Subwoofer wiki!

Status
Not open for further replies.
Jump to Latest
#21
damo21 said:
Hey Paul, some dude from Netherlands seems to have a more accurate model of GD threshold... GDmax(ms) = 400 / f (hz)
Click to expand...

Greets!

Bell Labs/W.E. did extensive testing in the mid '30s and concluded that anything < 3 ms in the 350-800 Hz BW where they XO'd their various studio and cinema horn combos was inaudible. Factor in that our ears are very non-linear and that I could hear as little as 0.1 ms difference at 2 kHz when I was much younger, I don't put much stock in this 'model'.

As always though, YMMV.

GM
 
#23
GM said:
Bell Labs/W.E. did extensive testing in the mid '30s and concluded that anything < 3 ms in the 350-800 Hz BW where they XO'd their various studio and cinema horn combos was inaudible.
Click to expand...
The model suggested by Rademakers: GDmax=400/f is in agreement with this, in fact it suggests a lower threshold in that frequency range to be even safer.

GM said:
Factor in that our ears are very non-linear and that I could hear as little as 0.1 ms difference at 2 kHz when I was much younger, I don't put much stock in this 'model'.
Click to expand...
The model suggested by Rademakers almost fits your experience, suggesting less than 0.2ms at 2kHz. The proposed model is quite non-linear aimed to match the differing ear sensitivity to group delay for various frequencies. Keep in mind, this thread was focusing on GD below 60Hz so the fact that the model almost fits your experience at 2kHz is pretty good.

Please see my earlier graph.
 
#24
GM,

If you have a look at the article by John Murphy you will see that the Bauert and laws study tends to agree at that level. As you go down in frequency, there is a relationship between GD and frequency that is not linear. Rather, in each range covered they looked at cycles, with about 16 cycles at the top end and about 1 or 2 down low.

If you look into it, the values I put down aren't unreasonable, however, I'm not going to stand behind them if someone does a credible study showing otherwise. I just think this is the best that we have to go with so far that I know of.
 
#26
damo21 said:

The model suggested by Rademakers: GDmax=400/f is in agreement with this, in fact it suggests a lower threshold in that frequency range to be even safer.


The model suggested by Rademakers almost fits your experience, suggesting less than 0.2ms at 2kHz. The proposed model is quite non-linear aimed to match the differing ear sensitivity to group delay for various frequencies. Keep in mind, this thread was focusing on GD below 60Hz so the fact that the model almost fits your experience at 2kHz is pretty good.

Please see my earlier graph.
Click to expand...

Greets!

Hmm, 600%/800 Hz, 100%/2 kHz difference respectively fits Bell Lab's and my experience?! I think not! Since our hearing acuity rapidly falls off below ~200 Hz, it seems reasonable to me that the formula can only increase in error with it, but like I said, YMMV.

As for 'safer', this implies that you're willing to give up considerable acoustic gain down low where we typically need it most to keep GD low, just on the off chance you might be able to hear it. This seems 'penny wise and pound foolish' to me, but again, YMMV.

GM
 
#27
paulspencer said:
GM,

If you look into it, the values I put down aren't unreasonable, however, I'm not going to stand behind them if someone does a credible study showing otherwise. I just think this is the best that we have to go with so far that I know of.
Click to expand...

Greets!

?! What is it with you guys? Where have I made any comments WRT what you've posted on the wiki, which I haven't browsed yet BTW?

GM
 
#28
Here is the original article on which those values are based:

http://www.trueaudio.com/post_010.htm

scroll down half way and you find a table of "cycles on delay"

8k > 2ms > 16 cycles
4 > 1.5ms > 6
2 > 1 > 2
1 > 2 > 2
.5k > 3.2 > 1.6

Notice how the cycles go down. Now we have 2 octaves to extrapolate down before we get to 120 Hz and 3 to get to 60 Hz.

John says:


Extrapolating the Blauert and Laws data to lower frequencies it could be that "one or two cycles" of delay continues to be the threshold as we move lower in frequency.
Click to expand...

and further ...

My current opinion would go something like this: "One cycle or so of delay in the bass range is probably just audible under controlled listening conditions (phones) and with the most challenging program material (clicks and pops). Under less well controlled listening conditions and with less difficult program material the audibility thresholds are somewhat higher."
Click to expand...

I have used the one cycle of delay guestimate to derive those figures.

Bell Labs/W.E. did extensive testing in the mid '30s and concluded that anything < 3 ms in the 350-800 Hz BW where they XO'd their various studio and cinema horn combos was inaudible.
Click to expand...

Bauert and Laws came up with 3ms for 500 Hz which seems to be fairly close agreement.

How the data can be extrapolated is another matter, and if indeed it can be extrapolated at all.

I find it surprising that no one has done a study on this for bass frequencies.

For now, this is the best that I can find. I'm wary of rule of thumb guidelines here as I don't know what they are based on.

I'd like to know how much GD is audible down low, since like GM indicated, there's no sense giving up low end extension to get low GD numbers if you can't hear the difference.

EDIT: I have a spreadsheet showing the extrapolation and anyone is welcome to see it and comment. Just shoot me an email. It should be noted that a different extrapolation can give very different results, and how you do this may effect whether or not the GD in a particular design is considered audible or not. The values in this wiki suggest that it's possible to get down low with a vented box and for the GD to be inaudible.
 
Status
Not open for further replies.
Top Bottom