It's hard to get good measurements! Measure multiple ways, in multiple places. Practice measuring, and understand what is happening. It's so easy to take a measurement and see something that is just a figment. A home environment makes it tricky (getting equipment out, putting equipment away, it's a barrier), I wish I had a big warehouse / gymnasium to do acoustic measurements in.
I think it depends, whether one is more thinking/theoretical vs doing/empirical. I'm kind of down the middle, attentive to (mathy) concepts but not afraid to make mistakes. With a smartphone and free Apps one can learn & do a lot in a short time. For example:
(correlate music and tone sweep)
https://www.diyaudio.com/community/...rements-are-the-way-to-go.414652/post-7726304
(free hearing-test-EQ-curve)
https://www.diyaudio.com/community/threads/full-range-speaker-photo-gallery.65061/post-7675576
(aligning phase and acoustic centers)
https://www.diyaudio.com/community/threads/stepped-baffles.413700/post-7705451
(measure impedance)
https://www.diyaudio.com/community/threads/crossover-inductors-roll-yer-own.395507/post-7261447
(minimalist method)
https://www.diyaudio.com/community/...nge-drivers-and-a-tweeter.391053/post-7143129
(transmission line quarterwave formula)
https://www.diyaudio.com/community/...ile-city-of-22m-laid-flat.393338/post-7267848
(my philosophy)
https://www.diyaudio.com/community/...gns-why-go-for-3-way-then.391639/post-7167945
(correlate music and tone sweep)
https://www.diyaudio.com/community/...rements-are-the-way-to-go.414652/post-7726304
(free hearing-test-EQ-curve)
https://www.diyaudio.com/community/threads/full-range-speaker-photo-gallery.65061/post-7675576
(aligning phase and acoustic centers)
https://www.diyaudio.com/community/threads/stepped-baffles.413700/post-7705451
(measure impedance)
https://www.diyaudio.com/community/threads/crossover-inductors-roll-yer-own.395507/post-7261447
(minimalist method)
https://www.diyaudio.com/community/...nge-drivers-and-a-tweeter.391053/post-7143129
(transmission line quarterwave formula)
https://www.diyaudio.com/community/...ile-city-of-22m-laid-flat.393338/post-7267848
(my philosophy)
https://www.diyaudio.com/community/...gns-why-go-for-3-way-then.391639/post-7167945
Last edited:
(measure coil inductance)
https://www.diyaudio.com/community/threads/values-of-air-core-inductors.407982/post-7572285
(minimalist method link, continued 2 msgs down)
https://www.diyaudio.com/community/threads/values-of-air-core-inductors.407982/post-7572285
(minimalist method link, continued 2 msgs down)
Last edited:
I personally would be wary of using threads from this forum as references for doing loudspeaker design--without some other adjudicated source to reign in the BS that I've personally encountered in the "rules of thumb" that get passed around as if they're truths. I've found that many of these rules of thumb are just plainly in error and are poorly thought out in their supporting design constraints.
Any source that talks about the use of DSP crossovers negatively (especially those that recommend continued use of passive balancing networks over DSP), older in-room acoustic measurement apps, and even general purpose simulation, as well as those sources that shy away from proposing horn loading--I advise avoiding in their entirety. This generally includes information older than 10-15 years when discussing DSP/multi-amping in loudspeaker design. For instance, the Rod Eliot references on DSP and horn loading I find are now seriously dated.
Chris
Any source that talks about the use of DSP crossovers negatively (especially those that recommend continued use of passive balancing networks over DSP), older in-room acoustic measurement apps, and even general purpose simulation, as well as those sources that shy away from proposing horn loading--I advise avoiding in their entirety. This generally includes information older than 10-15 years when discussing DSP/multi-amping in loudspeaker design. For instance, the Rod Eliot references on DSP and horn loading I find are now seriously dated.
Chris
Last edited:
Buy a well respected kit.build it and learn along the way. Also some knowledge is helpfull. Speaker building is easier than speaker designing. The chances of you designing and building a good sounding speaker as a newbie is nxt to nil.
