Achieving "perfect" power response

[diy speaker guy]

I think it's agreed that the "perfect" power response curve looks like it has gentle, uniformly decreasing spl as frequency rises. Dantheman has posted several measurements of commercial mini monitors that seem to do this extremely well. What's the secret?

I'm guessing that it probably has a lot to do with shallow xo slopes, I'm not sure how else I might begin to combat "tweeter bloom".

[Zilch]

Re: Achieving "perfect" power response

It's about directivity.

It defines, that, actually -- the difference between the on-axis response and the power response is the directivity index. Thus, your "gently sloping power response" equates to a gently rising DI and implicitly, dispersion gently narrowing with increasing frequency.

We cannot "fix" directivity with EQ.

We can, however, control directivity, and thus power response, via other means.

In this, it's important to distinguish between anechoic vs. in-room response. DI and sound power are anechoic parameters....

[diy speaker guy]

Re: Achieving "perfect" power response

Ok, so is it possible to control our directivity without using waveguides?

[Æ]

Re: Achieving "perfect" power response

Perfect power response, would be an omni directional loudspeaker. If there is such a thing as perfect, it would be point source and radiate equally in all directions, a bit like real sound does.

[Zilch]

Re: Achieving "perfect" power response

Ok, so is it possible to control our directivity without using waveguides?

Yes, with varying success. See Linkwitz Orion and Revel Salon2. Both apparently fail in the top octave, Linkwitz by default to a dome tweeter and Revel by design with a waveguide, though whether that is to significant detriment is arguable.

Perfect power response, would be an omni directional loudspeaker. If there is such a thing as perfect, it would be point source and radiate equally in all directions, a bit like real sound does.

[Deward Hastings]

Re: Achieving "perfect" power response

so is it possible to control our directivity without using waveguides?

Yes, in a number of (quite different) ways. Dipole cancellation "controls" directivity quite well, especially at lower frequencies, though only to one "pattern". Line arrays "control" directivity. and with some flexibility if you include time delay to different drivers. Even the simple MTM "controls" directivity to some extent by selective reinforcement and cancellation at different angles/frequencies. Selective absorption can control directivity, especially at higher frequencies. And individual diaphragm radiators "control" directivity as the wavelength radiated approaches 1/2 the diaphragm diameter. The simple expedient of putting a large(r) baffle on a sealed box can "control" directivity by lowering the baffle step frequency. Even the positioning of the speaker in the room (corner loading, for example) can be said to "control" directivity to some extent, as can room "treatments" which absorb radiation at undesired angles (a special case of selective absorption).

And I'm sure there are others I've missed . . .

[timw]

Re: Achieving "perfect" power response

One of biggest and most audible problems was caused by crossing over too high for the woofer, in the frequency range where its dispersion was rapidly decreasing. The off axis response would have a sag around the crossover point, for example say 3 to 4 kHz. As the tweeter took over with its greater dispersion, the off axis response would rise to about 5 kHz then fall again after 10k.

The larger the woofer, the lower the crossover point should be for good power response.

[billfitzmaurice]

Re: Achieving "perfect" power response

Perfect power response, would be an omni directional loudspeaker. If there is such a thing as perfect, it would be point source and radiate equally in all directions, a bit like real sound does.

'Real' sound isn't omnidirectional. Natural sound sources are subject to the same frequency versus polar response limitations as loudspeakers. That makes achieving 'perfect' reproduction and power response an impossible task, so one shouldn't lose sleep over the inability to do so.

[jeff_free69]

Re: Achieving "perfect" power response

'Real' sound isn't omnidirectional. Natural sound sources are subject to the same frequency versus polar response limitations as loudspeakers. That makes achieving 'perfect' reproduction and power response an impossible task, so one shouldn't lose sleep over the inability to do so.

Another important consideration is the fact that more often than not the recording was not mic-ed and recorded in a "perfectly natural" way.

Maybe that would mean using a minimalist stereo-coincident mic at the distance similar to listening position.

Such recordings are rare. While they tend to have a great imaging and soundstage, they often lack articulation and clarity most people expect. Even a typical acoustic piano jazz recording usually has mics up close on each instrument, in addition to room mics to fill in the ambience. The point is that most of what you hear is not the perfect power response of the instrument; it is usually focused on the point of sound generation eg hammers on strings, this is supplemented by another set of microphone several feet away or more.

If the engineer did a good job, the illusion of a real performance in a room is funneled through 2 little speakers. And I think illusion is an appropriate term to always keep in mind. That said, I'm currently interested in exploring dipoles. I'm not up on the math, but i don't know if the illusion requires mathematical perfection.

