lineamp at each frequency.

Pano, if you stumble accross the thread you mention, would you please post it here? It would be fun to try if I could learn how to do the math

Regards and merry chrismas

Dont trust atom's. They make up everything.
Last edited: