Strange thing - why do my DIY designs always sound dark, and a bit dull?

[fjhuerta]

I hope the title of this thread is a bit descriptive. :s

I've built 5 designs so far - the Natalies, a slim column with Focal woofers and a Neo3 tweeter, a tower with Tang Band (aluminum) woofers and the Neo3 tweeter,a Peerless small bookshelf, and a bookshelf with Tang Band (titanium) woofers and the RS-28 tweeter.

Although I think I'm getting better at the measuring / designing a crossover part, one problem still shows up all the time, everytime - the speakers sound pretty dull.

It doesn't make much sense. My measurements (always on-axis) seem to be OK. When I test the finished speakers, the on-axis response is mostly flat. But the speakers sound dull when compared to my Energy C-3's, or my Monitor Audio Silver RS-8 speakers.

I don't fully understand what is it that I'm doing wrong. The midrange is really great, and the bass response is mostly nice, but the highs seem to be recessed and dull.

Anyway, I tried measuring my Energy speakers, and I was surprised to see a rising frequency response, starting at around 2 Khz and ending around 5-6 dB's up at 20 KHz.

Now, here's what I've been thinking... I suppose this is done in order to deliver a flatter frequency response at 30° or 45° off-axis. This is a mere speculation on my part, but that's the only thing I can think about. I don't know what's the correct way of designing a speaker.... flat on-axis, or flat off-axis?

As it is, I "cloned" the Energy response curve on my latest project (the Tang Band titaniums) and the speakers now similar to the Energies. But I don't know what the right thing to do is... or whether my measurement protocol is wrong.

Any ideas would be very appreciated!

Javier.

[ThomasW]

If the NatP's sound dull, there's a problem with the crossover or the phase of the drivers.

[Dennis H]

Anyway, I tried measuring my Energy speakers, and I was surprised to see a rising frequency response, starting at around 2 Khz and ending around 5-6 dB's up at 20 KHz.

Well, you've answered your own question, haven't you? You're used to "colored" speakers so more accurate speakers just don't sound right to you. Put the Energies away for a while and give your ear (brain) a chance to adjust to the Nats. Then when you go back to the Energies after a month or so, they'll probably sound too bright.

[fjhuerta]

If the NatP's sound dull, there's a problem with the crossover or the phase of the drivers.

Hi Thomas!

I should have been more specific. The NatP's sound awesome. I haven't compared them to my commercial speakers, but I have never thought about them as dull.

I haven't measured them yet (I guess I should), but I assume they are flat on-axis. At least, that's what I've been led to believe by Vance Dickason's book - he doesn't aim for flat off-axis response, but for a flat on-axis one, with a crossover point optimized for off-axis targets.

It's my designs the ones that sound a bit dull. I don't get it!

EDIT: Dennis, you may be right. I have been listening to these speakers for ages - it would be posible that I'm simply not used to flat frequency response anymore.

[fjhuerta]

I suppose it may be a good idea to include the schematic of my latest project (Tang Band titanium / RS-28 MTM)... just in case someone figures out I'm doing something wrong

[cjd]

Is your mic calibrated?

While I suspect the Energies may indeed be hot on the top end (very very common) it is also possible that it's somewhere in between, if your mic isn't calibrated and happens to roll off the top end of your measurements.

C

[fjhuerta]

Is your mic calibrated?

While I suspect the Energies may indeed be hot on the top end (very very common) it is also possible that it's somewhere in between, if your mic isn't calibrated and happens to roll off the top end of your measurements.

C

Hi C,

No, it isn't. It's a Behringer ECM-8000. I also thought it may be uncalibrated, and there's no easy way for me to get a calibrated one, so I bought a second one (as a backup, and to check out calibration). Both mics measure the exact same thing all the time.

[Dennis H]

You can send the mic to Kim Girardin. Last I heard, he charges about $40 to calibrate it.

Kim Girardin
Wadenhome Sound
1400 Homer Rd. Suite 2
Winona, MN, 55987, USA

Phone: 507-454-8844

E-mail: kmgrdn@luminet.net

[ThomasW]

We compared 3 different ECM-8000's against Jon's B&K4133 (same mic that Kim uses) the differences between the 8000's and the 4133 weren't much.

My thought is use any of the online cal files for the 8000 and you'll be fine.

[AJINFLA]

It doesn't make much sense. My measurements (always on-axis) seem to be OK. When I test the finished speakers, the on-axis response is mostly flat. But the speakers sound dull when compared to my Energy C-3's, or my Monitor Audio Silver RS-8 speakers.

You have answered your own question. On axis measurements alone will tell you little if anything about what you ears will hear sitting in your listening chair. What you hear (unlike what a microphone measures at 1m on axis) is the speakers radiated power. Which is why the speakers measured response at (almost) every angle is important (The Harman link I use to post appears to no longer work).
The neo3 and RS28 both have drooping power radiation in the uppermost octaves (as do all drivers, some more than others. Some actually prefer that type of response).
A quick search revealed - Quote from S-pile on the RS-6: ..."Note that the tweeter also has wider-than-usual dispersion in the top octave".

On axis shows both a rise at 10k and a large resonance peak as well, which will definately add some sparkle. The reviewer made no complaints of harshness.


The RS28

http://www.partsexpress.com/pdf/275-130g.pdf

Neo3

If you are looking for tweeters with more uppermost octave energy, you might want to look at some other than the RS28 or Neo3.
Do start measuring off-axis as well.

cheers,

AJ

[Landroval]

If you are looking for tweeters with more uppermost octave energy, you might want to look at some other than the RS28 or Neo3.
Do start measuring off-axis as well.

Yep. Both of those are very smooth and they have biggish radiating area which means the off-axis responce isn't good.

I don't know what's the correct way of designing a speaker.... flat on-axis, or flat off-axis?

The correct way is: What sounds best to your own ear.

In my opinion speaker design needs to take into account both the technical and psychological side. If you don't like flat on-axis then don't do it. It might be your room, your amp, or most likely your ears. Also frequency responce is not the only thing one should look for. I for instance seem to find musical satisfaction from quite different designs. A 3" fullrange might do it for me, while a $70000 4-way might be just ok or plain boring.

[thylantyr]

I've built 5 designs so far - the Natalies, a slim column with Focal woofers and a Neo3 tweeter, a tower with Tang Band (aluminum) woofers and the Neo3 tweeter,a Peerless small bookshelf, and a bookshelf with Tang Band (titanium) woofers and the RS-28 tweeter.

Although I think I'm getting better at the measuring / designing a crossover part, one problem still shows up all the time, everytime - the speakers sound pretty dull.

It doesn't make much sense.

Random thoughts;

1. Many commerical loudspeakers claim technical superiority,
but they all have a different sound.

2. Many DIY'ers design to be technically perfect, but this
is no guarantee that you will like it.

Try a different approach because you already isolated the fact
that those systems you built were dull - perhaps you have reached
the advance stages of audio where you trust your ears.

Convert the design to an active system and let your ears guide
you when adjusting the crossover {slope and frequency}, all you
need is the basic starting point, then skew up/down until you feel it's right.

Next {optional}, install a graphic equalizer to alter response to
see if it sparks interest. If not, remove it.

If none of this works, then the drivers chosen aren't
interesting to you, seek a better design.

[fjhuerta]

AJ, thanks a lot for your detailed explanation. somehow, I suspected the Neo3 was a "soft" sounding tweeter, but that the RS28 may have a harder edge to it. Apparently, this is not the case. One thing is for sure, though - the RS28 is *a lot* flatter than the Neo3, at least according to my own measurements. It's also very interesting to see the Monitor Audio response graph - because it's almost identical to the Energy FR graph I got. It makes me think they are either voiced a bit too hot, or they are taking into consideration off-axis response.

Landroval, I do believe I'd not be completely satisfied if I designed a speaker that measured great but didn't sound that good. I still wonder why a flat frequency response doesn't really satisfy me, but I do understand what you mean with the 3" full-range example. I guess I still have a lot to learn.

Thylantyr, thanks for your comments. I'll try to trust my ears a bit more - so far, trusting measurements hasn't gotten me too far, satisfaction-wise. I think I'll be playing quite a bit with the LSPCad Pro demo, since it lets me model crossovers with multiple channel output without a time limitation. At least this way I'll be able to listen before I build a crossover.

[AJINFLA]

Fj,
not sure why it wasn't working before http://www.harman.com/wp/pdf/Loudspeakers&RoomsPt2.pdf

cheers,

AJ

[knifeinthesink]

Pardon me If Im threadjacking, but would someone point in the direction of one of those ecm8000 calibration files. I searched but I dont want the nagging doubt that Im using a file created by someone with no clue.

[cjd]

One other thing I just though about - flat response can be achieved many ways, and not all are equal. Poor phase relationship, even negative summing is possible, with a net result of a flat on-axis measurement, yet bad sound.

I'm not sure this is the case, but it's just one of those other things to ponder as you work.

C

[JJones]

Well, you've answered your own question, haven't you? You're used to "colored" speakers so more accurate speakers just don't sound right to you. Put the Energies away for a while and give your ear (brain) a chance to adjust to the Nats. Then when you go back to the Energies after a month or so, they'll probably sound too bright.

When the output is not omni-directional, then, a flat response measured "on axis" by definition will be inaccurate in terms of overall "power response".

So, the matter of which speaker is "colored" is subjective, to say the least.

Though I do generally agree with your post, for the most part.

[fjhuerta]

Just to make matters clear...

The Energy speakers are, indeed, very flat. I just re-measured them, and now my measurements agree with those from SoundStage! (http://www.soundstage.com/revequip/energy_c3.htm).

It's funny to see I had already found out about my problem two years ago. I can't believe I didn't trust my gut feeling back then - I should have investigated further!

[mikela]

I just re-measured them, and now my measurements agree with those from SoundStage!

What changed?

[ThomasW]

His microphone now has a more accurate correction file. See page 3 of this thread

http://www.htguide.com/forum/showthread.php4?t=23919

[mikela]

Aah...now it makes sense.

[Dave Bullet]

With the mic calibration issues out of the way.... that means the energy's and the Nats are flat - therefore what makes the Energy's brighter?

My guess is more distortion. Accurate speakers (ie. those that most closely reproduce the input signal without distortion) will sound dull compared to those that add harmonics.

David.

[fjhuerta]

Not really. My Nat-P build has a strange behavior - it has a serious bump at 2 KHz. It seems to be baffle-related.

My Nat-P's were modified by me, taking into consideration the measurements I took with the non-calibrated microphone. This is what made the Nats sound duller than the Energies.

The 2KHz issue doesn't allow me to build the "real" Nat-P, so I modified the crossover yet again to fix out this issue.

Now, the Energies and Nat-Ps sound similar, although the Nats have a far better midrange sound...

[Undefinition]

I think this is an interesting thread. it takes a lot of guts to admit this, Javier.

I have to ask, though: what would you like your DIY designs to sound like? What would your "dream speakers" sound like? I mean, it's called DIY, and part of the fun is that you get to "build your own sound." Maybe looking away from the mic'ed Frequency Response, and using your ears might help. I'm sure they did it with the commercial designs (though I'm sure they own plenty of nice measurement equipment as well).