LoudspeakerLab 10-Minute Speaker Design Tutorial

[technodanvan]

Happy to help! Couple more thoughts since I have a little caffeine flowing now, but I'll be playing with it more today.

Something else that might be nice is the ability to 'lock' individual dimensions as, sometimes, the software seems to change those itself to maintain volume or critical dimensions. It may be better to just provide a warning to the user that something is not going to work. In the previous screenshot, if I choose a 5" baffle for a woofer with a 5.25" radius I think the software should let me do that - it should just pop up with a warning saying how that is not going to work.

Right now it kind of seems like the minimum width (in this case) might not let me go below the driver width + wall thickness? If this is the case I don't think this is correct since driver flanges can be quite large sometimes and overlap walls on narrow baffles.

Showing dimensions in BOTH imperial AND metric simultaneously might actually be nice too, at least where it would fit.

Point 0,0 seems to be dead center on the baffle. Might be nice to be able to change that to bottom left of the baffle, or top left, or even something arbitrary?

What does "dims in in" mean and/or do below the diagram? Might be a place to change dimension units?

Baffle shape seems to be limited to rectangular currently. Would be nice to model trapezoids (such as an 'A' shape) and circular baffles. Perhaps others as well (think the Linkwitz Lab LX521!)

Would also be nice to have multiple baffles. Several recent designs use the mid/tweeter plus a bass bin - and the bass bin could be wider than the mid cab to accommodate larger drivers. That probably wouldn't be an 'ideal' arrangement but some might be interested in that.

Port length seems to be limited to that which fits in the cabinet. I think there should be a check box to allow longer ports. Several designers have integrated a longer external port in speaker stands for bookshelf speakers.

Ports can be down-firing (which is cool!) but in this configuration I think it would be beneficial to show what height the speaker would need to be raised off a floor to ensure that tuning does not change. Alternatively, or perhaps additionally, show how you can use the floor (or even a rear wall) to adjust the tuning of your speaker by virtually 'extending' the port with a solid barrier.

Edit: Does the software include a bass 'boost' for adjacent barriers? I haven't used this myself, but my understanding is one will sometimes place a woofer closer to the floor (assuming CTC distance is not compromised - a 2.5-way would work well for this) to boost the bass response.

 

[technodanvan]

Do calculations take place locally (as in, on the computer running the browser) or remotely (on the hose server for the website)? A combination of both?

I think it might be best to have a block that is simply called 'notes' and have any informative statements and/or warnings appear there. This section could be minimized or expanded depending on the size of the browser window and user preference.

I really think calculations should not happen until a button is pushed. The constant loading and updating - especially if using the up/down arrows for things like box tuning, length of port, etc - is a bit tedious. By using a button we could set (and perhaps lock) baffle dimensions and desired tuning and then hit 'go' to determine the other values. Perhaps having toggle switch between 'real time' calculations and 'user initiated' calculations would be helpful, especially for those with slower internet?

Just spit balling here, it may be valuable to not show the enclosure right off the bat. Let me explain: if you omitted the visual box at first you could select the bass frequency response you want by adjusting sliders for volume (sealed and vented) and Fb (vented) showing how the freq. response changes in real-time. Check boxes or drop downs could be used for stuffing, box leakage, port end correction, etc. Perhaps being able to temporarily save the box types could be handy as well so one could directly compare a sealed enclosure vs. vented at Fb1 vs. vented at Fb2, etc. After the box is 'tuned' to what the user wants THEN we could go to the enclosure creation screen to visualize the box, change the baffle size, etc.

You could also have predefined tunings for vented and sealed boxes, I don't see where Qtc is mentioned, for instance, or common alignment types for vented enclosures. I understand wanting to make this approachable for new users but unless people see these options (and unless good tooltips are available for said options) I'm not sure how much learning will occur.

Passive radiators: I assume this will be added at some point and it could get complicated fast. I do think that it would be VERY valuable to be able to select woofer and tweeter and begin your box modeling WITHOUT a predefined passive radiator. Rather, one could determine the desired tuning and volume, perhaps by using the sliders, and the tool could suggest appropriate passive radiators given those criteria once you go to the enclosure creation, maybe within, say 10 Hz and 20% volume range? Passive radiators like those from Dayton have information available for each weight added, if that could be loaded in the tool that would be handy. If it could also derive values from other weights that could be useful too.

