Sealed versus Ported Bass, can the difference be measured?

[Paul McNeil]

Ported and sealed pictures are exactly the same

Yes, sorry, see corrections below.

 

[ryanosaur]

There is little convincing evidence anywhere of group delay audibility in low frequencies. I have always presupposed that the perceived ‘tightness’ people hear from sealed subwoofers is really just from the absence of very low frequencies that, by nature, rumble in a way that doesn’t sound tight.

Agreed, the lower the frequency, the harder it is to hear GD. I think the rule of thumb is it is not audible below 1.5 cycles... Regardless, I think you are absolutely correct in suggesting that very low and infrasonic bass becomes more a feeling than is directly heard as a fundamental frequency. The efficiency with which Ported Subs can reproduce lower frequencies is a major part of the draw compared with what is needed to make a Sealed box produce the same.

From Larson's recent review of the PB17 Ultra:

Group delay is the measurement of how much time it takes for individual frequency bands of an input signal to be produced by the speaker. It can indicate that some frequency components are developing slower than others or are taking longer to decay. It is generally thought that 1.5 sound cycles are needed for group delay to be audible at bass frequencies, although there is an argument that group delay should remain under 20ms to be completely unnoticeable, but that is likely meant for mid and upper bass frequencies.

The PB17-Ultra puts up a very good showing in this measurement and improves upon its predecessor. Group delay doesn’t cross 20ms until around 30Hz below which human hearing is very insensitive. Above 40Hz, it hovers at or below 10ms. That is very good time domain performance. Many people would assume that a big ported sub wouldn’t be great for crisp transients in music, but that wasn’t my experience with the PB17-Ultra, and that is not what our measurements indicate. The different operating modes do not make a big impact on group delay either. Port-generated output does jump up to a cycle of delay, as would be expected. However, it’s far too low in frequency to affect anyone in musical range or likely anything that human hearing would perceive as amiss.

 

[Paul McNeil]

You have made too many concurrent changes to draw any conclusions. It would be good to do this test in a controlled manner.

Yes, I agree, as a (retired) scientist like me would say, I'm comparing Apples to Oranges!

 

[Paul McNeil]

Agreed, the lower the frequency, the harder it is to hear GD. I think the rule of thumb is it is not audible below 1.5 cycles... Regardless, I think you are absolutely correct in suggesting that very low and infrasonic bass becomes more a feeling than is directly heard as a fundamental frequency. The efficiency with which Ported Subs can reproduce lower frequencies is a major part of the draw compared with what is needed to make a Sealed box produce the same.

From Larson's recent review of the PB17 Ultra:

OK, but my sealed arrangement goes lower, for sure. See data above.

 

[Paul McNeil]

Correction for the SPL and PHASE measurements. My apologies, just learning REW...

 

[Blueprint]

OK, but my sealed arrangement goes lower, for sure. See data above.

On a graph, sure. But that’s not how we hear. We need greater SPL at lower frequencies to compensate for our reduced sensitivity. So you don’t hear any of that infrasonic bass. Which is why, in my view, it sounds tighter.

This is why I believe ports and passive radiators are ideal solutions for subwoofers and full range speakers because the port tuning can be manipulated to give a nice SPL boost at the lowest frequency so you can hear it.

 

[ryanosaur]

OK, but my sealed arrangement goes lower, for sure. See data above.

I didn't say that Sealed Subs cannot go lower.

You are also comparing one of the newest-baddest Subs on the market against Ported Subs that are much older and not so well designed.

Apples to oranges.

 

[Oddball]

As you noted, it can be measured. It will depend on the room and your preferences. If the question is what is better - than there is no real answer to it.

I would measure decay before any other parameter suggested above though (other that FQ response). Although they are all interconnected.

 

[Paul McNeil]

I didn't say that Sealed Subs cannot go lower.

You are also comparing one of the newest-baddest Subs on the market against Ported Subs that are much older and not so well designed.

Apples to oranges.

I don't think I did accuse you of saying that. I was just pointing out that I am hear the lower end, with, I agree, a badass sealed sub.

 

[Paul McNeil]

On a graph, sure. But that’s not how we hear. We need greater SPL at lower frequencies to compensate for our reduced sensitivity. So you don’t hear any of that infrasonic bass. Which is why, in my view, it sounds tighter.

This is why I believe ports and passive radiators are ideal solutions for subwoofers and full range speakers because the port tuning can be manipulated to give a nice SPL boost at the lowest frequency so you can hear it.

Hmm, the SPL on a graph does not reflect what we hear? This is the audio science review forum, right?!

 

[ryanosaur]

OK, but my sealed arrangement goes lower, for sure. See data above.

Consider that in a Sealed Box, which will always have a higher tuning above Driver Fs, you achieve lower frequencies by combining power and excursion. To force the driver to play lower than tuning requires much more power and a Driver that cannot only handle that power but also has the excursion to do so gracefully. This is usually where distortion measurements start to skyrocket because most Drivers do not like to be pushed that hard. Even a certain well known manufacturer of Extreme Sub Drivers has claimed that you can't push his Drivers too close to Xmax without them starting to sound like cr@p.
On the other hand, there is another well known designer in Canada whose Drivers are so capable, he encouraged people to model them to within a few mm of Xmech because they could actually handle that and still perform excellently.

Both of these still require gobs of power to perform beyond the normal physics of Driver Fs and Box tune (Fb).

 

[Blueprint]

Hmm, the SPL on a graph does not reflect what we hear? This is the audio science review forum, right?!

