Jump to content

dB Addition Math - Does not seem to correlate with actual listening - Why?


radiogram

Recommended Posts

The way the dB output of two drivers sum by math, does not seem to correlate with listening when using higher slope networks Vs lower order in terms of tonality. Of course, there are other advantages of higher order networks that consequent to reduced region of overlap lead to reduced combing effects and such. But for this thread I am confining the scope to ONLY SPL Vs Freq Response, since that primarily dictates perceived tonal balance.

For example, Let us take the following hypothetical example for simplicity:

- A woofer with a sensitivity of 90dB with a flat bandwidth of say 20-1000Hz.

- A midrange driver with a sensitivity of 90dB with a flat bandwidth of say 300-5000Hz.

- The midrange has a first order high pass circuit with a crossover (-3dB) freq of say 700Hz which is well within the drivers’ inherent passband.

1. Let us say we first implement some first order network for the woofer and let us say in doing so the output of the woofer is about 84dB (-6dB down) at say 1000Hz.

2. Let us take another case where we implement a higher order low pass such that the woofer’s output at 1000Hz is say 80dB (-10dB down).

3. Let us take another case where we implement a even higher order low pass such that the woofer’s output at 1000Hz is say 70dB (-20dB down).

4. Let us take another case of extreme slope, where we implement a higher order low pass such that the woofer’s output is say 40dB (-50dB down)

Now let us sum the total spl in both cases:

1. In case 1, the total output is 84dB + 90dB = 90.97

2. In case 2, the total output is 80dB + 90dB = 90.41

3. In case 3, the total output is 70dB + 90dB = 90.04

4. In case 4, the total output is 40dB + 90dB = 90

According to the above, whether the woofer is at 10db, 40dB, or 70dB or 80dB the total spl is always 90dB, implying that a higher order network has no effect on ultimate spl due to overlapping driver output. But, clearly I hear difference between a third order and 1st order in such a situation. Of Course, we have less of combing effect and such at the overlap region with higher order that contributes to listening differences. But I am, as I said confining to tonality issues alone that is primarily determined by SP Vs Freq and I do hear difference in tonality.

Can anyone explain this?

Link to comment
Share on other sites

For example, Let us take the following hypothetical example for simplicity...

Are you just doing calculations, or are you implying that these are measurements? Are you assuming broadband pink noise?

Chris

Link to comment
Share on other sites

I do not have the equipment to do proper A/B compariosn. But in any case My listeing is not just the woofer, it is integrtaed as a whole.

Anyway, the main point here is if adding 80dB or 60dB to 90dB results in mathematically to only 90dB, what is is the advantage of building complex higher order network from a freq response stand point?

Link to comment
Share on other sites

I assume from this that you are using a hand-held SPL meter, right? Is it on the "A" or "C" scale? Are you using music or pink noise?

Anyway, the main point here is if adding 80dB or 60dB to 90dB results in mathematically to only 90dB, what is is the advantage of building complex higher order network from a freq response stand point?

I think that you are assuming that a static measurement SPL is a good measure for higher-order crossover design. I'd use band-pass-limited diffraction patterns on SPL in the bandpass region of the crossed drivers only. You can do this by moving around the room, looking for differences in SPL (azimuth, elevation from the centerline of the speaker(s)) and using pink noise input.

Chris

Link to comment
Share on other sites

The sensitivities are probably based on something like 1 watt @ 1 meter. Well when you build up a high order network like and ESN you have more losses in the network than a first order network for instance because you have many more components. You are probably not feeding the same power levels to the drivers in each of the tests. That may be why you hear the differences. I definitely needed more power when I used to use ALK's ESNs. Maybe I should say it took more power to achieve the same listening levels when using the ESNs over lower order networks. I know nothing changed in my drviers so the power had to be eaten up in the network.



But what would you hear if you could assure the drivers each saw the same exact power? Your math and your ears might both be right. You need ALK to answer this. He will know.

Link to comment
Share on other sites

I'm not sure I follow everything in your post, but I would care to guess there's a lot of stuff going on that's going to effect the result. At the very least, I would point to the phase relationship between the woofer and midrange. This will dictate the amount of "coupling" going on between the two drivers in the crossover band. As you change the slope of the woofer filter, the phase relationship of the two drivers will change (relative to each other) such that the combined output will vary according to whether the drivers are "in" or "out" of phase with each other.

dbspl

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...