Sealed speakers VS ported

[voodooless]

no whining, I'm laughing. you put on voodoo as soon as you can.
Physics and mathematics? have you or haven't you?

Why don’t you just be constructive and get to the point? If I’m wrong, tell me why. It’s not hard, I can take it…

 

[YSDR]

The first 10ms look exactly the same.

In theory, they cannot be the same. In my case, the sealed one is even slightly worse, but theoretically the ported one should be worse (as we can see in the green circle):

 

[voodooless]

In theory, they cannot be the same. In my case, the sealed one is even worse, but theoretically the ported one should be worse (the green circle):View attachment 237413

There is not really a better or worse. All of this is just a function of frequency response. The steeper falloff of the reflex will probably give the higher swing and larger decay time.

What really matters however is how this does in a real room at the listening position. All the fancy theory can be thrown out of the window once the room starts interacting. Only treatment, special configurations and EQ can save (some of) this.

 

[YSDR]

There is not really a better or worse. All of this is just a function of frequency response. The steeper falloff of the reflex will probably give the higher swing and larger decay time.

Come on man, that's what theory says but in my case the sealed box had higher swing!!! Of course the ported had the longer decay time, because the steeper roll-off.

 

[voodooless]

Come on man, that's what theory says but in my case the sealed box had higher swing!!! Of course the ported had the longer decay time, because the steeper roll-off.

Do you have frequency response of both? The answer should be there. Or what driver, box size? Let’s see what the sims do?

 

[fpitas]

In theory, they cannot be the same. In my case, the sealed one is even slightly worse, but theoretically the ported one should be worse (as we can see in the green circle):
View attachment 237414

Since it undershoots, I have to think the sealed box has a pretty high Q. Some stuffing would probably help.

 

[YSDR]

Do you have frequency response of both? The answer should be there. Or what driver, box size? Let’s see what the sims do?

Sims just use simple maths, they cannot add sealed air pressure change non-linearity to the equation, at least that loudspeaker box sim doesn't exist that I know. But okay, the driver was an Eminence Kappalite 3012LF in net 75 liter box, port tuned to 34Hz. The box size and driver was the same (one box) just once sealed and once ported. Unfortunately I don't have the measured frequency responses, just saved those impulse responses.

 

[voodooless]

Sims just use simple maths, they cannot add sealed air pressure change non-linearity to the equation, at least that loudspeaker box sim doesn't exist that I know. But okay, the driver was an Eminence Kappalite 3012LF in net 75 liter box, port tuned to 34Hz. The box size and driver was the same (one box) just once sealed and once ported.

Thanks, will do some sims tomorrow. Let’s see how close they are to what you measured.

Unfortunately I don't have the measured frequency responses, just saved those impulse responses.

Do you only have the pictures?

 

[YSDR]

Since it undershoots, I have to think the sealed box has a pretty high Q. Some stuffing would probably help.

The sealed Q was 0.5 (according to the T/S spec of the driver), that's pretty low I think.

 

[YSDR]

Thanks, will do some sims tomorrow. Let’s see how close they are to what you measured.

Do you only have the pictures?

Only have those two impulse responses that I posted. Here is the sim for the sealed vs ported:

 

[pseudoid]

TL&DR version: #1)Thiele/Small parameters and #2)Quality Factor.
IMHO: These are the first 2 initial speaker design factors which (should) determine most all other subsequent calculations (dimensions, specs, ports, sound, etc.) by a passive speaker builder.
Should I put a YMMV disclaimer here; so that I am not tar'd and feather'd????

 

[fpitas]

The sealed Q was 0.5 (according to the T/S spec of the driver), that's pretty low I think.

Yes, it is. Odd then it should have that response. A Q = 0.5 2nd order highpass filter has a very damped response.

Impulse response, nevermind. I should learn to read lol.

 

[voodooless]

Only have those two impulse responses that I posted. Here is the sim for the sealed vs ported:
View attachment 237438View attachment 237439
View attachment 237436View attachment 237437

This is what Hornresp gives:

Closed:

Ported:

Looks not very similar to yours, these are much faster. Neither is nowhere near similar enough to your measurements though.

 

[YSDR]

Looks not very similar to yours, these are much faster. Neither is nowhere near similar enough to your measurements though.

