In the same "6 inch driver playing music pop" I got to -24dB just by turning headphone volume down. The problem seems clear, distortion products of the test become "buried" under the distortion products of my headphones.
I also experienced the same "discontinuity" in the 12dB or 15dB region, it seems that the distortion products of my headphones become similar in magnitude and opposed in phase to the ones from the test resulting in distortion cancellation that fools you. Reversing the polarity of the headphones may change things for even harmonics...
I also experienced the same "discontinuity" in the 12dB or 15dB region, it seems that the distortion products of my headphones become similar in magnitude and opposed in phase to the ones from the test resulting in distortion cancellation that fools you. Reversing the polarity of the headphones may change things for even harmonics...
While I completely agree with you cancellation likely occurs it's a very contentious issue because of what it says regarding ideal amplifier performance and speaker matching. Are you aware of any papers or studies examining it?
I managed to get all the way up to the -36... without missing a single trial (on my second run through the test...)
It was strange... I couldn't always "hear" the difference... but It was always obvious to me which was distorted (to a lesser degree on the highest settings)... One just sounded "wrong"...
M-audio external sound card... BeyerDynamic DT770 headphones- reasonably loud...
It was strange... I couldn't always "hear" the difference... but It was always obvious to me which was distorted (to a lesser degree on the highest settings)... One just sounded "wrong"...
M-audio external sound card... BeyerDynamic DT770 headphones- reasonably loud...
Attachments
Excellent!
That's the hard one too. I can only ever get to -18 on that one.
How well did you do on the first test? (music-pop).
That's the hard one too. I can only ever get to -18 on that one.
How well did you do on the first test? (music-pop).
It has to be easily repeatable in order to discard the purely random probability of reaching a lower level, that is not so low because during the first approach you move in 6dB steps.
You will appear 6dB better than what you are in one out of every two tests.
You will appear 12dB better than what you are in one out of every four tests.
You will appear 18dB better than what you are in one out of every eight tests.
Everybody can get to the minimum level by repeating the test enough times, but it's the worse result the one that tells how good actually you are.
With the proper headphone volume setting, I think that now I can get to -18dB without mistakes almost every time.
You will appear 6dB better than what you are in one out of every two tests.
You will appear 12dB better than what you are in one out of every four tests.
You will appear 18dB better than what you are in one out of every eight tests.
Everybody can get to the minimum level by repeating the test enough times, but it's the worse result the one that tells how good actually you are.
With the proper headphone volume setting, I think that now I can get to -18dB without mistakes almost every time.
Eva said:
I'm not sure that I agree with this. There are time when I've taken this test and I've been zoned in - I can easily hear the difference. Other times, due to a list of reasons such as distraction, I can't determine a difference at all.
I think your best score, if you get it more than 10% of the time, is how good you are at distinguishing distortion.
Eva said:
This test is tough to do over and over again. You're ability to distinguish distortion will match your concentration and focus.
Here's a thought: the average of 5 consecutive tests should give a good indication of your ability.
The best I can do with the "octave", Audigy2 and crapy Sony headphones:
On music and 6" speaker I did stop at -24dB...
An externally hosted image should be here but it was not working when we last tested it.
On music and 6" speaker I did stop at -24dB...
Very interesting test. The lower I go, although a difference is heard, which is distorted is in question. M-Audio headphones that came with a card was used. If I guess wrong the first time, normally I can go lower on the next run. I would be interesting to see how speakers work.
The gig is up!
Here are my results:
-45 dB on "octave - 6 inch driver"
Onboard sound on Dell Optiplex GX280
Koss UR-19 Headphones ($5 at TJMAXX) with intermittent left speaker.
Fan running in adjacent cubicle.
How did I do it?
There's a glitch in this particular setup. Judging by the disproportionate number of users who attained the -45 level in this test, I don't think I'm the only one who noticed.
The undistorted sound track has a series of two clicks just a few seconds after the start. The distorted one does not have these clicks.
I sent an email to their webmaster.
On a valid test "pop music on 6 inch driver" I reached -21dB on second try on the same setup as above. Can't wait to try it out at home with a lower noise floor.
Thanks for the link John. I'm eagerly following your 3 way thread.
Regards,
David
Here are my results:
-45 dB on "octave - 6 inch driver"
Onboard sound on Dell Optiplex GX280
Koss UR-19 Headphones ($5 at TJMAXX) with intermittent left speaker.
Fan running in adjacent cubicle.
How did I do it?
There's a glitch in this particular setup. Judging by the disproportionate number of users who attained the -45 level in this test, I don't think I'm the only one who noticed.
The undistorted sound track has a series of two clicks just a few seconds after the start. The distorted one does not have these clicks.
I sent an email to their webmaster.
On a valid test "pop music on 6 inch driver" I reached -21dB on second try on the same setup as above. Can't wait to try it out at home with a lower noise floor.
Thanks for the link John. I'm eagerly following your 3 way thread.
Regards,
David
Attachments
67" Samsung HL-R6768W dlp hdtv that I use as my computer monitor. Today I had this old HP from around the year 2000 hooked up with its horrible onboard sound, which was used for the test. The TV speakers were used for the test.. two 3" mids and one 2 or 3" paper cone tweeter on each bottom side. Despite the original price tag of this TV back in 2005 ($7,000 i think, but I didn't pay anywhere near that when I bought it refurbished on eBay in 2006), its audio is absolutely terrible. Oh, and the built in EQ via the setup menu is a little screwed up too, as it sounds horrible no matter how I adjust it. There's also a Magnepan MMG-W leaned against each side edge of the TV, completely blocking the tweeters. Another pair of speakers I've been toying with are sitting less than a foot in front of those, almost completely blocking the TV speakers..
-18 dB on the 6 inch driver with pop music test.
I bookmarked it for a later time when I'm MUCH more prepared. 😀
-18 dB on the 6 inch driver with pop music test.
I bookmarked it for a later time when I'm MUCH more prepared. 😀
the "bimodal" distribution with the 2nd peak at the test floor suggests that several tests here likely have this sort of "clue" that allows determination independent of the varying distortion level
Somehow made -30db using my laptop's Realtek AC97 audio, Sure Class D amplifier with a cheap wall wart, and a really old 3-way speaker that even has a cracked surround on the woofer.
Now to try this with headphones and a decent sound card...
Now to try this with headphones and a decent sound card...
It is a shame about those "clues". I don't know how much confidence you can put in the results with errors like that. I guess that is where statisticians step in and do their magic.
Some people should be ashamed for exploiting these loopholes and misrepresenting their hearing acuity.
Some people should be ashamed for exploiting these loopholes and misrepresenting their hearing acuity.
From what I can tell, the "pop music on 6 inch driver" test contains no such clues.
The "octave" test as well as a couple other ones do however have unwanted clues. You can tell from the stats too.
The "octave" test as well as a couple other ones do however have unwanted clues. You can tell from the stats too.
- Status
- Not open for further replies.