almikel Posted April 23, 2022 Posted April 23, 2022 (edited) So I get that to best measure speakers, outdoors away from boundaries and as high off the ground as possible is recommended to provide as much distance from 1st reflections as possible... ...a notable exception being measuring subs outdoors, where a common approach is to have the sub and mic on the ground, so no ground bounce. We've all seen this diagram before: The 1st reflection arrives at the mic in (2*y/343) seconds, and plenty of sources on the interwebs discuss the "lowest reflection free frequency" as = 1/(2*y/343) - maybe fine...especially when h is larger than x ...but my question relates to what happens when x is much larger than h, where the path difference (ie 2y-2x) is relatively small. Take an example where the driver and mic are 1m off the ground (h) and the mic is 20m from the driver (x=10m) Pythagoras says y= 10.05m The previous logic of the lowest reflection free frequency = 1/(2*y/343) would give the "lowest reflection free frequency" as 17Hz...but without catering for the long propagation time of the direct wave. The direct wave gets to the mic in 58.31 ms The ground bounce wave gets to the mic in 58.60 ms - with less than (edit) 0.30ms of real data between the direct wave and the 1st reflection. Are the interweb sources that say "the lowest reflection free frequency = 1/(2*y/343)" oversimplifying? Is there really any accuracy down to 17Hz? Is there a better calc that incorporates the difference in propagation times between direct and reflected sound to determine low freq accuracy, rather than using the inverse of the propagation time for the 1st reflection?... ...or are we coming back full circle where h can reduce to 0 like measuring subs on the ground plane as above? Obviously the goal for measuring speakers outdoors is to have x at an appropriate distance to incorporate diffraction effects from baffles etc, and have h (and distances from other boundaries) as high as possible. Clearly the key data is captured only after the direct sound arrives at the mic and decisions need to be made on how the measurement is used after the 1st reflection arrives at the mic (eg choose to make or not make an EQ correction knowing reflections are in the measurement). I need some schooling...thanks in advance... cheers Mike Edited April 24, 2022 by almikel typo
frankn Posted April 24, 2022 Posted April 24, 2022 The biggest problem is how large in Msec can you make the window for the analysis of the reflection free data. That’s why you want the largest “free-field” setup. That can be problematic in the real world - getting a really quiet area to record in and having the vertical and horizontal space to setup the experiment. There are other techniques to accomplish this if you want. these are discussed in articles on AudioXpress - they have quite a library. This one I think will help. There are several others discussing some of the questions you raise. https://audioxpress.com/article/measuring-loudspeaker-low-frequency-response 1
almikel Posted April 24, 2022 Author Posted April 24, 2022 7 hours ago, frankn said: This one I think will help. There are several others discussing some of the questions you raise. https://audioxpress.com/article/measuring-loudspeaker-low-frequency-response Thanks @frankn That link shows the lowest resolution freq = 1/(time diff between 1st reflection and direct wave) That makes more sense. cheers Mike 1
davewantsmoore Posted April 26, 2022 Posted April 26, 2022 (edited) On 23/04/2022 at 11:10 PM, almikel said: I need some schooling...thanks in advance... What is it you are exactly trying to do, by capturing "reflection free" data from a subwoofer? Confirm that the subwoofer you built matches what you designed? ... You can do this with a reasonable amount of reflections included in the data well enough. Trying to choose some correction filter for the (free field) subwoofer..... I would base this on the simulation of the woofer, not on the actual measurement...... and to a fairly big degree, I really wouldn't even bother with this (just do in room correction) Trying to choose an in-room correction for the subwoofer? Use in room data. Otherwise.... like you have already calculated, you can see that with a reasonably big X (eg. >5m) and a reasonable small Y (eg. the subwoofer just in its cabinet sitting on the ground as normal.... so H is << 1m) ...... then you are going to get a frequency that is close enough to the lowest mode of a room, that it's fine (as the room completely dominated the SPL propagation down that low). Said another way.... instead of asking questions like "how to I really accurately measure my subwoofer" ..... try asking "how do I figure out X" (the thing you actually wanted to solve using the "really accurate" data). Edited April 26, 2022 by davewantsmoore
almikel Posted April 28, 2022 Author Posted April 28, 2022 On 27/04/2022 at 9:24 AM, davewantsmoore said: What is it you are exactly trying to do, by capturing "reflection free" data from a subwoofer? Hi Dave, My question was not directed at subs specifically, more related to the measurement setup required to achieve reflection free measurements for main speaker correction...and some interweb sources had me confused...now corrected based on @frankn's link . Other interweb sources I'd read only discussed when the 1st reflection arrived at the mic as defining the lower limit of accuracy...which didn't sound correct... The article linked by @frankn https://audioxpress.com/article/measuring-loudspeaker-low-frequency-response, written by the respected Joe D'Appolito, defined the lower limit of accuracy as the time difference between 1st reflection and direct wave, which of course makes sense. On 27/04/2022 at 9:24 AM, davewantsmoore said: Trying to choose an in-room correction for the subwoofer? Use in room data. agreed cheers Mike
Recommended Posts