When you're building a two-way speaker, what degree of attenuation do you prefer? Let's say you have a tweeter with 110 db sensitivity and a woofer with 95 db, a difference of 15 db.
Would you attenuate...?
- The entire 15 db for a mechanically flat response, which will probably sound muddy.
- About 9 to 12 db so that the resulting speaker is a little "brighter", but not overpowering.
- Something entirely different - explain why.
Would you attenuate...?
- The entire 15 db for a mechanically flat response, which will probably sound muddy.
- About 9 to 12 db so that the resulting speaker is a little "brighter", but not overpowering.
- Something entirely different - explain why.
Why does it matter what the original sensitivity of the tweeter is? does a 15db attenuation sound any different than a 3db attenuation resulting in the same frequency response/power response? Is it measurable as distortion? not necessarily disagreeing with you just like to get this stuff since I have thought about it myself.
The L-Pad adjustment has only to do with the volume of the tweeter relative to the woofer.
It matters because you drive a speaker with one source, and if the tweeter has an additional 15 dB of sensitivity over the woofer, it will be playing 300% louder, so you'd have a very "bright" speaker which most people would describe as tiring to listen to. It will literally give most people a headache.
A 15 dB attenuation, assuming that was the difference in sensitivity between the woofer and tweeter, would flatten out the mechanical response, making the speakers ideal for tweaking with an EQ. That is what I did with a recent project, knowing that I was going to be using an EQ to tune the speakers to the room.
I did end up doing that, but I don't think it's the best way to go for every situation. I think you want 3-6 dB of additional sensitivity in the tweeter over the woofer unless you are certain you will be using an EQ.
The relationship is something like this:
Supose your speaker can do 10 dB @ 1M at 5,000 Hz:
One octave up to 10,000 Hz will mean the SPL is 7 dB @ 1M
One octave down to 2,500 Hz will mean the SPL is 13 dB @ 1M
Two octaves up to 20,000 Hz means SPL is 4 dB @ 1M
Two octaves down to 1,250 Hz means SPL is 16 dB @ 1M
When you go up an octave the relative volume of the sound is halved as the frequency is doubled, assuming a fixed listening point and power source.
So I think you'll want a multi-way speaker to have an "average" sensitivity rather than a perfectly matched sensitivity. In the example above, allowing the tweeter to retain 6 dB of sensitivity over the woofer should yield a far better sounding speaker without the need for an EQ.
Also, I think human ears tend to pick low frequencies with greater ease than high frequencies, so, generally speaking, 3-6 dB of additional sensitivity on the tweeter may be optimal.
I do not know what you mean by BSC.
I thought you meant that the amount of attenuation in db makes it sound worse not depending on the freq curve. like something in the electrical function of a heavy l-pad that introduces distortion.
Since i get now that your talking about a desired fequency response my thought would be that a 3db hot tweeter should sound really bright, but if you like that go with it. It's your speakers. Have you measured the speakers? It could be more about how the x-over sums
BSC is baffle step compensation
Since i get now that your talking about a desired fequency response my thought would be that a 3db hot tweeter should sound really bright, but if you like that go with it. It's your speakers. Have you measured the speakers? It could be more about how the x-over sums
BSC is baffle step compensation
Last edited:
No, I was just curious to see how others were approaching this in terms of how much L-Pad attenuation they would add. If done correctly the L-Pad should not introduce any distortion since it's just resistors in line with the tweeter. To me, muddy sounding speakers have poor high freq response...so the sound is dull and lacks that spacious quality that I like.
It might sound bright, but I think +3 dB is a modest gap. I find that I'm always cranking up treble at least +3 dB on any speakers, so I guess that means I prefer my sound to have some extra zing on the top end.
I have not measured the speakers, but I am very pleased with how they sound. As mentioned, what I ended up building was a xover whose L-Pad attenuated the total gap between the woofer and tweeter. My reasoning was that it would make it easier to adjust the sound via EQ to get it "just right".
Good to know. I might pull my xovers out to try this and see if it improves the low end performance.
Get a pair of variable Lpads and wire them in. If you have a ported box all the better, then you can have the wires through the ports and adjust the Lpads outside the box. Adjust for best sound.
Sorry to say that, but IMHO this is pure nonsense. A speaker that works as you say is awful, and isn't working in a natural way but needs to be equalized to do what you're saying. For a normal non-junk tweeter, say one that has a FS of 500-1000 Hz, if it will do 10 dB at 5KHz, then it will also do 10 dB at 10KHz, 20KHz, 2.5KHz, and probably the same or a bit less at 1250Hz. For a woofer it is another thing as we are talking about a range where breakup and rolloff are predominant, every driver is different and we need to look at the FR.
For your main question, not counting the BSC, if the tweeter has 15dB more sensitivity than the woofer, then in order to have a natural sounding speaker you need to lower the tweeter by 15dB. Some people argue that the best sounding speaker don't have a flat FR, but slightly tilted with the tweeter 2-3dB less than the woofer, so in this case you need 17-18dB of attenuation. If you feel that this attenuation makes the sound muddy, then probably you should chose a different tweeter. I know (*) that if the tweeter is some 2-3dB more than the woofer, then it will be bright, appealing at first but tiring after some time, giving listening fatigue. So a question comes to my mind, do you have some form of hearing loss at high frequencies?
Ralf
(*) By experience, playing with attenuation on some crossovers.
Higher frequencies use less energy to produce the same sound pressure levels, so you can expect a driver to be flat in its passband.
This will depend on the driver being used correctly and on the shape of the cabinet.
With manufacturers rated sensitivities, they should match when you use the drivers in the same way as each other.
So you would make the tweeter sensitivity a few dB below that of the woofer instead of BSC? Because it seems like if you did that along with BSC the speaker would be heavily biased to the lower end. Would the ideal setup be matching tweeter to woofer and using BSC?
No hearing loss as far as I know. I've always been sensitive to high frequencies, and I think my preferences for extra treble is just that - a preference. Too much and it causes fatigue for me as well.
I never bothered with BSC but then all my speakers find themselves with their back to a wall, on a desk or even in a corner but never free-standing.
I usually have to reduce the bass output to get a flat in-room response.
My main speakers now sit in corners and I had to knock nearly 9dB off the bass to get a +-2dB response at my listening position.
I usually have to reduce the bass output to get a flat in-room response.
My main speakers now sit in corners and I had to knock nearly 9dB off the bass to get a +-2dB response at my listening position.
The amount of BSC is not always 6dB, even if the loss amounts to that value. Depending of the speaker location a "normal" BSC could be about 3-4dB, as there will be bass reinforcement from walls. You can aim for a full BSC with a big room and speakers far from any boundary. On the contrary no BSC means quite often if not always a thin bass sounding speaker.
With the BSC you set, you are setting also the woofer level. Say you have a 90dB woofer and decide that (*) 3dB of BSC is enough, then you'll end up with a 87dB woofer, and this should be the reference for the tweeter, or a couple dB less. As an example of a FR of this kind see here in the Measurements section: Compact-Studio-Monitor
Ralf
(*) Depending on the FR and placement.
- Status
- Not open for further replies.