Just suggestions of course. Not everyone would benefit from these and I understand priorities might lie elsewhere. I'm becoming something of a box design nerd so this is kind of my focus right now.

 

[wigginjs]

@technodanvan This is great. I made some changes and implemented them to (hopefully) fix the issues you were having with the Enclosure Designer automatically changing or reverting dimensions. Now, when you take a dimension off of 'auto', it should stay exactly what you set it to. Could you try it out and let me know what you think?

Yes, you are right about limitations with frame size of a driver versus cutouts affecting the minimum baffle dimensions. About half of the drivers in the database have a cutout metric. If it does, the enclosure designer will let you narrow the baffle to the cutout size. If not, the designer limits you to the frame diameter to prevent you designing an enclosure that the driver can't actually fit into.

 

[technodanvan]

I jumped in there this evening, it feels a lot better!

Edit:

Discovered another instance of mixed units when I was playing with larger curves in an attempt to simulate a baffle akin to the Audio First Fidelia. The warning is in metric whereas the set radius is in imperial.

Edit 2:

I think we should see vertical lines in the graph for every 10 Hz between 20-100 Hz. F6/F10 would also be interesting. Being able to manually set the range on the X/Y axis would be nice, especially if one were to have very large subwoofers in the mix. Scanning along the graph seems to 'jump' the cursor by a few Hz as you move horizontally across the graph, it'd be nice if that were smoother.

Edit 3:

Something I think would be really neat is a way to automatically incorporate the golden ratio in box dimensions.

 

[technodanvan]

I was playing with the front slot and noticed that once the dimensions of the baffle were sufficiently large to activate the slot option, it basically let me make the slot as large as I wanted. Personally I think this is fine, but a warning might be prudent to state it would not fit as placed. See below.

Also note the word 'woofer' obscures the area between the tweeter and midwoofer in this instance. Likewise, 'tweeter' obscures the edge of the baffle. It'd be nice if it didn't do that so I could be sure about appropriate spacing. I also think labeling them as tweeter and woofer is redundant since the model numbers are already present to identify them.

Also regarding slot ports: While not seen very often, one could have multiple slots, slots along the vertical edges, a rear firing slot, etc. Just something to think about, might be beyond your intent for the tool.

It would be informative if we had an option to see the following:

• Approximate distance from the woofer/mid/tweeter edge to the edges of the baffle. As an example, I sometimes like to place a woofer equidistant from the vertical edges and bottom edge of the baffle for aesthetics.
• Center-to-center distance between drivers and the quarter wavelength frequency that corresponds to. This could be carried over into the crossover design.
• Clicking on either image of the cabinet should 'maximize' the image so it's easier to see. Perhaps that could show some of the other dimensions discussed then?

 

[technodanvan]

I'm a little concerned the response graph is a bit on the optimistic side? I don't know what all goes into it but here's a comparison for the FaitalPRO 5FE125. I confirmed the T/S parameters were the same between LouspeakerLab and VituixCAD. VituixCAD shows a shelved response (which I think is correct for this driver).

 

[Azrael]

Nor can you seem to tell it you have chamfered the interior of the baffle on the cone woofer(s) / mid(s) or if you countersunk the drivers flush with the baffle. Not sure if there is a calculation that would determine the effect of these decisions.

In VituixCAD (and I assume it’s no different in LoudspeakerLab), the software apparently assumes that drivers are milled flush with the surface, resulting in a front baffle surface that is as free of irregularities as possible.

Behind the drivers, it apparently assumes optimal flow conditions.

Deviations from these assumptions—which can indeed occur in practice—significantly increase the complexity of the simulation.

However, if you properly mill all the drivers flush in practice and, if necessary, mill the baffle behind the driver to optimize airflow, you won’t have to deal with these simulation issues.

 

[wigginjs]

I'm a little concerned the response graph is a bit on the optimistic side? I don't know what all goes into it but here's a comparison for the FaitalPRO 5FE125. I confirmed the T/S parameters were the same between LouspeakerLab and VituixCAD. VituixCAD shows a shelved response (which I think is correct for this driver).