I've also found that many offered kits (flat packs, etc.) are based on designs that are quite dated. Many of these are designed around passive crossover networks only (i.e., no option for DSP use). I would personally recommend sticking to kit designs that were originally created less than 10 years ago.
If the kits once employed horns and the kits themselves are more than 10 years old, I think you will see that many of these horns are no longer available. The use of DSP and in-room measurements using something like REW are enablers of truly outstanding sound quality--even if the supplied horns for the kit are different than originally proposed.
I personally believe that direct radiating drivers in loudspeakers (as opposed to horn-loaded drivers) is something that needs to be minimized in order to create a truly 21st-century loudspeaker design. Another point of poor design practice that I've seen is the reliance on passive network crossover designs, especially for novices. DSP and in-room acoustic measurements are the real enablers of success. The novice DIYer can produce very good sounding loudspeakers first try because of in-room acoustic measurements and DSP.
Conversely, trying to design a loudspeaker around a particular set of existing/owned acoustic drivers on hand and/or driving or dividing electrical/electronic components is usually the formula for high rates of failure.
Chris
If the kits once employed horns and the kits themselves are more than 10 years old, I think you will see that many of these horns are no longer available. The use of DSP and in-room measurements using something like REW are enablers of truly outstanding sound quality--even if the supplied horns for the kit are different than originally proposed.
I personally believe that direct radiating drivers in loudspeakers (as opposed to horn-loaded drivers) is something that needs to be minimized in order to create a truly 21st-century loudspeaker design. Another point of poor design practice that I've seen is the reliance on passive network crossover designs, especially for novices. DSP and in-room acoustic measurements are the real enablers of success. The novice DIYer can produce very good sounding loudspeakers first try because of in-room acoustic measurements and DSP.
Conversely, trying to design a loudspeaker around a particular set of existing/owned acoustic drivers on hand and/or driving or dividing electrical/electronic components is usually the formula for high rates of failure.
Chris
Last edited:
But still extremely valuable with some personal conventions applied.
This thread is very much about orienting.
Quoting Col. John Boyd's OODA loop is about as far from a good DIY loudspeaker design as you can get. I recommend shortening the number of design and manufacturing steps as possible to increase the probability of novice success. Boyd's writing was about aircraft combat maneuvering (something that I spent some years in--in a prior lifetime).
Designing loudspeakers first by the way that they look is another well-known recipe for failure (acoustically and/or economically). I recommend STEM courses in night school instead (and drugs to treat any existing OCPD that comes with judging things and people by the way they look).
Chris
Designing loudspeakers first by the way that they look is another well-known recipe for failure (acoustically and/or economically). I recommend STEM courses in night school instead (and drugs to treat any existing OCPD that comes with judging things and people by the way they look).
Chris
Last edited:
Math can greatly speed you up -- or slow you down. I'm probably the only one on this forum to divide 20khz by capacitance uF without a calculator (1st-order HPF frequency for 8ohm impedance, off by 0.5% from textbook), but these shortcuts allowed me to improvise, test, and tweak a simple crossover in much less time than it took someone to labor over a measurement and computer modeling setup. (Which I would love to do if my situation allowed!) If one wanted to design & diy a mainstream multi-way BR, I think a dedicated workshop/lab is needed to iterate and refine in-test-cabinet measurements, repeatedly. Just grabbing someone else's measurement in a different environment of a different driver specimen, seems to me not that much more reliable than listening to a tone-sweep (very deliberately). And once in room...
... just use a smartphone as source and its built-in earphones-hearing-test-generated parametric EQ profile and voila! Do one for each listening position and switch on-the-fly. My retired ex-flagship phone streaming multi-TB lossless/hirez files to my diy-of-the-evening can sound almost as good as my best ex-high-end gear.
I don't actually recommend passive crossovers for novice DIYers--at least until they have measured and dialed-in their DIY loudspeaker using DSP/multiamping.
