VituixCAD

Nothing wrong with starting with traced data as long as ultimately its replaced with measured data.

So let me get on the bandwagon - don't even think about reading the sticky thread on this forum about designing crossovers without measurements!

Agreed. I started into DIY speaker design as a teenager. From many mistakes and lessons learned, there is an urge to dissuade new designers to use manufacturer or 3rd party data, and to get a mic, build a 2 channel jig and start generating your own data. I wish I hadn't thought single axis design was good enough (mind you the DIY tools were much more limited back then), I wish I hadn't started designing with a USB mic, and I definitely wish I was never told that tracing a manufacturer sheet is good enough.

The lesson to be learned on manufacturer data is that some manufacturer data is relatively accurate, some is pretty useless. How is a newcomer to know? Dayton is one of the worst offenders for poor datasheet information to be honest. And still, if the data is simply traced, acoustic offset is a missing factor completely, so only a guess can be used for design. I certainly wouldn't want to spend all the time to build a cabinet and purchase drivers and have the crossover performance dependent on a guess.

There is a lot to be learned from simply using a mic and 2-channel setup whether for the purpose of designing a loudspeaker or not. Locking FFT window start provides understanding of how phase changes with acoustic delay. Gating measurements provides understanding of reflections and room interaction. Measuring your own speakers provides reference information of measurement information vs the acoustic impressions you have.
 
@DcibeL

Very similar story here, but instead I am actually really happy I made all those mistakes.
And I certainly don't regret going that route and taking the "perfect" short-cut instead.

Why? Well, because it gives very valuable inside what things can go wrong, but even more so how much it can go wrong, or how big certain errors can be.
In other words, you develop (practical) experience and knowledge actually THINKING about problems.

Doing things the right way first try is nice, except one doesn't learn anything from it.
And you certainly don't develop experience from it.
When it comes to knowledge and experience; failing more the better.

About useless data, I can literally write whole books about it (there has been one actually on the backburner / to-do list for a while).
But there are many many examples, even from the more reputable brands.

Many datasheets still contains an insane amount of reflections in the freq resp.
Resulting in many weird dips and peaks. Often masking all kinds other issues.
Take the datasheets from SB Acoustics for example. As much as I like their quality, performance and price, their datasheets are pretty useless.

Something I don't understand.
If one has an anechoic chamber, why not combining them with gated measurements?
That will literally just only improve the data, never worsen it.
 
Now that we've got all that out of the way can I ask a different question of you guys. I respect all of your experiences above any I have.

Where I'm at. I've build others designs in the past with varying success. Sometimes great success but never designed my own. I am an artist and spend all day in a studio listing to music. I have built the amp and the flac streamer and current speakers (many years ago) which were repaired badly recently and now sound crap. I was close to just building a kit as most of the speakers I've built so far were better that spending the same money on a commercial product. I like to diy.

Then it occured to me that I work in various locations in the room ( 12m x 7m 3.5m-4m high) so something like an Omni would maybe give better sound in more locations in the room. I'm pretty sure this is the right direction but there are no kits I can find for an Omni...in my current budget. I have spoken to those in other groups who have amazing Omni builds. But these are way too complex and expensive for me.

I quite understand you have been discussing here the correct method of design for the best results but I'm pretty sure that even wanting Omni is not a correct decision for the 'best" results. I just want better than I have and not a single point in the room to listen. Sound will be bouncing around on shelves and tables causing all kinds of accoustic mayhem .

So I came across vituixCAD. I tried a few others but I came closer to understanding vituix than other software.
Now I'm usually an all in type of guy but I don't think speaker building will be my life's passion. I have a lot going on, but I'm gonna build something. Hence me not currently owning test equipment. I want it, but have other priorities right now. At some point Ill need to get test equipment if I build my own design.

So maybe you've all convinced me to give up because I can't do it perfectly. Well this time I'm not after something I'd put in my lounge. If there was an existing design id build it. I will build something and was just after some help from you guys who have the knowledge .

To make things difficult I'm in NZ where there are very few drivers and components available. I'd consider freighting from anywhere if the solution was perfect but in my budget the freight often is more than the parts. So that is why I chose those specific cheap drivers as I can source from Australia..

So CAN anyone offer advice or point me somewhere else maybe more suited to my problem. Maybe I should give up vituix and my own designs for now? Maybe I can build some compromised design that is still better than my current work studio speakers?
 
Last edited:
Where I'm at. I've build others designs in the past with varying success. Sometimes great success but never designed my own. I am an artist and spend all day in a studio listing to music. I have built the amp and the flac streamer and current speakers (many years ago) which were repaired badly recently and now sound crap.

That is a lot to unpack:) so please bare with us. Since a number of "us" have been doing this for a very long time.........even for us it is difficult to get "correct" from scratch.

I've found one of the harder questions to ask is what topology will "do it" for me? Unless you've heard all topologies you don't know. What SPL do you require? If you are listening to all points in the room then you will probably leave critical listening out of the equation since stereo needs 2 speakers at near equal distances to your ears. Can your room handle corner horns? Too many questions to ask in this thread but maybe you point us to your speaker thread and many would be more than happy to nail down the requirements of a speaker.

There are many speakers designs that would do it for you but hard to nail down what you need for your space. That could be anything but we'd need a starting place. That and how is bass going to be handled? In order for things to get close to a nice sounding the system must be considered as a whole and not just the components of the loudspeaker. Room modes, distributed bass modules, SPL requirements ect.

