Ok so a quick search proved my point: most serious reference ( from university research or AES published) only take into account the 500hz to 8khz ( more or less telephone bandwidth, human speech inteligibility range, iow the medium range). There is no question about audibility of gd in these range or lower, but there is no real results for threshold of low freq afaik.
Other answers about it are either opinion based, non verified talks or suppositions ( which can be great or interesting but are not significant).
And we are interested in low end in this case.
Other answers about it are either opinion based, non verified talks or suppositions ( which can be great or interesting but are not significant).
And we are interested in low end in this case.
This is exactly how most PA systems sound. You should be able to hear the drumstick hit the skin and the impact together. I have had someone tell me that on my PA (sealed 18" subs) you can tell if the drum skin is made of plastic or goat skin...
1 to 1.5 cycles is not resulting in your equation.
1 cycle at 30Hz is 33ms. But referring to your equation, 33 x 30 is 990, almost 2.5x.
I'm still interested in the source for the 400 number. It sounds plausible but there has to be some reasonable method to arrive at it. I looked at half cycle as being a tight target (not always achievable) out of instinct, your formula seemed to indicate it was about right. Is it just a rule of thumb then?
Or it could simply be that GD below a certain limit does not matter as wavelength is too long to reliably detect any additional delay. Thanks for the pdf. It is quite useful though I am not convinced of its rigor, however it's definitely a good starting point.
Allen I think the 1978 paper found that GD detection depends on environment as well as frequency, so there are detection threshold differences between semi-anechoic and reverberant spaces.
1 cycle at 30Hz is 33ms. But referring to your equation, 33 x 30 is 990, almost 2.5x.
I'm still interested in the source for the 400 number. It sounds plausible but there has to be some reasonable method to arrive at it. I looked at half cycle as being a tight target (not always achievable) out of instinct, your formula seemed to indicate it was about right. Is it just a rule of thumb then?
Or it could simply be that GD below a certain limit does not matter as wavelength is too long to reliably detect any additional delay. Thanks for the pdf. It is quite useful though I am not convinced of its rigor, however it's definitely a good starting point.
Allen I think the 1978 paper found that GD detection depends on environment as well as frequency, so there are detection threshold differences between semi-anechoic and reverberant spaces.
Every sort of filter adds GD. Starting with the box. Every manipulation of driver phase adds GD. For a recent project I experimented with a Peerless STW-350F in a simulator. Peerless recommend a sealed enclosure with added boost and a highpass. If you follow their recommendations to the letter, you get 23ms at ~24Hz. Without the recommended filters/EQ, the GD drops to 7ms at the same frequency. The EQ triples the GD.
For the Parts Express ported recommendations, we get 36ms at the same frequency, but that's not the peak. After removing the highpass and lowpass it drops to 24ms at 24Hz. Nobody seems to be complaining of "slow bass".
For the Parts Express ported recommendations, we get 36ms at the same frequency, but that's not the peak. After removing the highpass and lowpass it drops to 24ms at 24Hz. Nobody seems to be complaining of "slow bass".
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.