Maybe that's where reality (a realworld loudspeaker) differs from simple math theory (the simulation) and maybe the main reason is the closed airspring non-linearity to pressure changes. Of course my measurements did not take place in a perfect anechoic environment, so they are only comparable to each other and not to other measurements or simulations.

 

[Frgirard]

Maybe that's where reality (a realworld loudspeaker) differs from simple math theory (the simulation) and maybe the main reason is the closed airspring non-linearity to pressure changes. Of course my measurements did not take place in a perfect anechoic environment, so they are only comparable to each other and not to other measurements or simulations.

the simulation is not the mathematics.
simulation is a plus when you don't know how to define the laws and mathematical calculations that characterize the operation.

Drivers and un speakers are a physical system: mechanics and electricity. these are two fields of physics that are deterministic and therefore predictable. mathematics is the foundation of physics.
A driver in the real world can only work according to the physics that describes it. where is the physics that describes and predicts how a driver works?

 

[YSDR]

A driver in the real world can only work according to the physics that describes it.

The other way around. The driver in the box works as it does and there are some quantifications to describe (or predict) how it works, but if some aspects are missing, the simulation will not be accurate. Does the Hornresp simulation (or other simulation) include the closed air spring nonlinearity against pressure changes?

 

[voodooless]

A driver in the real world can only work according to the physics that describes it. where is the physics that describes and predicts how a driver works?

All of these are just models to predict real-world implications. Some models are simpler than others, and can therefore be used in specific use cases. They can predict one thing well but fail at certain others. Predicting "all" real-world implications will make for a very complex model. Usually, people don't bother.

 

[risandipra]

Thanks, I hear ya but I only have my ears in a room to judge.

It tried this to fix my room modes:

Acoustic calculator: Room modes

And this for the reverberation time : https://www.omnicalculator.com/physics/reverberation-time

To be honest; I did not measure it pre and post-treatment, just by ear; the room modes got better. (I applied them into JRiver Media Center Parametric EQ features).


FYI: I played unamplified classical guitar (and bass-guitar ocassionally). And yes, I don't like the boomy sound as well.

Hope this helps a bit.
Cheers.

 

[fineMen]

This is what Hornresp gives:

Closed:
View attachment 237560
Ported:
View attachment 237561

Looks not very similar to yours, these are much faster. Neither is nowhere near similar enough to your measurements though.

Please acknowledge that the amplitude ./. time behaviour of a speaker doesn't tell much to the naked eye. It is way better to describe the speaker's time behavior with concepts like 'group delay' and such. The above given graphics don't tell anything significant.

Maybe that's where reality (a realworld loudspeaker) differs from simple math theory (the simulation) and maybe the main reason is the closed airspring non-linearity to pressure changes. Of course my measurements did not take place in a perfect anechoic environment, so they are only comparable to each other and not to other measurements or simulations.

No need to take physics into scrutiny. Please acknowledge that the Thiele/Small model is a linear one. Only lately Klippel (patented) invented a machine to investigate the driver's actual operation beyond, into the non-linear regime. There are significant deviations from the elder Thiele/Small model to be observed. They also affect the difference between ported versus sealed.
Let alone the design of an actual port which most easily suffers from 'chuffing', further non-linearities and so forth, and agan again with drones aka passive radiators. It becomes a challenge actually to combine the urge for small(er) speaker with that for (cheap) deep bass via a port.

It s all in the current literature, and can be found in praxis, when eventually evaluating the achievements critically by honest measurement.

ps: regarding Klippel, this one, not the near filed scanner used by Amir: https://www.klippel.de/products/rd-system/modules/scn-scanning-vibrometer-system.html

 

[ppataki]

So, here is goes. As an example the SB15CAC, closed and BR. Here the frequency response of both (green BR, blue closed):
View attachment 236624
And the group delay:
View attachment 236625

Now lets EQ the closed one very close to the BR version:
View attachment 236626
Close enough...

Now what does the group delay look like:
View attachment 236627
Well, well, closed does "worse", even though they look the same. Why? because the closed version has a tiny bit steeper downward trend. If you were to EQ that exactly the same, they would be the same as well.

But I would not bother. The room has so much influence on the response, that these tiny differences are irrelevant anyway. So the main goal should be to create a flat in-room response, regardless of the type of alignment you have. This will always yield the best group delay.

@voodooless
Would it be possible to run the same simulation but with linear phase filters?
Thank you