Why simple crossovers, tuned by ear, don’t work

Passives can provide gain in certain situation(s)

korpberget: From a general approach, basic truth. However, say in a three way you achieve a true 2nd order solution, you can gain as much as 2.45 db from a passive crossover for the mid range driver. Gotta watch those absolute statements.
 
Just my point. My hearing is a mess and I have tinnitus on both ears. A flat measuring speaker will not sound good in my ears.
Thats why I tweak by ears, just using simple clacs as a starting point to work from.

Without a correctly designed crossover, the frequency response could be going in and out of phase throughout the crossover region and cause peaks and dips in random places. I'd rather start with a flat response, with the phase tracking nicely through the crossover and then adjust the response to suite my ears.
 
You build yours your way, I'll build mine my way. We will both enjoy our hobby together and be happy for each other doing things the way that we each see fit. There is room for many perspectives here.
While that's too relativistic for my tastes, we can agree their is a hierarchy of goodness with trade-offs of time, effort, cost, etc.

But these have changed over time as gear has changed. I'd say passive crossovers and systems that aren't bi-amped have moved way down the scale as DSP, acoustic testing, and cheap bi-amping have become available.

Likewise and in reply to fatmarley, it becomes less and less wise to fuss and waste time over passive crossovers as their results move down the scale and as systems to show off to your cousin, no better than a table radio (if they still exist). So, just use "6.3" and "160", wire it up... and then learn more about DSP.

You are only half way done when you screw the back on your speaker cabs.

B.
 
Last edited:
The one thing that I can't help but notice about all the comments in support of being able to design a crossover without measurements is that they are coming from people with a lot of history in audio and in designing and building crossovers: people who have the experience to know what they are hearing and how to tweak the design for improvement.

I doubt that those are the people the OP is referring to.

How is the amateur going to take it if the speakers they build don't sound good to them? Most of the "I want to build a speaker, but I don't know what I'm doing" threads I've seen have been by people who want to build *a* set of speakers, not take up a new hobby.

A designer can voice their speakers however they want, but is an amateur working on their first project more likely to identify that, say, their speakers sound weird because the crossover isn't summing properly more likely to come to that conclusion by ear or by graph? At the very least, a free RTA smartphone app to see if crossover slopes are anywhere near right would be a huge help to the first time designer.
 
The one thing that I can't help but notice about all the comments in support of being able to design a crossover without measurements is that they are coming from people with a lot of history in audio and in designing and building crossovers: people who have the experience to know what they are hearing and how to tweak the design for improvement.

I doubt that those are the people the OP is referring to.

How is the amateur going to take it if the speakers they build don't sound good to them? Most of the "I want to build a speaker, but I don't know what I'm doing" threads I've seen have been by people who want to build *a* set of speakers, not take up a new hobby.

A designer can voice their speakers however they want, but is an amateur working on their first project more likely to identify that, say, their speakers sound weird because the crossover isn't summing properly more likely to come to that conclusion by ear or by graph? At the very least, a free RTA smartphone app to see if crossover slopes are anywhere near right would be a huge help to the first time designer.

Yes, thank you. That's exactly who i'm aiming this thread at. Noobs who want to build "a" set of speakers.
 
Without measurements I won't be able to create a very good sounding loudspeaker and also without listening I won't be able to create a very good sounding loudspeaker. It's a combination of the two that works best for me.

Measurements let you see where the problems are: suppose the highs sounds too prominent or fatigue. Without measurements you probably attenuate the tweeter but in reality it's the cone breakup of the mid that should be attenuated. Measurements let you also design good phase integration between the drivers at the crossover region, check the impedance, check the frequency range of the individual drivers, their slope, ...

After getting the drivers to play in their optimal FR range, suppression of the cone breakups and good phase integration between the drivers, then I listen to a variety of music (100 different songs) to set the level of the different drivers and their FR range so it sounds natural and very good to me. But still using measurements to keep the phase correct, impedance not too low, no strange FR peaks or dips.

As a result I often come to a frequency response with 2-3db more bass then mids and highs. It's in the line of the response curves like Troels in his designs

Measurements let you see in the dark and listening let you create a good sounding loudspeaker.