Not to discourage but there is a very steep learning curve. If you don't have the time or want to learn then the outcome will be subpar. If you do want to learn it is a very rewarding journey.
 
So what is good enough? My current (id say broken) speakers are a full range (local driver) and a vifa super tweeter. It used to sound at least acceptable for a work system but now it falls off above 100Hz and has a nasty breakup pek in the 5-7 k area since the recone, recoil. Music just sounds thin and scratchy...though clear, through these. I listen to all types of music though these from pop, rock, classical, jazz folk, disco, funk, world and everything else except death metal.

Although I wont be doing critical listening in this room I completed formal training as an audio engineer when younger and did run my ownrecording studio, which for a while used my DIY near fields succesfully... So when i say it falls off above 100Hz and it peaks 5-7k...thats just my ears in the room saying that...with a few test tones.


My amp (a gain clone) would like 8 ohms at as high a sensitivity as i can get but id settle for 86-87 dB sensitivity at the lowest, I'd aim for 88-90... hopefully in this budget.

I'm never playing party levels but sometimes loud enough so conversing is not easy.
I dont want a sub at this stage. I want the amp to handle it, and if i get to 40Hz I'd be happy...maybe even 50... anything lower is a bonus. I dont want or require a room shaker.

The room is foremost a workspace and location of things within it are decided not by acoustic requirements. The speakers are currently high up a few metres from the corner on the short end (7m) of the room angling down giving aquite large "OK" spot in the centre but im rarely in that spot. looks is also not an issue, so I'll spend zero time making these look good. Id thought that omnis might get a better overall dispersion at halfway along each long walls (12m)
 
In my experience most people ask the wrong question.

The question should never be "what is better".
But you just have to approach each situation as something unique
I am personally also no believer at all that there is only one great or "perfect" system.
It all depends on context.

So just start making a list with requirements and start to prioritize them.
You will see that quite quickly a big mountain of options starts to actually be very manageable.

I used to work an live in NZ as well, getting proper stuff can be quite the challenge.
Although it seems that Wagner sells SB Acoustics as well as Dayton and stuff.
(that wasn't available when I was there :( )
That shouldn't be to hard to ship to kiwiland?

btw, maybe open your own topic (or just send a PM :) )
 
I agree with those who say manufacturers frequency response graphs are good for initial eyeballing only not for design, so tracing the measurements and importing them is of marginal value, even in the early "will this work or won't it" stage of planning when you're still choosing drivers.

With a bit more expensive you can usually look at the data provided with drivers by eyeball and see if they are suitable or not or will work together or not. No need to simulate based on potentially iffy data as garbage in means garbage out.

As was noted further up thread - in a passive design two very important criteria are sensitivity matching and the "complexity" of the response of the drivers.

You never want to be in a situation where you have to attenuate a woofer in a passive design - if you are, you chose the wrong drivers. The woofer should always be the least sensitive drivers with both midrange and tweeter having a few dB higher sensitivity so you have "room to work" with a passive crossover to shape the response of those drivers as you can (mostly, except for resonances that cause impedance dips at the input) only subtract from the drivers raw response with the crossover not add to it.

A smoother less complex response is also fairly important for passive crossovers. All crossovers double as "shaping" filters to modify the non-flat (in a cabinet) response of drivers to be closer to flat. With higher order crossovers its often possible by deliberately "detuning" the crossovers to achieve the desired correction without adding additional components or only a small number of additional components.

The smoother and less complex the response of the driver is the easier this is to do with a passive network. If the driver has complex response aberrations theses quickly get infeasible to deal with passively and you have to put up with them in the final result.

As also inferred further up the thread - even if manufacturers data was accurate, high resolution (not "airbrushed" to look smoother) and measured under standardised controlled conditions such as an infinite baffle that still isn't enough to design a high quality crossover from.

The cabinet modifies the response of all drivers so it is impossible to design a crossover until the cabinet design is finished because you need the measurements from the drivers in the finished cabinet to do the design as the cabinet is an integral factor in the drivers response.

If you're doing a prototype cabinet that is "near" what you think the final design is that's fine - design a crossover based on that to test your prototype cabinet and establish the validity of the overall design. But be aware that if you make any changes to the cabinet between the prototype and final design the crossover will have to be redone.

If you're lucky you may be able to simply re-optimise all the component values without actually changing the circuit design as such (and this is quite easy to do in Vituixcad even if you hand optimise the values as the graphs change in real time) but if cabinet changes are more extensive including interior changes like midrange enclosure volume you may have to do considerable design rework on the crossover.

Finally, if you're new to speaker design you have to accept that you may end up choosing drivers and a cabinet design that just don't work out in the end no matter what you do with the crossover.

Sometimes it will be because the drivers don't live up to their claims, sometimes because you're inexperienced. Sometimes you just can't be sure how it's going to turn out until you can measure/listen to the drivers yourself and by that time you've already bought the drivers and you're stuck with them.

Take it in your stride, optimise and finish the design as best you can (it's amazing what a good crossover can do with less than stellar drivers) then start a new design with different drivers having learnt from the failure of the previous design.

Like anything, speaker design is not all about theory, there's a lot of experience and real world practice that goes into getting a good result.

PS: Hello from a Northland NZ expat... :)
 
Last edited:
  • Like
Reactions: 1 user