After getting that right, they can then easily use the dialed-in DSP settings to propose a much better passive crossover design (all except the minimum input impedance problems introduced by passive monoamping)--but only if the use of passives is absolutely required. The only cases where I've seen that passives are preferred by the buying authorities are in fixed install PA duty and a few cinema surround loudspeakers located at extreme positions from the central electronics/amplifiers, where the length of the wires and the copper involved to connect to a single driving amplifier becomes a nuisance (but not a requirement).
But note that nowadays, active loudspeakers with on-board DSP and integral amplifiers are displacing passive loudspeakers in those type of applications. All that is needed is a small wired signal line or even wireless signal, and electric power (110/220 VAC, etc.).
Chris
After getting that right, they can then easily use the dialed-in DSP settings to propose a much better passive crossover design (all except the minimum input impedance problems introduced by passive monoamping)--but only if the use of passives is absolutely required. The only cases where I've seen that passives are preferred by the buying authorities are in fixed install PA duty and a few cinema surround loudspeakers located at extreme positions from the central electronics/amplifiers, where the length of the wires and the copper involved to connect to a single driving amplifier becomes a nuisance (but not a requirement).
But note that nowadays, active loudspeakers with on-board DSP and integral amplifiers are displacing passive loudspeakers in those type of applications. All that is needed is a small wired signal line or even wireless signal, and electric power (110/220 VAC, etc.).
Chris
OP is asking or orientation to resources based on observation that this forum tends to go hard-mode quickly. Based on that he will presumably make a decision on what to do with the feedback in this thread, possibly put learnings into action, then observe the quality of results and repeat accordingly. My point is that the value of this forum does not need to be disregarded entirely if it's placed within a larger understanding of one's own frameworks for learning.
It's by no means a recipe or speaker design. Substitute OODA loop with any number of frameworks/diagrams for iterative development. Or use none at all and just enjoy the hobby. I've always loved the visual metaphor of the OODA loop and it's been used as a reference in a number of frameworks (scrum, etc.) for that reason. Take it with a grain of salt.
I don’t get it. You are proposing a beginner, a novice, have a computer, software of various kinds which: may not run on their existing computer, may not be free, and they know nothing about it or the underlying science of why it does what it does. They also need a microphone of reasonable quality, two stereo amplifiers instead of one, lots of cables, and a crossover and eq box. Beginners in your world may have unlimited funding and resources but I bet most in the real world do not. For the first kit this is simply too much for the average noob.
I think it would be far more realistic (and better for the hobby) if the budding hobbyist, who enjoys music more than computers, would buy a good proven kit (like a Frugal Horn) using a single high quality driver. Then use reasonable room treatments like drapes, bookshelves and carpets instead of crazy eq and phase dances they know nothing about. You don’t want them chasing shadows instead of learning. Then they can do the most important thing ever, listen.
There are many good kits on the market and a novice can build most with limited knowledge and resources. Then they can enjoy music at a level they wouldn’t have been able to by buying speakers on Amazon.
It depends on what the OP wants from the hobby. If he wants to build just one pair of speakers then putting his stamp (mildy modifying) an existing design that is close to what he is after would seem a good way to go. The kind of speaker and the intended modifications would likely guide the required reading matter which could be fairly specific.
Alternatively if he wants to invest more time and effort in the hobby and sees himself designing and build several speakers with a view to high technical performance then DSP and active crossovers are almost certainly the way to go. They are more flexible, higher performance, quickly reconfigurable for no cost as part of the learning process, reusable with the next speakers and hence likely to be cheaper. Getting on top of DSP requires a fair bit less learning than getting on top of passive crossover design subject to having the basic maths to understand what is going on. Now some people don't have it and don't want to acquire it, fair enough, but the OP has an engineering background and so likely has most if not all of it already.
Opting to learn about active DSP crossovers rather than passive crossovers is likely to be a wise move for many newbies. Not all depending on what they want from hobby but likely most that want to get serious about putting together their own high technical performance designs. Of course learning about speaker design involves more than crossovers but opting to start with active or passive has a significant influcence on the initial hardware purchased, the longer term costs and the learning programme followed.