[billfitzmaurice]

Re: Achieving "perfect" power response

Another important consideration is the fact that more often than not the recording was not mic-ed and recorded in a "perfectly natural" way.

Quite true. IMO far too much attention is paid to the playback system and not enough to the system via which the media was created. As the saying goes, you can't polish a turd.
As for dipoles, or any other esoteric solution to a perceived problem that may or may not exist, media is mixed by engineers listening on playback systems. If you want to hear it as they mixed it you need the same system. And room. That's obviously totally impractical. And that's why the Yamaha NS-10 was the best selling studio monitor ever. Not because it was a good speaker, but because it was a typical speaker. Engineers and producers liked it because they felt that if their mix sounded good through an NS-10 it would sound good through anything.

[Pete Schumacher]

Re: Achieving "perfect" power response

Quite true. IMO far too much attention is paid to the playback system and not enough to the system via which the media was created. As the saying goes, you can't polish a turd.
As for dipoles, or any other esoteric solution to a perceived problem that may or may not exist, media is mixed by engineers listening on playback systems. If you want to hear it as they mixed it you need the same system. And room. That's obviously totally impractical. And that's why the Yamaha NS-10 was the best selling studio monitor ever. Not because it was a good speaker, but because it was a typical speaker. Engineers and producers liked it because they felt that if their mix sounded good through an NS-10 it would sound good through anything.

Well, actually, you can polish a turd to quite a sheen.

However, it still stinks.

[diy speaker guy]

Re: Achieving "perfect" power response

Thanks for all the responses. Deward gives some good, real examples of directivity control but unfortunately I don't think any of them answer my particular question or suit my situation.

One of biggest and most audible problems was caused by crossing over too high for the woofer, in the frequency range where its dispersion was rapidly decreasing. The off axis response would have a sag around the crossover point, for example say 3 to 4 kHz. As the tweeter took over with its greater dispersion, the off axis response would rise to about 5 kHz then fall again after 10k.

The larger the woofer, the lower the crossover point should be for good power response.

This is exactly the "tweeter bloom" I was talking about. I've been playing with a monopole mtm in PCD and off axis, the mids start to droop at 350 hz, fully 7 db down off axis by 1khz, and then at 3 khz the tweeter busts in like a loud drunken mother in law, off axis just as loud as on axis for an octave or so. I'm particularly aware of this because the terrible off axis response of my current single driver fullrange speakers is annoying, and one of the biggest reasons for my shift into mult way design.

I suspect there's little that can be done about that in an mtm. So I'm going to look into a 3 way instead. I haven't done anything with the 3 way in PCD yet, I'm still researching available mid drivers at this point.

I still suspect a shallow xo will help with the "tweeter bloom" but probably not nearly enough, and from this conversation so far it appears that a waveguide might be the only way to flatten the power response right out.

[billfitzmaurice]

Re: Achieving "perfect" power response

Thanks for all the responses. Deward gives some good, real examples of directivity control but unfortunately I don't think any of them answer my particular question or suit my situation.

One way to get it 'perfect' would be to use a driver to cover every octave, each half the diameter of that before it. And once you did that you'd say, well, why not one every 1/2 octave, or 1/3 octave? Or how about a 33 foot high ribbon? (Yes, it has been done). Every speaker is a compromise.

[diy speaker guy]

Re: Achieving "perfect" power response

Sure, I get your point but at the same time Dantheman has posted measurements of at least a couple small commercial 2 way mini monitor type speakers with absolutely lovely power response. How did they do that? IIRC at least one of them had a very shallow waveguide on the tweeter.

Guess I'm going to have to look up Dan's old posts and do a bit of research into those speakers to see what they did.

[DDF]

Re: Achieving "perfect" power response

I still suspect a shallow xo will help with the "tweeter bloom" but probably not nearly enough, and from this conversation so far it appears that a waveguide might be the only way to flatten the power response right out.

Not sure if you've read my posts on this technique in the past, but I've written about it at length and a number of times. It works.

One way to get there and still get good driver resonance suppression (either tweeter fundamental resonance or mid range break up resonance) is to use an overdamped electrical third order. Design the filter so the initial acoustic roll off through the blending region is shallow slope (more like a first order) but then it quickly asymptotes to third order outside the blend region.

It's my favoured topology for a few reasons. One is that it calls for more output from the "non blooming" driver higher up, second is that it calls for less output from the "blooming" driver through xover, third is that its not an "in phase type" so that you are not sacrificing as much off axis for on axis, 4th is that power handling and driver break up are still dealt with.

To me, this is the best compromise, and I've pretty much tried them all, over 30+ years.

Dave