Tools for building better speakers

[a4eaudio]

CLIO is fast and easy to use. ...support for motorized table,

The motorized table support is attractive as well.

I'm not arguing for REW, but as useful information....

I believe REW added functionality to send a signal to motorized turntables in 2024.

Another cool thing (maybe other software does this too) is the ability to use multiple microphones simultaneously (with the Pro version that has a small fee). (Although the naming convention for VituixCAD doesn't work so you have to manually rename.) I'm working on a setup with 4 mics, at zero, 50, 100, and 150 degrees. Zero to 180 in 10 degree increments can be done in 5, rather than 19, sweeps. If I get it working I'll post the setup on ASR.

 

[a4eaudio]

for ARTA, could see that it had a lot of good science behind it. I did not find it very intuitive though and has yet another dated UI...

Yep. I recognized that ARTA had a few technical advantages, but I switched to REW for the UI and just general user-friendliness.

 

[TimVG]

Never know when the next Microsoft update breaks stuff and if nobody is around to fix,

I bought a used €25 laptop that runs offline for this eventuality

It would be convenient if a similar workflow could be achieved one day in REW.

 

[RickS]

I bought a used €25 laptop that runs offline for this eventuality

It would be convenient if a similar workflow could be achieved one day in REW.

Only got as far as seeing whether ARTA would work with my hardware. What does REW need?

 

[TimVG]

Only got as far as seeing whether ARTA would work with my hardware. What does REW need?

Simply put: ARTA can auto-save every measurement as a seperate file (live), appropriately named. So right after your measurement set, you can then simply select your folder in Vituixcad and load them in through the IR conversion tool. Further adjustments are done solely in vcad (gate, mic call, and SPL compensation if needed).

 

[umr]

In the one video I watched, he had claimed that the woofer, cabinet and port do their thing there so why bother to measure.

Amirm,

I'd love to see your take on ported vs. sealed speaker designs regarding transient response in the bass region.

 

[antcollinet]

Amirm,

I'd love to see your take on ported vs. sealed speaker designs regarding transient response in the bass region.

Not sure what you mean by "transient response in the bass region". A transient - by defintion is a full bandwidth signal. To reproduce it, you need the full frequency range of the speaker.

 

[Keith_W]

It has always been my belief that ported speakers have inherently more group delay, until it was pointed out to me that the extra group delay is simply the minimum-phase nature of the subwoofer. Extend the bass response = get more group delay.

I had a hard time believing it. So I simulated a ported and sealed subwoofer in WinISD then tried to design a Linkwitz transform to get the sealed sub to match the ported sub. It is very difficult to design a Linkwitz transform since it's a trial-and-error process, but despite my inability to make the sealed sub exactly match the ported sub, the trend was clear. The extra group delay IS simply due to more bass extension, and that's all there is to it.

That insight alone made me have to unload a whole bunch of other erroneous beliefs. So: the impulse/step response of a ported and sealed speaker would be the same, IF the frequency response is made to be the same. This may involve applying a Linkwitz transform to the sealed speaker. I too, would like to hear what Amir thinks.

 

[dfuller]

The extra group delay IS simply due to more bass extension, and that's all there is to it.

Yes and no; there's extra group delay in a ported cabinet just by virtue of ports being fourth order systems vs sealed being second order (yes, Q plays a role here of course, but that's part of frequency response).

Of course it's all minimum phase down there so if you match linear extension and low end slope it'll produce the same group delay result.

I actually just looked this up for the sake of it.
This the group delay for the KH310 - a Linkwitz Transformed sealed box. Its low end roll off is certainly steeper than a sealed box normally would be (there's a 6dB/oct electronic filter with F3 of 15hz).

Compare this to the KH150, which is ported and has an extremely steep roll-off (12dB/oct sub filter with F3 of 38hz as well as the port's 24dB-ish/oct inherent rolloff)... Note that the group delay is higher for the ported speaker even in regions they're both linear.

Obviously we have to take into account the earlier roll-off for the KH150, but even if we do that with the KH420 (which extends lower than the 310) it still bears out somewhat, with slightly higher GD at 40hz than the 310.

Is any of this audible? In the vast majority of spaces, almost certainly not. Even in the best of them in terms of low end response, the answer is still probably no; none of these show anything even remotely resembling excess GD.