Alternatively if he wants to invest more time and effort in the hobby and sees himself designing and build several speakers with a view to high technical performance then DSP and active crossovers are almost certainly the way to go. They are more flexible, higher performance, quickly reconfigurable for no cost as part of the learning process, reusable with the next speakers and hence likely to be cheaper. Getting on top of DSP requires a fair bit less learning than getting on top of passive crossover design subject to having the basic maths to understand what is going on. Now some people don't have it and don't want to acquire it, fair enough, but the OP has an engineering background and so likely has most if not all of it already.
Opting to learn about active DSP crossovers rather than passive crossovers is likely to be a wise move for many newbies. Not all depending on what they want from hobby but likely most that want to get serious about putting together their own high technical performance designs. Of course learning about speaker design involves more than crossovers but opting to start with active or passive has a significant influcence on the initial hardware purchased, the longer term costs and the learning programme followed.
Contrary to your comment here, I actually think you do...and you don't like where that takes DIYers nowadays--which likely is away from a comfort zone that you see.
The last time I looked, it is the 21st century (not 1985 as I see many guys here think the current DIY paradigm happens to be).
All the software I use is free or shareware. I think you probably already know that all the software is free, too.
This is a pretty unconvincing counter-argument, since the OP wants to know what sources can take him farther.
Two amplifiers is a lot? Amplifiers for the type of loudspeakers I advocate can easily be chip amplifiers - and cost $5-$10 each. The OP said he's been in it for 30+ years. I don't believe this is an issue. DSP crossovers can now be had for the price of a measurement microphone, and I would never advocate DIY loudspeaker design/builds without a measurement microphone nowadays. REW is free--and powerful. DSP crossovers do all the EQ as-is.
So now you're worried that it's too hard? I'm not. And the OP isn't an "average noob".
I think you went wide of the mark here. Computers are in everyone's pocket. And my focus is on better sound quality than what you propose. The chances of true success in achieving good sound quality is clearly higher than what you apparently propose, too.
And my reason for spending time in this pastime is to listen--not to fiddle about just to fiddle about (as I suspect many others here are actually doing, and that's okay if they want to use their own time on activities that may or may not result in better sound quality for far less money). My background is as a musician (classical).
Chris
Last edited:
Three things in life are guaranteed... death, taxes and active/DSP guys will promote active/DSP.
I remember the thread where someone posted that the OP should go active...in a thread titled "help with passive crossover".
I remember the thread where someone posted that the OP should go active...in a thread titled "help with passive crossover".
One of the good things about using acoustic measurements and DSP crossovers is that the DIYer can measure what they are seeking from the loudspeakers (i.e., better sound quality) and even post their measurements here or on something like Google Drive, etc. for others here to help them dial everything in better, I've been helping dozens of guys do this over the past decade or so. Not one of them failed to achieve much better sound quality with advice on the DSP settings based on their own in-room acoustic measurements.
Furthermore, I find that packages like REW itself provide ways to help the "experts on-line" to see the quality of the novice's acoustic measurements, as well as the macroscopic acoustic conditions in-room (RT measurements, etc.). This is a great ENABLER of better results from DIY efforts.
Chris
Furthermore, I find that packages like REW itself provide ways to help the "experts on-line" to see the quality of the novice's acoustic measurements, as well as the macroscopic acoustic conditions in-room (RT measurements, etc.). This is a great ENABLER of better results from DIY efforts.
Chris
Wow! Defensive much? I’m simply proposing a different approach for a novice to attain a high level of performance with out a large financial commitment, without tech he may not understand, and get it done quickly. I firmly believe in the KISS Principle. A good 5-6” full range driver can provide exceptional performance for very little cost. This can get the novice excited about the hobby, they now have a speaker system better than 95% of people in the country. Then they can experiment, read all the literature, and buy more stuff as the budget allows.
Lets get back on topic and stop the philosophy debate.
Lets get back on topic and stop the philosophy debate.
