Where to get the most accurate test tones?

steveinaz

Ok, thinking my curve looks way to good I tried some tones from the www.snapbug.ws/bfd.htm site. The resulting curve was not nearly as flat as the tones I got from another site (the one Pjdami posted a link to). Though the "pattern" was similar, amplitude of the peaks and vallleys are more intense. The sinewaves from snapbug were kinda flaky, some 'popping" noises and such.

What's the best source for GOOD sinewaves?

pjdami

That is a very interesting observation Steve. I was kind of wondering if the different sine waves could make a difference if they were not originally recorded properly.

The http://beyond_gomer.tripod.com/

1/12 octave test tones from the above website is what I have been using and if I'm not mistaken that link was posted by Doc himself a while back. Not sure what Doc is using now... TrueRTA something I believe. He has a pretty fancy test rig that does the chirp and entire sweep and graph in a matter of seconds.

I'm sure that Doc will chime in soon. At best what we are using is crude (but still effective) compared to the sophisticated setup that he has.

steveinaz

Pjdami, the link you posted above is where my original tones came from, and what my curves posted thus far are based upon. I guess another solution would be Stereophiles test CD which has frequency sweeps that I would trust in a second. I may just go ahead and buy the Stereophile CD, I've always wanted it anyway...

steveinaz

Here's the 2 waves, first is "beyond gomer" second is "snapbug":

Dr. Spec

Yes, the method of plotting burned test tones with the RS meter and correction factors is "OK", but not great. There are too many variables and opportunities for error.

Clearly, your own experiment proves this. Nothing changed except the source of your test tones, and your FR now looks considerably different. The only logical conclusion is that the two sources of test tones are different.

This method is far better than nothing; I used it with success in the past, and it can definitely illustrate large peaks and nulls and allow correction of the same.

If you want to measure the "true" FR fairly accurately in both an absolute and a relative sense, then you need to spend some money - probably around $700-$800 for a basic test rig and associated support equipment. And even then, the professional rigs are better still.