View attachment 545434



View attachment 545433

I think probably the difference is coming from your Qa and Ql. In LoudspeakerLab, these are hard-coded to Qa 100 and Ql 7. I'm not sure if you chose Qa 5 and Ql 15, but those will affect the curve probably in the way you've shown. These variables have to do with estimated box air leakage and absorption, so they vary based on lots of factors. LoudspeakerLabs are actually more conservative than what you've picked in VituixCAD, estimating a little more leakage and significantly less absorption. You could only achieve a Qa number like 5 in a vented box if you heavily stuffed it with polyfill, which isn't typically what folks do when they build a vented enclosure. Not saying it's wrong, just not common.

 

[ernperkins]

For a little more explanation, here are descriptions from VCAD (in typical terse format) and Jeff Bagby:

Qa:

VituixCad: No fill: Qa = 120, Minimal fill: Qa = 80, Walls covered: Qa = 20, Heavy fill: Qa = 5

Jeff Bagby, Passive Crossover Designer (PCD):

Qa is the Q at due to box losses from absorption. As Qa approaches infinity losses will approach 0.
High values >100 will represent very low losses. Normal values will range from 120 to about 5. As a guideline, an enclosure with no fill at all would have a Qa value of 120, a small amount of dacron in the box would bring it to about 100-80, walls covered with foam would be about 40, covered with thick fiberglass or a heavy damping pad would bring it to about 20. A fully stuffed box would be 5-10 depending on stuffing density.

Ql:

VituixCad: No leaks: Ql = 30, Minimal leaks: Ql = 15, Minor Leaks: Ql = 10

Jeff Bagby, PCD:

Ql is the Q at Fb due to box losses from leakage. As Ql approaches infinity losses will approach 0. A system with no leakage will have a Ql of 50. Very small leaks will have a value of 30, small minor leaks will be about 20, and significant leaks will have values of 5-10. Normal values will range from 40 to 10, with 20-30 being a fairly typical system. Values in the range of 10 will represent many porous cabinet and cones, 5 will represent leakage like you would see when using an Aperiodic vent. Use these numbers as a guideline for adjusting these values.

 

[technodanvan]

I think probably the difference is coming from your Qa and Ql. In LoudspeakerLab, these are hard-coded to Qa 100 and Ql 7. I'm not sure if you chose Qa 5 and Ql 15, but those will affect the curve probably in the way you've shown. These variables have to do with estimated box air leakage and absorption, so they vary based on lots of factors. LoudspeakerLabs are actually more conservative than what you've picked in VituixCAD, estimating a little more leakage and significantly less absorption. You could only achieve a Qa number like 5 in a vented box if you heavily stuffed it with polyfill, which isn't typically what folks do when they build a vented enclosure. Not saying it's wrong, just not common.

Interesting, do you mind explaining why these are hard coded at these values? Or is that explained somewhere? For example, and as @ernperkins mentions, VituixCAD suggests a Qa of 20 when walls are covered with foam, which would be pretty common for a vented enclosure.

 

[wigginjs]

Interesting, do you mind explaining why these are hard coded at these values? Or is that explained somewhere? For example, and as @ernperkins mentions, VituixCAD suggests a Qa of 20 when walls are covered with foam, which would be pretty common for a vented enclosure.

I chose what I thought were common values for the way I've seen folks typically build DIY speakers. This is a tradeoff between tool flexibility and UX simplicity. Tools like VituixCAD go for maximum flexibility, but IMO, at the cost of UX and approachability. If you model out what the response of an enclosure looks like with Qa = 100 vs. Qa = 20, all else being equal, the difference in the response is typically quite small. Actually, lower than the expected error from things like actual box volume, port diameter, port length, stuff fill when you consider that folks don't typically exactly measure driver volume displacement, etc.

In short, my philosophy when designing LoudspeakerLab is to only expose what's REALLY necessary to the user for kind of the 80% of DIY speaker design with the hope of making the tool more approachable, easier, and faster to use.

Curious your feedback about that.

 

[ernperkins]

I like how WinISd handles Qa, Ql and Qp. Default values are used unless you select the "Advanced" entry. From there you can change the values.

