Crossover Group Delay Audibility Testing - please take part!

charlielaub

Hello All!

I'm doing well and far from dead and gone, I'm just living far away in Germany! I wanted to post this for all of the dedicated DIY loudspeaker builders and designers that frequent this forum, since this topic should be relevant.

I have just launched a suite of online tests where you can try to distinguish between various levels of group delay that have been added to several different signals (both synthetic and music types).

These tests are quite difficult, even with the synthetic signals. I have added group delay in levels from inaudible to audible, based on the published literature.

PLEASE ENCOURAGE OTHERS TO ALSO TAKE THESE TESTS.
WHY?
ANSWER: The audibility level for group delay in loudspeakers is an important parameter that puts strict limitations on crossover filter design. Once I have a sufficient number of respondents I will analyze and share the data and full details on the adulterations used to generate the test signals. This information will hopefully be useful to the loudspeaker design community at large.

GROUP DELAY TESTING MAIN PAGE:
http://audio.claub.net/GD_testing/GD_Testing_Main_Page.htm

Steve_Lee

Nice test and query UI - very intuitive/easy to follow.
The lower frequencies are the only places I could really spot shifts worth selecting as "different".
Looking forward to your test results and ultimate decision concerning the audibility threshold of GD.

Eggguy

I do not have a thorough understanding of group delay. It seems to me that a better term would be cumulative delay.

Eggguy

According to what I have read, it can be introduced by microphones and electronics with a low end roll off, and it accumulates until it hits our ears. It is measured in the time domain, milliseconds.

4thtry

@Eggguy said:
I do not have a thorough understanding of group delay. It seems to me that a better term would be cumulative delay.

Yes, the more phase rotations you have, the higher the cumulative "group" delay. A speaker with an 8th order crossover at 1kHz will have roughly 0.5ms of group delay throughout the crossover region, whereas the same speaker with a 4th order xover will only have about 0.25ms of group delay. Group delay, when measured in milliseconds, then starts to increase substantially below 100Hz due to the much longer wavelengths involved. I don't know if I can actually hear this, I have not taken the test yet. Before taking the test, I need to get my system hooked up to some better speakers, so if there is a difference, I can actually hear it.

Steve_Lee

I really think that this needs to be listened too with a near-zero GD system --> ie// headphones or a sealed full range driver alone but that's just me thinking again . . .

Our systems have their own GD and it will either add or subtract from the contrived GD in the sound tests, is my thinking.

charlielaub

Thanks for all the comments. The testing is coming along slowly. I could still use a few more responses to get better statistics. You are welcome to retake the tests, and please encourage new people to give it a try.

These first tests rely on audio memory (a few seconds of that) in order to compare the reference and modified audio to try an identify audible differences. Because the differences are subtle this adds some extra difficulty. I am currently trying to create the ability to instantaneously switch between the two tracks, and I have a demo/test page up with that working but without any voting. Link:
http://audio.claub.net/GD_testing/test3/PinkImpulse_InstantSwitching.html
The way this one works is that each line corresponds to a different modification of the audio. When you click the begin playback button the unmodified (reference) audio begins to play. Click the TOGGLE button to switch back and forth between that and the line's modified audio track. When you are done on that line, click the Stop Playback button. Eventually there will be voting checkboxes as well.

It should be easier to hear differences when you can quickly compare in this way.

Let me know what you think!

charlielaub

I have computed the statistics for TEST #1 (the square wave) since it reached 50 returns. Sorry to just post a link here but that is easier than reposting everything! See:
https://www.diyaudio.com/community/threads/crossover-group-delay-audibility-testing-please-take-part.416625/post-7774676