There's no reason not to use measurements, the price of a microphone is low and excellent software like REW is free.
 
.......... The better you get at crossover design, the more these bad recordings become good recordings.

I very much agree with that.:)

I do not even understand when someone comments on the result of their construction as "sounds great with good recordings, but highlights the imperfections of old recordings" ...:rolleyes:
I think they expected their work to include a contract with a full-time magician....:D
 
Is this whole thread in response to the long running sticky? Introduction to designing crossovers without measurement

No. I usually leave a link to that for noobs because it shows that speaker design is a bit more complicated than they usually think. Although I think it gives a good grounding for noobs, I still think they need to know that to beat even cheap (but good) budget design takes more than that (assuming same box size).

I'd be amazed if anyone could beat a good, classic, budget design like the Kef coda 8 / Mission 760i or Mission 731i with a similar sized speaker and without measurements.

I'll probably get a million responses saying these designs i've mentioned a are rubbish but that's the beauty of subjective opinion.
 
<snip> We were asked about the phone cart used, and told to use a Decca. We were asked about what amplifier, and told which one to use. Finally, we were told what album to play. On that album, they sounded rather OK...... Not on anything else. This is one of the likely failures of tuning by ear.....

This is exactly how not to listen. I have found over the decades that one needs to listen to pink noise, the spoken voice and varied music. I find violin and piano very revealing.(I play the former, so am familiar with the sound) And I always compare with a decent commercial speaker, and decent headphones.

For me, the point of making speakers, is making speakers I like the sound of. That is the reason for diy. If I wanted something that just measured well, I'd buy a good pair of commercial speakers 2nd hand.
 
I usually leave a link to that for noobs because it shows that speaker design is a bit more complicated than they usually think.

I totally agree. And FWIW, I mostly agree with post#1. It's not that they don't work, it's that they may not work well.


This is what led to my input, and I think it's a frightening abyss for a newbie to be referred to, unless your goal is to scare them away:

This page is a good start...

Quote: "The question is whether we can make a crossover at all without measurements - and the answer is NO.
It cannot be done, and crossovers cannot be calculated."

Yes, you need a DCR measurement at minimum, and no they won't be optimum, but don't say it cannot be done. That's simply not true and I guarantee there's more than one in this group who started out with no more than that.
 
This is exactly how not to listen. I have found over the decades that one needs to listen to pink noise, the spoken voice and varied music. I find violin and piano very revealing.(I play the former, so am familiar with the sound) And I always compare with a decent commercial speaker, and decent headphones.

For me, the point of making speakers, is making speakers I like the sound of. That is the reason for diy. If I wanted something that just measured well, I'd buy a good pair of commercial speakers 2nd hand.
Yep, that's the reason why I diy. I want the speakers to sound exactly how I want. I couldn't do it without the help of measurements and software though (Maybe I could do it on paper but I couldn't be bothered)
 
This is what led to my input, and I think it's a frightening abyss for a newbie to be referred to, unless your goal is to scare them away:

No, I don't want to scare them away - the more the better. I just don't like the idea of someone spending a fortune on drivers and only getting a bad sound because they wanted instant, excellent results and then giving up when they realise it's more complicated that they first thought.
 
The above "how to design crossovers without measurements" is very useful, but it assumes someone is committed to designing a proper crossover.

I think we could come up with a "Read me first: I want to build my own speaker FAQ".

Some starting questions:
  • Can I just choose my drivers and worry about the crossover later?
  • Can I use on-line crossover calculators to find my values for inductors, capacitors and resistors?
  • Can I buy a pre-made crossover designed for a specific crossover frequency that matches the one I intend?
  • I want to build my own speaker and not follow a kit or someone elses design. Should I?
  • I don't have any graphs of impedance or frequency response for my drivers. Can I just design a crossover anyway?
  • I want to start simple and just design a 2 way, but want really good bass. Can I use a large 8" or greater woofer?
  • I hear 3 ways are better than 2 ways. Can I start my first design with a 3 way speaker?
  • I see some capacitors are really expensive. Is it worth it?