I am talking generally. You may have a monster sealed box that can produce low frequencies. Such boxes exist. More power to you. I’m speaking to the idea that sealed subs produce infrasonics better than ported due to their roll off. If you compare two subs, one which has 2-3db boost over the other at 20hz, but fourth order roll off, and one which is 2-3db down at 20hz but has second order roll off. 100% of people comparing such boxes would say the former produces more bass. It’s just how we hear.

 

[ryanosaur]

Hmm, the SPL on a graph does not reflect what we hear? This is the audio science review forum, right?!

Remember Equal Loudness? Our hearing is less efficient at lower frequencies which a simple SPL graph does not account for.

 

[Paul McNeil]

I am talking generally. You may have a monster sealed box that can produce low frequencies. Such boxes exist. More power to you. I’m speaking to the idea that sealed subs produce infrasonics better than ported due to their roll off. If you compare two subs, one which has 2-3db boost over the other at 20hz, but fourth order roll off, and one which is 2-3db down at 20hz but has second order roll off. 100% of people comparing such boxes would say the former produces more bass. It’s just how we hear.

Yeah, I know that ported subs are more efficient in the lower registers. It's simple physics.

 

[TurtlePaul]

Remember Equal Loudness? Our hearing is less efficient at lower frequencies which a simple SPL graph does not account for.

But then you are in circle of confusion territory. If you are playing your subs sufficiently loud, then material which was mastered on a system with flat bass response should be played back on a system with flat bass response.

 

[ryanosaur]

But then you are in circle of confusion territory. If you are playing your subs sufficiently loud, then material which was mastered on a system with flat bass response should be played back on a system with flat bass response.

Right?

 

[ppataki]

Correction for the SPL and PHASE measurements. My apologies, just learning REW...

You can unwrap the phase curves (right click, Unwrap option), use 1/12 smoothing
Then go to Overlay --> Phase and you will then see the two phase curves nicely on top of each other

 

[levimax]

You can unwrap the phase curves (right click, Unwrap option), use 1/12 smoothing
Then go to Overlay --> Phase and you will then see the two phase curves nicely on top of each other

Unfortunately, trying to measure subs in a room is extremely difficult and unreliable beyond basic FR measurements or nearfield measurements (which don't include the port). The combination of long wavelengths, reduced resolution, and the inability use to gated measurements makes any in room time domain measurements suspect at best. I just went through this with a very simple "time alignment" project for some co-located subs and mains. I tried all the standard methods including REW time alignment tool and 2 different "MSO" programs as well as the time staggered impulse response methods and "nearfield and measuring tape" method and all gave different answers and some programs even said the polarity was reversed and some others confirmed it was OK. I had previously used DIRAC DLBC and that gave wildly inaccurate results (over 100 ms off) but I just tried the latest updates and the results appeared reasonable and agreed with my "nearfield and measuring tape" method so went with those delays and based on listening to kick drum loops it appear to be "OK". At this point "maybe" DIRAC is working correctly in some cases by using multiple measurements from know locations and deriving some useful time domain information from these measurements but for the average person using REW and one or even multiple average sweeps, in many/most cases the time domain information is contaminated by multiple reflections and is not reliable. I know @gnarly has experienced this and takes his subs outside to large open spaces to get reliable measurements. My assertion is that much of the confusion and dissatisfaction with subs can be traced to suboptimal integration but since in room measurements are so unreliable I don't know what the answer to better integration is, especially with multiple subs in a MC system.

 

[Keith_W]

Hmm, the SPL on a graph does not reflect what we hear? This is the audio science review forum, right?!

Absolutely the measurement you see does not reflect what you hear. Floyd Toole said so - an omnidirectional microphone does not hear the same as a human with two ears and a brain. This does not mean we throw the baby out with the bath water, what this means is that we have to interpret the measurement with caution.

 

[Paul McNeil]

Unfortunately, trying to measure subs in a room is extremely difficult and unreliable beyond basic FR measurements or nearfield measurements (which don't include the port). The combination of long wavelengths, reduced resolution, and the inability use to gated measurements makes any in room time domain measurements suspect at best. I just went through this with a very simple "time alignment" project for some co-located subs and mains. I tried all the standard methods including REW time alignment tool and 2 different "MSO" programs as well as the time staggered impulse response methods and "nearfield and measuring tape" method and all gave different answers and some programs even said the polarity was reversed and some others confirmed it was OK. I had previously used DIRAC DLBC and that gave wildly inaccurate results (over 100 ms off) but I just tried the latest updates and the results appeared reasonable and agreed with my "nearfield and measuring tape" method so went with those delays and based on listening to kick drum loops it appear to be "OK". At this point "maybe" DIRAC is working correctly in some cases by using multiple measurements from know locations and deriving some useful time domain information from these measurements but for the average person using REW and one or even multiple average sweeps, in many/most cases the time domain information is contaminated by multiple reflections and is not reliable. I know @gnarly has experienced this and takes his subs outside to large open spaces to get reliable measurements. My assertion is that much of the confusion and dissatisfaction with subs can be traced to suboptimal integration but since in room measurements are so unreliable I don't know what the answer to better integration is, especially with multiple subs in a MC system.

Thanks, levimax, for an answer to my question!? Do you have a link for the kick drum loop, I really like that idea. With the two sealed sub configuration, one in each corner, I now get a 3 dB boost across the spectrum, inlcluding the infrasonic. Isn't that an indication of 'time alignment'?