 

[voodooless]

Extend the bass response = get more group delay.

Well, usually it’s the steeper slope that gives you more group delay, but not exclusively.

But let’s not forget that a reflex system is not only minimum phase. But with good design the additional ringing can be minimal.

And please people, don’t forget the room! Group delay is equally influenced by room effects. Fix the in-room response, and you fix most group delay issues.

 

[Kevinfc]

this is sadly standard fare in American economics the last few decades.... "the free market" goes unchecked in predatory practices ..I see this in home improvement ads, things like large air conditioning companies declaring that ac units "life expectancy" is 10 -12 years .. the truth is all the parts can be replaced as long as they exist .. if you can diy you could make a unit last decades and work well ... but that cuts into profits ... so units have expected lifespans and very inflated service prices (labor) to encourage replacement... Danny follows the same business model , preying on the audio illiterate ...

Life Expectancy can also mean the end of a unit’s technological efficiency. Air conditioners have become enormously more efficient over the last 25 years, some using as little as a third of the energy of previous units. They are also considerably quieter. Your computer would be a useful comparison. Yes you could continue to use a 25 year old PC, repairing it with vintage parts, but doing so would still leave you with something that is less useful and efficient.

 

[Davey2156]

I had a hard time believing it. So I simulated a ported and sealed subwoofer in WinISD then tried to design a Linkwitz transform to get the sealed sub to match the ported sub. It is very difficult to design a Linkwitz transform since it's a trial-and-error process, but despite my inability to make the sealed sub exactly match the ported sub, the trend was clear. The extra group delay IS simply due to more bass extension, and that's all there is to it.

A Linkwitz Transform isn't designed to convert a second-order roll-off into a fourth-order roll-off.
A ported box utilizes stored energy to extend the response lower, but at the cost of increased roll-off rate......relative to a closed box.
It's relatively easy to convert a ported box to a sealed box.....but not the other way round.

 

[kimmosto]

a reflex system is not only minimum phase.

Simplest simulators say it's minimum-phase but there is transmission delay in the port signal. I've never seen measured vented to be minimum-phase, but poor SNR at very LF is part of the puzzle causing error to calculated excess phase. So sealed is minimum-phase with much higher confidence.

 

[CDMC]

I know a handful of small shops being in the HiFi audio circle over the years, many of them are started by DIYers who got very serious and are close to retirement from their day job and decided to take that leap and started their own business. In fact, I met one couple months ago in my local audio society club in our last meetup.

Some of these small shops would do ground plane or quasi-anechoic, until they get their speakers near the final stages of design, then they would send it into an anechoic chamber for measurement to validate; some don't even bother, because they believe their ground plane or quasi-anechoic is sufficient. The ones who do, they would send it in maybe twice or 3 times depending on the design cycle. From what they tell me, the cost ranges from $800 to $2000 per time (price may have gone up since).



Of course the decision is up to the manufacturer to spend that money or not. In the last 7 years or so, 70% of my speaker purchasing decision is based on the measurement results, if it doesn't cut mustard, it's an automatic no. If no measurement data is available, then it's also an automatic no.

In 2026, if you designed and sell speakers and I can't find the measurements either through you or a third party, I, for one, will not even give your products the time of the day. . .full stop, end of story, case close. And I am pretty darn sure I am not the only one.

Ever since spin data has been widely available, I have never encountered, a pair of speakers that measures good yet sounded bad, not one. . .unless there are other measurements that wasn't publish such as distortion level.



It was between the Philharmonic HT Tower and the Ascend ELX RAAL. I have heard the Philharmonic at CAF one year, I think that was one of few years where Dennis did a show, but one can never judge accurately from shows alone, for reasons outside the scope of this post.



Yes, I do know, though not always clear and reliable. But anyway, I reference Ascend showing that cabinet resonance with a vibrometer to highlight the number of tools and data that was published by Ascend on their products, which ultimately pushed me to buy the Ascend over Philharmonics.

I will just add to this, that in the old days, dealers would have stock on hand, so you could bring potential purchases home to hear them in your room. Those days are gone. Buying a speaker with a good response and good spin is far easier and gets better results. You start with a speaker that is going to have a better in room response, then apply adjustments to correct for your room deficiencies and preference. It is so much easier, less expensive, and at least in my experience nets a better listening experience.