Well this thread only highlights that an evidence based guide on the engineering of loudspeakers as a whole almost certainly doesn't exist or else it would have been referenced. There are a few good guides on aspects but none that I am aware of addressing the whole. The technical performance side is straightforward for those with an engineering background but the relevant information is in text books and not guides for nontechnical speaker hobbyists. Of course many are enthusiastic about aspects of loudspeaker design which are in conflict with what is required for a high technical performance (e.g. wideband drivers, passive crossovers, small cabinets, etc...) and it would be a rather dry and dull hobby if they weren't.
So if a newbie wanted to know beforehand how much technical performance would be lost by getting their hands dirty with a small wideband speaker or a small 2 way passive where would they look for reliable guidance? Lots of sources for unreliable enthusiastic subjective advice as can be seen in this thread but where to look for reliable quantitative guidance? If instead they wanted to go all in with a high technical performance full range design where to look to get reliable advice on what is cost effective rather than enthusiasm for expensive prestige components? The latter is of course a valued part of the hobby for a fair few but I would suggest it rather goes against the spirit of DIY for many which involves maximising bang for the buck.
So to answer the op's question there are good guides around on details but there are no reliable guides of which I am aware on how best to get going from scratch. It would make a good project but is a pretty challenging one. It tend's to be why the more senior professors are often noticeably better at introductory lectures than junior ones but is less the case with the more specialised subjects in later years.
So if a newbie wanted to know beforehand how much technical performance would be lost by getting their hands dirty with a small wideband speaker or a small 2 way passive where would they look for reliable guidance? Lots of sources for unreliable enthusiastic subjective advice as can be seen in this thread but where to look for reliable quantitative guidance? If instead they wanted to go all in with a high technical performance full range design where to look to get reliable advice on what is cost effective rather than enthusiasm for expensive prestige components? The latter is of course a valued part of the hobby for a fair few but I would suggest it rather goes against the spirit of DIY for many which involves maximising bang for the buck.
So to answer the op's question there are good guides around on details but there are no reliable guides of which I am aware on how best to get going from scratch. It would make a good project but is a pretty challenging one. It tend's to be why the more senior professors are often noticeably better at introductory lectures than junior ones but is less the case with the more specialised subjects in later years.
Isn't it "phase" that's highly non-intuitive and its sonic effects unteachable on paper alone? So the lessons have to be high-fidelity aural, which only became mass-available with smartphones (trillion-dollar R&D/manufacuring-at-scale) and earpods (are we there yet?).
And the first part of such a lesson has to be instructing how to run the phone's built-in hearing test in order to EQ-flat the lesson material. Kind of a chicken-and-egg problem no? (Ironic the rest of the lesson becomes mostly unnecessary.)
And the first part of such a lesson has to be instructing how to run the phone's built-in hearing test in order to EQ-flat the lesson material. Kind of a chicken-and-egg problem no? (Ironic the rest of the lesson becomes mostly unnecessary.)
Last edited:
Hey friends, as the OP I can speak up a little to clarify MY intentions:
I am very much looking to dive into this hobby and am willing to invest some dough to get a respectable workbench going to have a greater chance of success. In fact, I have bought a calibrated measurement mic which I intend to use with the Audiotool app. I'd love to get an oscilloscope, signal generator and LCM meter but some workarounds exist. For the moment I'm going to keep it simple. I downloaded XSim and started playing with it this week.
I probably have more knowledge than most newbs being a musician and engineer and know how to solder. There are lots of good points made by folks here and some healthy (?) debate.
I will consider DSP but for the moment sticking to passive XO on my current project. I am aware of the advantages as I rewired my van with a Kicker Key for the front and it's rather cool!
I am very much looking to dive into this hobby and am willing to invest some dough to get a respectable workbench going to have a greater chance of success. In fact, I have bought a calibrated measurement mic which I intend to use with the Audiotool app. I'd love to get an oscilloscope, signal generator and LCM meter but some workarounds exist. For the moment I'm going to keep it simple. I downloaded XSim and started playing with it this week.
I probably have more knowledge than most newbs being a musician and engineer and know how to solder. There are lots of good points made by folks here and some healthy (?) debate.
I will consider DSP but for the moment sticking to passive XO on my current project. I am aware of the advantages as I rewired my van with a Kicker Key for the front and it's rather cool!