 

[technodanvan]

In short, my philosophy when designing LoudspeakerLab is to only expose what's REALLY necessary to the user for kind of the 80% of DIY speaker design with the hope of making the tool more approachable, easier, and faster to use.

Curious your feedback about that.

I guess this seems limiting to me since you'll be training people on the basics but not exposing them to the details that push the boundaries that extra 20%. Some folks might be okay with that, but others will seek out other tools. If that meets your goals for the tool then great! But in general, I agree with @ernperkins that an 'advanced' menu could be used in several places in the tool so people would have the option to go past the default settings.

I was under the impression part of the reason for this tool was to combine everything in one place so folks wouldn't need to hit several pieces of software to get the job done. While we seem to be exceeding the minimum necessary details to meet that goal I worry that it will only end up segregating users without including advanced options. For example, I enjoy designing enclosures and enjoy the details that can go into those things. I like to think I'm reasonably experienced in that after however many years (though I have a ton to learn still). Because of that I might be inclined to pass on LoudspeakerLab in favor of other software for box design and simulation. However, I find crossover design pretty intimidating and usually not very fun because of it, and having a starting point (or even several starting points) could be super useful. This is where I could see LoudspeakerLab helping me out. Now I'm using at least two pieces of software again though.

I agree that software like VituixCAD is extremely powerful but not particularly intuitive. Creating something that caters to those new to the field AND to those that already use the VituixCADs of the world may be impossible. But I don't see much out there trying to bridge that gap currently and think there is value in doing so.

 

[wigginjs]

@ernperkins @technodanvan

These are really good points and suggestions. I've added a 'view assumptions detail' modal to the Enclosure Design screen that at least exposes these details to the user for now. In the future, I'll look into have them be configurable as kind of an 'advanced' panel.

I also made some improvements to the vent tuning UX.

 

[ernperkins]

I do applaud you on keeping the UI as simple as possible. I think there is a segment of (potential) DIYers who want a tool that can create a design without knowing all the "gory" details. This forum is useful for getting feedback, but most of us understand/want to know those details. I keep thinking you should try contacting some of the popular Youtube channels (e.g. Toids, DIY Audio Guy, etc.) and have them try out Loudspeaker Lab. It could be a way to get greater exposure to that segment of DIYers. So keep up the good work!

 

[wigginjs]

I do applaud you on keeping the UI as simple as possible. I think there is a segment of (potential) DIYers who want a tool that can create a design without knowing all the "gory" details. This forum is useful for getting feedback, but most of us understand/want to know those details. I keep thinking you should try contacting some of the popular Youtube channels (e.g. Toids, DIY Audio Guy, etc.) and have them try out Loudspeaker Lab. It could be a way to get greater exposure to that segment of DIYers. So keep up the good work!

@ernperkins thanks for the feedback and compliments. One thing I’ve noticed is that when I talk about LoudspeakerLab there some skepticism because it comes across as self-promotion. I suppose it is by definition, but when you create something new, initially there is no one else to talk about it but you.

All of that to say, I agree it would be great to hear feedback and content from some of these respected creators on LoudspeakerLab. Would you be interested in reaching out? I think it would be more meaningful coming from passionate people who aren’t me.

 

[daniii88]

@wigginjs I am trying to simulate my current speaker and have an issue with the cabinet design. It is a cabinet with two chambers, but both are ported - tweeter + midwoofer and woofer. I cannot figure out how do i achieve that, as only the woofer chamber can be ported, the other one is always sealed and I cannot change this.

 

[technodanvan]

I don't think it allows that option currently. Can you share your design components? It's not very common to port a midwoofer.

 

[daniii88]

Its SBA 761 from Troels.

 

[wigginjs]

New video's up: Driver Measurements, High-Value Bookshelf Part 5.
Finally got around to putting the drivers on the infinite baffle rig and running them through the full battery of tests. This round covers the Dayton Audio PCS175-8 woofer, the ND13FA-4 tweeter, and the TD24F-4 tweeter. Frequency response, off axis, distortion, impedance, the whole thing.
Once the data was in, I built new driver profiles for all three in LoudspeakerLab so I can use them for crossover work going forward.
Let me know if any of you have used these drivers and what you thought of them.
https://loudspeakerlab.io/