 

[Paul7777x]

I will just add to this, that in the old days, dealers would have stock on hand, so you could bring potential purchases home to hear them in your room. Those days are gone. Buying a speaker with a good response and good spin is far easier and gets better results. You start with a speaker that is going to have a better in room response, then apply adjustments to correct for your room deficiencies and preference. It is so much easier, less expensive, and at least in my experience nets a better listening experience.

Exactly this.

A lot of erm… the hard of thinking on so many other sites, fail to realise (or don’t want to) that a ‘properly’ measuring speaker is a godsend because it is entirely predictable in that there will be no random, hard to understand nonsense noise coming from it in room.

And therefore correcting for the room is a breeze.

 

[umr]

Not sure what you mean by "transient response in the bass region". A transient - by defintion is a full bandwidth signal. To reproduce it, you need the full frequency range of the speaker.

I am talking about the leading response for something like a bass drum or bass guitar. A lot of the analysis for speakers is made for steady-state response of a speaker which does not represent many actual sounds that we attempt to reproduce.

 

[antcollinet]

I am talking about the leading response for something like a bass drum or bass guitar.

Download a sample from a bass drum, or bass guitar (actual sample, not synthesized). Load it into audacity. Zoom into the first couple of milliseconds of the "leading response". You will find it is nothing like an impulse - it will be a number of cycles of some midrange frequency - some higher frequnecies on top, gently ramping up to the peak of the "strike"

It has nothing to do with the impulse response of a speaker "in the bass region" or any other region. Just like every other real world object, speakers, and instruments - even percussive instruments - are band limited.

A lot of the analysis for speakers is made for steady-state response of a speaker which does not represent many actual sounds that we attempt to reproduce.

Test signals are not steady state. Some of them go from full negateive to full positive 20,000 times per second. You don't get more dynamic than that.

Also : Superposition Principle.

 

[illusonic]

The most perfect speakers I know, the Neumann monitors, are designed using anechoic chamber and DSP.

When I was at the Klein&Hummel facility in Ostfildern, where for example the O500 was designed, they had a small room with wedges on walls, ceiling and most of the floor. It was definitely not an anechoic chamber. And I'm fairly certain that the majority of the acoustic design work, especially manual waveguide modeling, was done on site. Maybe with an external chamber used for final validation and tuning of DSP filter settings. Don't know if they have an anechoic chamber today.
Anyways, especially an anechoic chamber and depending on the task also an NFS may help to speed things up. But they are no guarantee for good products. If an acoustic engineer is not able to design a good speaker without such tools, they will not magically enable him to do so. I think the majority of speaker companies has neither anechoic chamber nor an NFS. But that is usually not the reason for bad products.

 

[amirm]

Anyways, especially an anechoic chamber and depending on the task also an NFS may help to speed things up. But they are no guarantee for good products. If an acoustic engineer is not able to design a good speaker without such tools, they will not magically enable him to do so. I think the majority of speaker companies has neither anechoic chamber nor an NFS. But that is usually not the reason for bad products.

I don't know why this point keeps getting made. Of course NFS, or anechoic chamber don't design speakers by themselves. And having one as tools, doesn't assure a good design.

I as a reviewer could have done my job without these as well. But with massive amount of work and possibility of inaccuracy. Heck, I even considered building a $100K anechoic chamber! But decided to get the NFS. Using it sharply increased my knowledge level and ability to analyze performance of speakers. And create a highly defensible objective barrier to people dismissing my work. There is no reason to argue about this and say that I could have been doing reviews without one.

 

[amirm]

When I was at the Klein&Hummel facility in Ostfildern, where for example the O500 was designed, they had a small room with wedges on walls, ceiling and most of the floor. It was definitely not an anechoic chamber. And I'm fairly certain that the majority of the acoustic design work, especially manual waveguide modeling, was done on site. Maybe with an external chamber used for final validation and tuning of DSP filter settings. Don't know if they have an anechoic chamber today.

None of the products I have reviewed from them are from that era. In every case, I am given anechoic measurements which I have replicated very closely with Klippel NFS. What the company did back then is neither here, nor there. Unless you can show no progress in using an anechoic chamber.